System and method for distributed utility service execution

Description

BACKGROUND

The present teachings relate generally to utility services. For example, the present teachings can relate to assisted delivery of goods originating at distributed establishments and destined for customers located in the vicinity of the distributed establishments. What is needed is a system that can accommodate trips of various lengths, and can solve the problem of short-distance assistance to customers. What is further needed is a system that can accommodate semi-autonomous and autonomous operation, and can deliver utility services economically.

SUMMARY

The utility system of the present teachings solves the problems stated herein and other problems by one or a combination of the features stated herein.

The system of the present teachings can be part of a fleet network of similar systems. The fleet network can also include trucks, planes, cars such as self-driving cars, and business establishments. All members of the fleet network can communicate seamlessly to share, for example, but not limited to, navigation data, dynamic objects, alternate routing, and utility requirements including utility characteristics, customer location, and destination. The system of the present teachings can interface with existing truck systems so that the fleet is seamlessly connected. Piloted utility vehicles can include technology disclosed in U.S. patent application Ser. No. 15/600,703 filed on May 20, 2017, entitled Mobility Device.

The utility robot of the present teachings can operate in an autonomous or semi-autonomous mode. The autonomous utility robot can, in conjunction with the network, control its movement without the assistance of an operator. The semi-autonomous utility robot can include technology that can receive and process input from the operator of the semi-autonomous utility robot. The input can, for example, but not limited to, override autonomous control of the utility robot, or be considered in controlling the utility robot, or be ignored. The utility robot can include a set of sensors appropriate for the location of the utility robot. For example, when the utility robot is deployed in an environment that includes many other members of the fleet network, the utility robot can include a first number of sensors. In some configurations, for example, in an environment that includes a relatively small number of members of the fleet network, the utility robot can include a second number of sensors. The sensors can operate in conjunction with sensors that are associated with other members of the fleet network. In some configurations, the utility robot can include enough physical storage space to accommodate delivery items from typical distributed sources such as pharmaceuticals, food, meals, and documents. The utility robot can operate on city sidewalks, and near and within buildings, among other places. The utility robot can include the capability to determine a current location and situation of the utility robot (localization), through the use of, for example, but not limited to, fiducials, sensors, external application data, operator input, beacons, and physical orientation of the utility robot. The utility robot can plan a route to reach a desired destination, detect obstacles along the route, and dynamically determine specific actions that the utility robot is to take based on the route, current location, and obstacles. Obstacles can include, but are not limited to including, dynamic (mobile) obstacles, such as, for example, but not limited to, pedestrians, vehicles, animals, and static obstacles such as, for example, but not limited to, trashcans, sidewalks, trees, buildings, and potholes. The utility robot can accommodate map matching including locating obstacles visually and matching them to other data such as, for example, satellite data. The utility robot can determine preferred routes and routes to be avoided. In some configurations, the utility robot can climb curbs. In some configurations, the utility robot can climb stairs. The utility robot can achieve stabilized operation while on four wheels, including while climbing stairs. The utility robot can maintain a pre-selected distance, which could vary along the route, from an obstacle such as, for example, but not limited to, a building. The utility robot of the present teachings can be driven by an operator who is seated upon a seating feature of the utility robot. In some configurations, the utility robot can take the form of a wheelchair, and can thus legally traverse sidewalks in all jurisdictions. The utility robot can accommodate disabled operators, and can include carrying capacity for, for example, but not limited to, pizzas and pharmaceuticals. In some configurations, the utility robot can follow rules of the road to maintain the safety of the utility robot, the operator of the utility robot (when present), and the people and obstacles encountered by the utility robot. The rules can include, for example, but not limited to, what to do when encountering an obstacle and what to do when crossing a road. For example, the rules can include prohibitions on rolling over someone or something, and traveling into unsafe places. The rules can also include prohibitions on stopping in unsafe locations, for example, the middle of an intersection. In general, safety protocols can be established and learned by the utility robot of the present teachings.

The utility robot of the present teachings can serve many purposes. The utility robot of the present teachings can be summoned to assist an individual in carrying heavy things, for example, to a bus stop. In some configurations, the utility robot of the present teachings can watch for threats and odd occurrences, and can be summoned to escort individuals from place to place. In some configurations, the utility robot of the present teachings can be summoned by a mobile device, to a location that can change between the summons and the rendez-vous of the utility robot and the mobile device. The utility vehicle can transport items from one location to another, for example, from a pharmacy to the residence of the person ordering the pharmaceuticals. The utility robot can communicate with pedestrians and vehicles, for example, by gesturing and providing awareness feedback.

In some configurations, the utility robot of the present teachings can travel at least fifteen miles at sixteen miles/hour on a single battery charge. The utility robot of the present teachings can use GPS, road signs, stereo cameras, cell phone repeaters, smart beacons with steerable RF beams that can direct the utility robot along a desired route, IMU data between beacons, and other beacon data to help the utility robot to recognize and traverse the desired route. In some configurations, at least one autonomous utility robot of the present teachings can be coupled, for example, electronically, with at least one semi-autonomous utility robot. Batteries can include quick change/quick charge batteries. In some configurations, batteries can be protected from being stolen. The batteries can be locked down, for example, or they can include an identification number that is required to enable the batteries.

The utility robot of the present teachings can accommodate such numbers and types of sensors as are necessary for the function of the utility robot. For example, the utility robot, when operating in an urban area, can expect to receive real time data relevant to its travel path from other members of the fleet network such as, for example, but not limited to, beacons and fiducials. Thus, the utility robot, when operating in an urban area, can include a sensor package appropriate for its environment. The same utility robot, when operating in an area that includes fewer fleet members can include a sensor package appropriate for its environment, and possibly different form the urban area sensor package. Sensors can be integrated with the utility robot of the present teachings. The sensors can access and/or collect street/building/curb data, and can include, for example, but not limited to, visual sensors, LIDAR, RADAR, ultrasonic sensors, and audio sensors, and data from GPS, Wifi and cell towers, commercial beacons, and painted curbs. The visual sensors can include stereoscopic visual sensors that can enable object classification and stop light classification, for example. In some configurations, visual sensors can detect curbs. Detection of curbs can be simplified by painting the curbs with substances that can include, but are not limited to including, reflective materials and colors. Curbs can also be painted with conductive materials that can trigger detection by appropriate sensors mounted on a fleet member such as the utility robot. LIDAR can enable the creation of a point cloud representation of the environment of the utility robot, and can be used for obstacle avoidance, object classification, and mapping/localization. Maps can contain static objects in the environment. Localization provides information about the locations of static objects, which can be useful in recognizing dynamic objects. Audio and/or ultrasonic sensors can be used to detect the presence of, for example, but not limited to, vehicles, pedestrians, crosswalk signals, and animals, and can enable collision avoidance and semi-autonomous driving. Ultrasonic sensors can enable calculation of the distance between the utility robot and the closest object. In some configurations, the utility robot can accommodate repositioning of the sensors upon the utility robot. For example, sensors can be positioned to accommodate the variable placement of storage containers on the utility robot.

In some configurations, vehicles, such as, for example, but not limited to, trucks and self-driving vehicles, can transport the utility robots of the present teachings closer to their starting locations and destinations, and can retrieve the utility robots to remove them to storage, charging, and service areas, for example. With respect to trucks, in some configurations, as the utility robots can enter the trucks, their batteries can be removed and be replaced with fully charged batteries so that the utility robots can continue their services. The truck can include the capability to swap out batteries and charge them. In some configurations, empty storage compartments can also be filled on the delivery truck, and the utility robot can be sent from the truck to perform further deliveries. The utility robots and trucks can locate each other wirelessly. A dispatching mechanism can couple trucks with services and batteries with utility robots that need them. The trucks can include at least one ramp to receive and discharge the utility robots of the present teachings.

In some configurations, the movement of trucks and utility robots of the present teachings can be coordinated to minimize one or more of service costs, service times, and occurrences of stranded utility robots. Service costs may include fuels for trucks, battery costs for utility robots, and maintenance/replacement costs of trucks and utility robots. The trucks can include on- and off-ramps that can accommodate rolling retrieval and discharge of the utility robots. The trucks can be parked in convenient places and the utility robots of the present teachings can perform services in conjunction with the trucks. In some configurations, the trucks and utility robots can be dynamically routed to meet at a location, where the location can be chosen based at least on, for example, but not limited to, the amount of time it would take for the fleet members to reach the location, availability of parking at the location, and routing efficiency. In some configurations, the utility robots of the present teachings can be moved from place to place, depending upon where they are needed the most, by, for example, the trucks. Daily schedules can control where the utility robots of the present teachings are transported. For example, a truck can pick up the utility robot of the present teachings when the utility robot has completed its services and/or when its batteries need to be charged, and/or when it needs service. The utility robot can automatically remain in the location of its final service until a truck arrives to retrieve it. A truck can be used to transport the utility robot of the present teachings from a station such as a store where goods and services have been purchased to a retirement home, for example, where the goods and services are to be delivered. The utility robot of the present teachings can be dropped off at, for example, the retirement home at which time the utility robot can deliver the goods and services. In some configurations, a first of the utility robots of the present teachings can deliver parcels to the truck, and those parcels can be removed from the first of the utility robots to the truck. The parcels can be picked up by a second of the utility robots of the present teachings that is heading towards the delivery destination of the parcel. The utility robots of the present teachings can be deployed from moving trucks or other moving vehicles.

In some configurations, self-driving vehicles can be fitted with controls and hardware that can accommodate the utility robot of the present teachings. Self-driving vehicles can be more ubiquitous in and adaptable to urban settings than trucks. For example, a utility robot of the present teachings can receive goods to be delivered, summon a nearby self-driving vehicle, move to meet the vehicle, enter the vehicle, and become docked in the vehicle. The battery of the utility robot of the present teachings can be charged during the delivery trip by the self-driving vehicle. The self-driving vehicle, as part of the fleet, can access the service information for the utility robot from which the summons came, and can move the utility robot of the present teachings to the service destination(s).

In some configurations, at least one semi-autonomous utility robot can be associated with at least one autonomous utility robot. The semi-autonomous utility robot and the autonomous utility robot can wirelessly communicate with each other to maintain synchronous behavior when desired. In some configurations, the group of utility robots can form a secure ad hoc network whose participants can change as autonomous utility robots enter and leave association with the semi-autonomous utility robot. The ad hoc network can communicate with the fleet network. In some configurations, the utility robots can communicate by, for example, wifi, through standard electronic means such as text, email, and phone. In some configurations, each of the utility robots can share features of the route upon which the group travels by individually measuring wheel rotations and inertial values and sharing those data. The group of utility robots of the present teachings can arrange to meet a truck. The arrangement can be made by a cellular telephone call to a dispatcher, for example. A dispatcher, which may be automatic or semi-automatic, can locate the truck that is nearest the group of utility robots of the present teachings and can route the truck to the location of the group. In some configurations, a meetup request can be generated by one or more utility robots of the group, and can be electronically transmitted to trucks that come within wifi and/or ad hoc network range of the group of utility robots. In some configurations, the group of utility robots can be in continuous electronic communication with the fleet of trucks, can monitor their whereabouts, and can summon the nearest truck and/or the truck with the appropriate specifications such as, for example, size and on/off ramps. In some configurations, summoning the one or more of the utility robots of the group of the present teachings can automatically involve summoning a utility robot with the correctly-sized storage compartment(s) for the parcel(s), and the utility robot that is geographically closest to the pickup point for the parcel(s).

In some configurations, the utility robot can include storage for items to be delivered, and can track the sizes of storage containers on each utility robot, as well as the sizes of the contents of the storage containers. The utility robot can receive the size of the package and can determine if the package can fit in any available storage in the fleet of utility robots of the present teachings. The storage can be compartmentalized for security and safety of the contents of the delivered goods. Each of the compartments can be separately secured, and the sizes of the compartments can vary according to the sizes of the parcels. Each of the compartments can include, for example, a sensor that can read the address on the parcel and ensure that the parcel is sized correctly for the storage container and the utility robot. For example, a drug store might require several small compartments to house prescription orders, while a restaurant might require pizza-sized compartments. In some configurations, the utility robot can include operator seating, and the storage compartments can be located behind, above, beside, in front of, and/or under the operator, for example. The storage containers can be sized according to the current parcel load. For example, the storage containers can include interlockable features that can enable increasing or decreasing the interior size of the storage containers. The storage containers can also include exterior features that can enable flexible mounting of the storage containers upon the chassis of the utility robot of the present teachings.

In some configurations, the utility robot can include storage compartments and can accommodate long-term storage, for example, overnight storage, that can be advantageously provided when the utility robot is securely located within an enclosure in proximity to a charging station. The storage compartments can actively or passively self-identify, and can include tamper and content status information. The storage compartments can automatically interface with the system controller to provide information such as, for example, but not limited to, the tamper information and the content status information. In some configurations, the storage compartments can include information that can be used when by the controller to command the utility robot. In some configurations, when contents within the storage compartments are destination-tagged, the storage compartment can sense the place where the contents are to be delivered and can direct the controller to drive the utility robot to the destination. In some configurations, the storage compartment can transmit destination information to other members of the delivery fleet. In some configurations, contents within the storage compartment can protrude from the storage compartment. Sensors can detect the orientation of the storage compartment and can maintain the storage compartment at a pre-selected angle with respect to the ground.

In some configurations, storage compartments can include temperature/humidity control that can accommodate extended storage, for example, but not limited to, overnight storage, of goods for delivery. In some configurations, storage of food and pharmaceuticals, for example, can be accommodated by temperature and or humidity control within the storage compartments of the present teachings. In some configurations, the storage compartments can include insulation and cold packs of ice, dry ice or other commercially available cold packs such as model S-12762 available from ULINE® in Pleasant Prairie, Wis. In some configurations, storage compartments can include electrically powered refrigerators and/or heaters. In some configurations, the electrically powered heater or cooler may be powered by mains AC. In some configurations, the power can be provided by the batteries of utility robot.

The storage compartments can include sensors mounted exteriorly and interiorly. The storage compartment sensors can detect when they have been touched and moved, and can provide that information to a controller executing in the utility robot. In some configurations, storage compartment sensors can monitor environmental factors, such as, for example, but not limited to, temperature and humidity as well as shock and vibration loads. In some configurations, storage compartment sensors can detect the size and weight of a package and can read information embedded in or on the package. The information can, for example, be embedded in an RFID tag or encoded into a barcode or QR code. The utility robot can compare the information embedded in or on the package to a manifest associated with the delivery, and can raise an alert and/or alarm if the information does not match the manifest.

In some configurations, one or more of the storage compartments can ride above the operator of the utility robot of the present teachings. In some configurations, the above-operator storage compartment(s) can ride on a telescoping device, and can be raised up and down to enable convenient access to the contents of the storage compartment(s), while at the same time enabling convenient entry and exit of the operator onto the utility robot of the present teachings. The telescoping device can include articulation. The storage compartments can ride on positioning rails, and can be positioned backwards, forwards, up, down, and from side to side, for example. The storage compartments can be maintained in a particular orientation automatically by the controller.

In some configurations, the storage containers can be positioned in various orientations and at various locations with respect to each other and the chassis of the utility robot. The storage compartment can accommodate weather barriers to protect the operator of the utility robot from inclement weather. In some configurations, curtains attached to an elevated storage container can protect an operator and possibly storage containers from inclement weather. Parts of the storage container can be articulated to accommodate storing and removing items, and to accommodate secure placement of the storage container. In some configurations, the utility robot can include active control of the storage container, for example, to maintain a particular orientation of the storage container. If the contents of the storage container must remain in a particular orientation to prevent destruction of the contents, active control of the orientation of the contents within the storage container can be enabled. In some configurations, each face of the contents of the storage container can be identified to enable proper orientation of the contents.

In some configurations, sensors can be mounted in various locations on/in the storage container, for example, to notify the utility robot when the storage container could be subject to an undesired collision. In some configurations, the storage container and/or the manifest can inform the utility robot to adjust accelerations according to a pre-selected threshold. The utility robot, which can determine the current rate of acceleration of the utility robot based on data collected from the utility robot's wheel counter and IMU, can limit commands to the drive wheels and/or brakes to adjust accelerations according to the pre-selected threshold.

In some configurations, one of the storage containers can be mounted behind the operator, and can be greater than or equal to about two feet tall. The storage containers can include snap-on features that can allow placement of the storage containers onto the chassis in various configurations. The storage containers can receive and process information from an electronic application, for example, open and close commands from a wireless device. In some configurations, when a parcel is loaded into a storage container, the utility robot can identify, for example by taking a photograph, the individual who loads the parcel and associate the parcel with the identification. In some configurations, the storage container of the present teachings can measure 30-40 inches by two feet. In some configurations, the utility robot can automatically poll the parcels it carries and automatically summon any needed assistance to deliver the parcels in a timely manner. The mounted storage containers can be interchangeable with storage containers of sizes suitable for the particular delivery and can be secured to the utility robot.

The utility robot of the present teachings can be docked proximal to where package delivery can originate. In some configurations, docking stations can include openings in the building where the packages are located. Packages can be deposited at stations within the buildings and near the openings, and can be automatically sorted. The sorted packages can be automatically loaded onto a utility robot of the present teachings through one of the openings. Sensors and/or transponders can detect the contents of the packages.

The utility robots of the present teachings can include technology to collect payment for services and retain payment records. The utility robot can notify the service target that the service has been completed, for example, by a cell phone notification or a text. The service target can move towards the utility robot to avoid challenging terrain such as, for example, stairs. In some configurations in which the service provided is a delivery service, storage compartments can include embedded RFID circuitry that can be broken when the delivery storage is opened. An RFID scanner could be used to reveal that the storage container has been opened. To maintain privacy, the contents of the storage container can be moved to a secure location before opening. The utility robot can receive information about the service target such as, for example, biometric information, to identify that the service is being delivered to the correct target. For example, the utility robot can secure the storage container until the target is recognized by, for example, facial recognition technology. The utility robot can receive personal information such as credit card and cell phone information, to, for example, unlock a storage container. In some configurations, the utility robot can include biometric sensors, for example, facial sensors and/or fingerprint sensors, that can, for example, detect if the contents of a storage container are associated with the person attempting to collect the contents. In some configurations, the utility robot can combine correct location information with correct code entry or other forms of identification to unlock the storage container.

The utility robots of the present teachings can detect tampering with the utility robot, and thus unsafe and dangerous conditions. In some configurations, the utility robot can detect a change in the center of mass that can indicate tampering. Adding or subtracting weight from the utility robot can change the center of mass. The utility robot can include an IMU, and can measure the location of center of mass based on the response of the vehicle to accelerations and changes in the attitude of the utility robot. The change of mass can indicate that the utility robot might be compromised. In some configurations in which packages are being transported, the utility robot can detect packages that do not include identification sufficient to couple the package with the delivery target. For example, the utility robot can detect an unapproved package because a loading authorization code does not match the expected code, or the RFID code is incorrect or missing, or there is a mismatch between the actual weight of the package and the weight listed on the manifest. The utility robot can generate an alert, the type of which can depend upon the probable cause of the suspected tampering. Some alerts can be directed to the state authorities, while others can be directed to an electronic record that can be accessed by the utility robot of the present teachings, the trucks, the smart beacons, and other possible participants in the provided service, possibly through the fleet network. Following an error condition, the utility robot can automatically or semi-automatically steer the utility robot to a safe location such as a charging station. In some configurations, the contents of storage containers can be secured.

Beacons can communicate with the utility robot, and the status of the utility robot and its current activities can be provided to the beacons and thus to the fleet network. In some configurations where the utility robot is delivering goods, beacons can communicate with the contents of the storage containers, and a list and status of the contents of the storage containers can be made available to other members of the delivery fleet through the fleet network. All of the members of the fleet can be recognized by each other. If a utility robot of the present teachings detects that it has been compromised, it can initiate a safety procedure in which its secure electronic information can be backed up and destroyed, and the contents of its storage containers can be safely locked down.

To facilitate mapping of the route traveled by the utility robot between the starting and ending points, whether the starting point is at a fixed location, such as a pickup station associated with a brick-and-mortar source, or whether the starting point is at a mobile location, such as a truck or a pedestrian, the utility robot can begin with a static map. In some configurations, the static map can be derived from an open source map. In some configurations, the fleet system can include at least one server that can manage static map activity. In some configurations, the utility robot can maintain a local version of the static map from which it can operate between updates from the version maintained by the server. In some configurations, the utility robot can augment the static map with, for example, but not limited to, indications of congested areas based on information from, for example, but not limited to, other fleet vehicles, cell phone applications, obstacles such as trees and trash cans, pedestrians, heat map data, and wifi signals. The static map can be used, in conjunction with utility robot sensor data and fleet data, to deduce the location of dynamic objects. The utility robot can collect navigation data while enroute to a target and can avoid the congested areas. The utility robot can, for example, detect fiducials and beacons installed at various places along the path, for example, but not limited to, street corners and street signs at street corners. The fiducials and beacons can be members of the fleet network and thus share data with and possibly receive information from members of the fleet network. The fiducials and beacons can be installed and maintained by any entity including, but not limited to, the item's source entity, the company managing the deliveries, and the city in which the deliveries are taking place. The utility robots can receive information from fiducials and beacons installed at street intersections and, in some configurations, can send information to the fiducials and beacons that are configured to receive information. The utility robot can also sense safety features such as traffic lights and walk/no-walk indicators that can generate alerts audibly, visually, another type/frequency of signal, and/or all of the alert generation methods. The utility robot can process traffic light data and follow the pre-established road rules that it has learned. For example, the utility robot can be taught to stop when the traffic light is red. Vehicles in an intersection can be detected. Route issues such as closures can be detected. The utility robot can update the fleet network's database with information such as, but not limited to, traffic light information, that can enrich the mapping utility robot available to the fleet network. In some configurations, the utility robot can make use of information collected by a body camera worn by the operator of a member of the fleet network.

Semi-autonomous utility robots of the present teachings can receive input from operators during each trip and can use that input to record locations of obstacles such as, for example, but not limited to, stairs, cross-walks, doors, ramps, escalators, and elevators. From these data and real-time and/or semi-real-time data, maps and dynamic navigation routes can be created and updated. Autonomous utility robots can use the maps for current and future deliveries. For each step in the dynamic navigation route, the utility robot of the present teachings can determine the obstacles in the navigation route, the amount of time required to complete a desired motion that the utility robot will have to accomplish to follow the navigation path, the space that will be occupied by the static and dynamic obstacles in the path at that time, and the space required to complete the desired motion. With respect to the obstacles, the utility robot can determine if there is an obstacle in the path, how big the obstacle is, whether or not the obstacle is moving, and how fast and in what direction the obstacle is moving and accelerating. The dynamic navigation path can be updated during navigation. The path with the fewest obstacles can be chosen, and dynamic route modifications can be made if a selected route becomes less optimal while the utility robot is in transit. For example, if a group of pedestrians moves to a position in the chosen route, the route can be modified to avoid the group of pedestrians. Likewise, if repairs begin on a sidewalk, for example, the route can be modified to avoid the construction zone. Stereo cameras and point cloud data can be used to locate and avoid obstacles. The distance from various obstacles can be determined by real-time sensing technology such as, for example, but not limited to planar LIDAR, ultrasonic sensor arrays, RADAR stereoscopic imaging, monocular imaging, and velodyne LIDAR. In some configurations, processing of sensor data by the utility robot can allow the utility robot to determine, for example, whether the utility robot is within an allowed envelope of the planned path, and whether the obstacles in the navigation path are behaving as predicted in the dynamic navigation path. The utility robot can accommodate trips of various lengths, solving the problem of short-distance delivery of services.

Information can be derived from commercially-available navigation tools that provide online mapping for pedestrians, for example. Commercially-available navigation tools such as, for example, but not limited to, GOOGLE® maps, BING® maps, and MAQUEST® maps, can provide pedestrian map data that can be combined with obstacle data to generate a clear path from source to destination as the utility robot travels from one place to another. Crowd-sourced data can augment both navigational and obstacle data. Operators who travel in the vicinity of the source of the goods and the target services area can be invited to wear cameras and upload data to the utility robot, and/or to upload an application that can, for example, but not limited to, track location, speed of movement, congestion, and/or user comments. Operators can perform the job of smart sensors, providing, for example, but not limited to, situational awareness and preferred speed to the utility robot. In some configurations, operator driven systems of the present teachings can generate training data for interactions with people including, but not limited to, acceptable approach distances, following distances, and passing distances. Cellular phone-type data, such as, for example, but not limited to, obstacles and their speed and local conditions, can be made available to the fleet's database to enable detailed and accurate navigation maps. The utility robot can include technology that can determine areas in which the GPS signal falls below a desired threshold so that other technologies can be used to maintain communications. Sidewalks can be painted with various substances, such as, for example, but not limited to, photo luminescent substances, that can be detected by sensors on the utility robot. The utility robot can use the data gathered from sensing the substances to create and augment navigation maps.

Wheel rotation and inertial measurement data can be combined to determine dead reckoning positions when creating the maps. Sensor data, such as data from visual sensors, can be used to determine dead reckoning positions. The utility robots of the present teachings can receive information about their routes from information collected by trucks, and members of the fleet can be used to create/improve pedestrian maps. The trucks can include portable utility robots, and the operators of the trucks can collect further data by use of body cameras and location sensors to map walking deliveries. Visual, audible, and thermal sensing mechanisms can be used on the trucks and in conjunction with the operator's movements. The utility robot can make use of optimized and/or preferred route information collected by trucks and operators. The utility robot can include a pedestrian route on the desired navigation map.

In some configurations, the utility robot can learn navigation paths independently and can share the navigation information with other members of the fleet network. In some configurations, the operator can select at least one optimum navigation route. The utility robot can also include cameras that can be used to augment navigation maps. Areas that can be located inside buildings such as, for example, but not limited to, doors, stairs, and elevators, and routes limited to pedestrians, can be candidates for body camera data collection. In subsequent journeys to the same location, the doors, stairs, and elevators may be navigable by the utility robot, and the utility robot can by-pass pedestrian-only paths, for example. The utility robot can follow a planned route. The utility robot can receive commands from the operator, and/or can self-command based on the desired route. Steering and location assistance can be provided by navigation tools combined with obstacle avoidance tools. The utility robot can accommodate ADA access rules, including, but not limited to, space requirements with respect to the utility robot's egress and ingress requirements.

In some configurations, the dynamic navigation path can be updated by the utility robot when the utility robot determines if an obstacle can be surmounted and/or avoided. For example, the utility robot can determine if the obstacle can be driven over, such as a curb, a rock, or a pothole, or can be driven around. The utility robot can determine if the obstacle can be expected to move out of the navigation path, and if there is a way that the utility robot can make progress along the planned navigation path. In some configurations, the utility robot of the present teachings can accommodate crossing roads with and without traffic signals, curbs, dynamic obstacles, and complete path obstruction. The utility robot can include routing technology that can avoid congested areas based on, for example, but not limited to, current congestion information from other utility robots of the present teachings, crowd-sourced congestion information, and historical congestion information from other utility robots of the present teachings and trucks. Historical congestion information can include, but is not limited to including, day and time of congestions from past traverses in the same area by utility robots of the present teachings, and data and time of congestion from delivery truck speed. Dynamic navigation paths can be created based on current path data and the maps. The utility robot can include training technology in which data from operators traveling a route can inform the utility robot of the present teachings how to interact with moving obstacles and how to behave in an environment having moving obstacles. In some configurations, data from fleet drivers traveling the route can be used as training data for machine learning on how to interact with moving people or in an environment of moving people. In some configurations, a heat map of pedestrian traffic can be used to update pedestrian density data. In some configurations, route planning can take into account the desired transit time, the estimated transit time, how much space obstacles are occupying on the planned route, and how much space the utility robot requires. The utility robot can determine its status with respect to the planned route, and can track what movement the obstacles in the planned route are making.

Each form of sensor data can provide a unique view of its surroundings, and fusing the various types of sensor data can help to specifically identify obstacles, including dynamic objects. Using these data, dynamic objects can be classified by methods including, but not limited to, semantic segmentation. Predicting the future position of a dynamic object, after it has been identified, can be accomplished by semantic scene segmentation which can color code a scene based on object type. The future position of a dynamic object can also be predicted by creating behavioral models of dynamic objects that can be processed by the utility robots of the present teachings. Neural networks, Kalman filters, and other machine learning techniques can also be used to train the utility robot of the present teachings to understand and react to its surroundings. If the utility robot encounters an obstacle with which it can interact, for example, a pedestrian, the utility robot can be trained to stop before encountering the pedestrian, greet the pedestrian, and avoid hitting the pedestrian, for example. In some configurations, planar LIDAR, visual sensors, and ultrasonic sensors can be used to detect pedestrians. A critical distance around a pedestrian can be defined based on the distance needed to stop based on sensor delays, and social norms, for example. The socially-acceptable interactions between the utility robot and humans may be defined by data from user-driven systems interacting with humans. In some configurations, the data collected by the user-driven systems can be used to train a neural network in the autonomous systems that can control the utility robot's interaction with humans. In some configurations, to avoid obstacles such as humans and vehicles when crossing a street, RADAR and/or LIDAR, combined with stereo cameras, can be used for long distance viewing and to reliably identify the obstacles and create a crossing strategy. In some configurations, the utility robot of the present teachings can communicate wirelessly with available electronic sources such as elevators and pedestrian crosswalks. Smart beacons can be used for this purpose. When obstacles such as construction zones are encountered, the utility robot of the present teachings can purposefully navigate the construction zone, and can inform other fleet members of the extent of the obstacle, giving the other fleet members an opportunity to avoid the obstacle. A neural network executing in the utility robot can train the utility robot to recognize crossing signals, for example, and to cross when safe.

The utility robot can receive information from smart beacons placed strategically along travel paths. In some configurations, information from the smart beacons can be encrypted, and/or information exchanged between the utility robot of the present teaching and the smart beacon can be encrypted to protect the utility robot from malicious hacking. In some configurations, the smart beacons can include cameras, RADAR, and/or LIDAR that can be used to map the local area. In some configurations, smart beacons can vary in complexity and specialization. For example, smart beacons that can manage network communications can be placed in areas where it is likely that network members will need communication services. Smart beacons that include mapping cameras can be placed in locations where mapping is required, and can be moved from place to place depending on current needs. In some configurations, smart beacons can include data transfer hot spot capability, or other networking capability to enable the fleet network of the present teachings to communicate among fleet members. In some configurations, smart beacons can recognize the travel path and be aware of the next navigation step required for the utility robot to reach its desired destination. Smart beacons can receive at least part of the utility robot's path and/or destination from a server. The smart beacons can identify the utility robots of the present teachings, possibly through the secure wireless exchange of identifying information, possibly through visual and/or audible identification techniques, or other means. Secure exchange of messages can include encryption, for example, and other forms of protection against in-flight message modification, man-in-the-middle threats such as eavesdropping and denial of service, third party application threats, and malicious/erroneous application threats. The utility robot can receive navigation information from the smart beacon, including homing, triangulation, and aiming signals. The utility robot can receive current mapping information including, but not limited to, congestion areas and path closures, from the smart beacon, and the utility robot can send the information it has collected to the smart beacon. The utility robot can make beacon information available to other utility robot fleet members at any time, for example, but not limited to, during a parcel delivery and/or pickup. The utility robot can receive information from the smart beacon that can be used to correct the utility robot's IMU dead reckoning and wheel rotation navigation. In some configurations, the utility robot can navigate entirely through information received from the smart beacon. For example, in a congested area, it is possible that some of the sensors located on the utility robot of the present teachings could be blocked. Sensors, for example, LIDAR sensors, on the smart beacon can provide navigation information to the utility robot of the present teachings that the utility robot could not itself have obtained with its on-board sensors. Sensors located on any of the utility robots of the present teachings, the trucks, and/or the smart beacons can provide current congestion information from cameras and/or thermal imaging to form heat maps. The utility robot can receive instructions from a steerable RF or laser beacon that can be controlled by another member of the fleet, a central control location, or by the utility robot itself. In some configurations, the utility robot can be configured with a minimum number of sensors if data are planned to be collected by other fleet members. The utility robot can receive these sensor data, for example, the heat maps, and recognize the location of groups of obstacles, possibly dynamic obstacles, within potential travel routes. In areas without various types of beacons, exploring utility robots with partial or full complements of sensors, can retrieve navigation and congestion data and make the data accessible to utility robots of the present teachings that are traveling the explored routes to deliver goods and services. The exploring systems can provide their sensor data and analyses to a central service, a cloud-based storage area, a smart beacon, and/or another exploring system, utility robot, and/or truck or other member of the delivery fleet, for example. Beacons can be used to facilitate data communications among the fleet members, and can be used to improve localization accuracy. In some configurations, beacons can include wireless access points generating signals, such as, for example, wifi and RF signals, that can be used to help navigate the utility robot in areas in which global positioning techniques are inadequate.

BRIEF DESCRIPTION OF THE DRAWINGS

The present teachings will be more readily understood by reference to the following description, taken with the accompanying drawings, in which:

FIG. 1 is a pictorial representation of the fleet network of the present teachings;

FIG. 2 is a schematic block diagram of the system of the present teachings;

FIG. 3 is a flowchart of the method of robot path processing of the present teachings;

FIG. 4 is a pictorial representation of the truck and autonomous vehicles of the present teachings;

FIG. 5 is a schematic block diagram of a second configuration of the system of the present teachings;

FIG. 6 is a schematic block diagram of the sensor system of the present teachings;

FIG. 7 is a schematic block diagram of the fleet network communications of the present teachings;

FIG. 8 is a schematic block diagram of a third configuration of the system of the present teachings;

FIG. 9 is a schematic block diagram of a configuration of a vehicle system including a localization subsystem;

FIG. 10 is a schematic block diagram of a configuration of a vehicle system including an obstacle subsystem;

FIG. 11 is a schematic block diagram of a configuration of a vehicle system including training and rules compliance subsystems;

FIG. 12 is a schematic block diagram of a configuration of a vehicle system including a preferred route subsystem;

FIG. 13 is a schematic block diagram of a configuration of a vehicle system including a road obstacle-climbing subsystem;

FIG. 14 is a schematic block diagram of a configuration of a vehicle system including a stair-climbing subsystem;

FIGS. 15A-15K are pictorial representations of a stair-climbing autonomous vehicle of the present teachings;

FIG. 16 is a schematic block diagram of a configuration of a vehicle system including a grouping subsystem;

FIG. 17 is a schematic block diagram of a fourth configuration of the system of the present teachings;

FIG. 18A is a schematic block diagram of the infrastructure of the system of the present teachings;

FIG. 18B is a schematic block diagram of robot path processing of the system of the present teachings;

FIG. 19 is a pictorial representation of perception processing of the present teachings;

FIGS. 20-22 are pictorial representations of object detection and classification of the present teachings;

FIG. 23 is a pictorial representation of object parameter estimation of the present teachings;

FIG. 24 is a pictorial representation of path planning processing of the present teachings;

FIG. 25 is a pictorial representation of path following processing of the present teachings;

FIG. 26 is a schematic block diagram of the robot path processing with map update;

FIG. 27 is a flowchart of the method for managing control of the vehicle of the present teachings;

FIG. 28 is a schematic block diagram of multi-robot path planning of the present teachings;

FIG. 29 is a pictorial representation of a subset of the steps involved in robot path processing;

FIG. 30 is a pictorial representation of static route map construction of the present teachings;

FIG. 31 is a flowchart of the method of map management of the present teachings; and

FIGS. 32A-32B are schematic block diagrams of fleet management components of the present teachings.

DETAILED DESCRIPTION

The utility system of the present teachings is discussed in detail herein in relation to commercial services. However, various types of applications may take advantage of the features of the present teachings.

Referring now to FIGS. 1 and 2, system 100 for moving a utility robot from at least one starting point to at least one utility execution point 128 can include, but is not limited to including, system collectors 119 that can form a communications network. System collectors 119 (FIG. 2) can access historic data 137 (FIG. 2) associated with a proposed path between at least one starting point and at least one ending point 128. System collectors 119 can include utility vehicles 113 (FIG. 2). At least one utility vehicle 113 (FIG. 2) can include, but is not limited to including, autonomous utility vehicle 119A (FIG. 1) and semi-autonomous utility vehicle 119B (FIG. 1). In some configurations, at least one utility vehicle 113 (FIG. 2) can include at least one sensor 118 and at least one storage container 101. In some configurations, at least one storage container 101 can house the goods that are to be delivered. Historic data 137 (FIG. 2) can include vehicle data 129 (FIG. 2) previously collected along the proposed path, which can be delivered to drive subsystem 111. Drive subsystem 111 can provide drive commands to utility vehicle 113 processors. System collectors 119 (FIG. 2) can collect real time data 127 (FIG. 2) about the proposed path before and while at least one utility vehicle 113 (FIG. 2) navigates the proposed path. System collectors 119 (FIG. 2) can update the proposed path based at least on vehicle data 129 (FIG. 2), historic data 137 (FIG. 2), and real time data 127 (FIG. 2). System 100 can include at least one processor that can execute in utility vehicle 113 (FIG. 2), and/or in a server such as, for example, fleet manager 601 (FIG. 1) communicating with system collectors 119 (FIG. 2), including utility vehicles 113 (FIG. 2), through communications network 115 (FIG. 2). The processors can continually update—based at least on historic data 137 (FIG. 2), real time data 127 (FIG. 2), and at least one sensor 118—the updated proposed path while utility vehicles 113 (FIG. 2) navigate the updated proposed path from at least one starting point to at least one utility execution point 128. In some configurations, system collectors 119 (FIG. 2) can optionally include airborne vehicles 2000 (FIG. 1) that can transport the goods to, for example, trucks 2001 (FIG. 1). In some configurations, self-driving cars 2001A can be included in the fleet network.

Referring now to FIG. 2, a group of utility vehicles 113 can travel together for several reasons. In some configurations, one member of the group can be “learning” a delivery path and can be “teaching” other members the path. In some configurations, multiple utility vehicles 113 can be required to deliver goods and/or perform services that are too numerous for a single utility vehicle 113 to accomplish. In some configurations, a method for delivering goods from at least one first location to at least one second location can include, but is not limited to including, coupling, by at least one of a plurality of utility vehicles, at least one of the plurality of utility vehicles with another of the plurality of utility vehicles through a communications network. The method can include receiving, by at least one of a plurality of utility vehicles 113, the goods from the at least one first location into at least one of the plurality of utility vehicles 113. The method can include determining, by at least one of a plurality of utility vehicles 113, a proposed path between the at least one first location and the at least one second location, and enabling, by at least one of a plurality of utility vehicles, the at least one of the plurality of utility vehicles 113 to follow the other of the plurality of utility vehicles 113 along the proposed path, and enabling, by at least one of the plurality of utility vehicles 113, the other of the plurality of utility vehicles 113 to deliver the goods at the second location. The method can optionally include (a) updating, by at least one of a plurality of utility vehicles 113 (FIG. 2), the proposed path based at least on information received in real time from the one at least one utility vehicle 113 and the other at least one utility vehicle 113, (b) enabling, by at least one of a plurality of utility vehicles 113, the one at least one utility vehicle 113 to proceed along the updated proposed path, and (c) repeating (a) and (b) until the one at least one utility vehicle 113 reaches the at least one second location. The coupling can optionally include a physical and/or an electronic coupling.

Continuing to refer to FIG. 2, a group of utility vehicles 113 can include at least one semi-autonomous utility vehicle 119B (FIG. 1) and/or at least one autonomous utility vehicle 119A (FIG. 1). At least one of utility vehicles 113 can optionally follow a different path from the rest of the group. At least one diverted utility vehicle 113 can provide services at a different location from the rest of the group, for example, or may have experienced mechanical or electronic problems and can seek help, or may have been summoned by a customer needing help with a package or a safe escort. Any members of the group can optionally update the fleet network with path and status information, for example, through communication network 115 (FIG. 1). In some configurations, when a customer at the first location needs assistance, the customer can summon a nearby one of utility vehicles 113 through, for example, fleet manager 621 (FIG. 1), or through, for example, direct communications with utility vehicle 113. Utility vehicle 113 can optionally be directed to a mobile destination or a fixed destination, or a destination that had been fixed but became mobile, for example, a parked vehicle that starts and moves or a pedestrian who is walking. In some configurations, one member of the group can be “learning” a travel path and “teach” other members the path. In some configurations, semi-autonomous utility vehicle 119B (FIG. 1) can create an electronic record of a traversed path based on sensor data 118. Autonomous vehicle 119A (FIG. 1) can follow the traversed path by steering according to the electronic record. In some configurations, utility vehicle 113 can transport goods. In some configurations, system 100 can include optional physical storage 101, and optional physical storage subsystem 103 that can provide optional physical storage control commands 131 to optional physical storage 101. Optional physical storage 101 can include at least one processor, for example, that can receive commands and respond to the commands. Optional physical storage subsystem can provide receive and send optional physical storage status 133 from/to delivery path subsystem 117 which can be tracking the status of the goods contained in optional physical storage 101.

Referring now to FIG. 3, method 150 of the present teachings for establishing a path for moving utility vehicle 113 (FIG. 2) from at least one starting point to at least one destination 128 (FIG. 2) can include, but is not limited to including (a) automatically determining 151, by fleet network 606 (FIG. 1) including system collectors 119 (FIG. 2), at least one proposed path between the at least one starting point and at least one destination 128 (FIG. 2). The proposed path can be selected from a set of pre-selected types of routes. In some configurations, the proposed path can include pedestrian route 602 (FIG. 1) including street crossings 604 (FIG. 1). System collectors 119 (FIG. 2) can include utility vehicles 113 (FIG. 2). Method 150 can include (b) accessing 153, by utility vehicle 113 (FIG. 2), historic data 137 (FIG. 2) associated with the proposed path. At least some of historic data 137 (FIG. 2) can be collected by at least one of system collectors 119 (FIG. 2). Method 150 can include (c) collecting 155, by at least one of system collectors 119 (FIG. 2), real time data 127 (FIG. 2) about the proposed path, and (d) updating 157, by fleet network 606 (FIG. 1), the proposed path based on historic data 137 (FIG. 2) from historical data subsystem 109 (FIG. 2) and collected real time data 127 (FIG. 2) from real time data subsystem 125 (FIG. 2). Method 150 can include (e) navigating 159, by utility vehicle 113 (FIG. 2), the updated proposed path, and (f) repeating 161 (c)-(e) until utility vehicle 113 (FIG. 2) reaches the at least one destination 128 (FIG. 2). Method 150 can optionally include authenticating and annotating the updated proposed path, by utility vehicle 113 (FIG. 2), as utility vehicle 113 (FIG. 2) navigates the updated proposed path, and providing, by utility vehicle 113 (FIG. 2), the authenticated, annotated, updated proposed path to fleet network 606 (FIG. 1). Method 150 can optionally include forming communications network 115 (FIG. 2) including system collectors 119 (FIG. 2), and sharing, by system collectors 119 (FIG. 2), historic data 137 (FIG. 2) and real time data 127 (FIG. 2) through communications network 115 (FIG. 2). Authenticating and annotating can include receiving, by utility vehicle 113 (FIG. 2), visually-collected information from a driver of utility vehicle 113 (FIG. 2). Historic data 137 (FIG. 2) can include, but is not limited to including, data from a plurality of sources. Fleet network 606 (FIG. 1) can include, but is not limited to including at least one server. Method 150 can include maintaining, by the at least one server, historic data 137 (FIG. 2) and the updated proposed path.

Referring now to FIG. 4, system collectors 119 (FIG. 2) can include trucks 2001 that can, for example transport goods to utility vehicles 113, and can transport utility vehicles 113 to the vicinity of delivery locations 128 (FIG. 1). Trucks 2001 can enable exchanging of spent batteries 1163 (FIG. 5) with charged batteries 1163 (FIG. 5) in utility vehicles 113. Trucks 2001 can include battery charging features that can charge spent batteries 1163 (FIG. 5). Trucks 2001 can include lift mechanisms that can enable ingress and egress of utility vehicles 113. Trucks 2001 can optionally include in-lift features 2003 and out-lift features 2005/2007 such as, for example, but not limited to, ramps, that can enable ingress and egress of utility vehicles 113 to/from trucks 2001. In some configurations, trucks 2001 can be moving while utility vehicles 113 enter and leave trucks 2001. In some configurations, utility vehicles 113 can receive packages from trucks 2001, and can drop packages such as, but not limited to, undeliverable packages, into trucks 2001.

Referring now to FIGS. 5 and 6, in some configurations, utility execution system 200 (FIG. 5) for moving utility vehicles from at least one first location to at least one second location can include, but is not limited to including, a network of system collectors 119 (FIG. 5) including at least one utility vehicle 113 (FIG. 5). Utility execution system 200 (FIG. 5) can include at least one processor A 114A. Utility vehicle 113 (FIG. 5) can optionally include sensor subsystem 105 (FIG. 5) that can process data from sensors 118 (FIG. 5). Sensors 118 (FIG. 5) can include, but are not limited to including, infrared (IR) sensors 201 (FIG. 6) that can sense, for example, pedestrians, cameras 203 (FIG. 6) that can sense object depth, and lasers 205 (FIG. 6) that can provide a point cloud representation of an object and distance measurements. Sensors 118 (FIG. 5) can include ultrasonic sensors 207 (FIG. 6) that can sense the distance to an object, radar 209 (FIG. 6) that can sense the speed of an object, as well as weather and traffic proximate to utility vehicle 113 (FIG. 5), and LIDAR 211 (FIG. 6) that can, for example, but not limited to, provide point cloud data. Sensor subsystem 105 (FIG. 5) can optionally include sensor fusion subsystem 108 (FIG. 6) that can integrate data from a plurality of sensors 118 (FIG. 5). Sensor fusion subsystem 108 (FIG. 6) can classify obstacles encountered by utility vehicle 113 (FIG. 6), and can validate the observations from unreliable sensors. Sensor subsystem 105 (FIG. 5) can optionally include behavior model subsystem 106 (FIG. 6) that can predict future positions of the obstacles. Sensor subsystem 105 can optionally expect sensor data 135 to arrive from at least two of sensors 118 (FIG. 5). Utility vehicle 113 (FIG. 5) can optionally include at least one battery 1163 (FIG. 5). Battery 1163 (FIG. 5) can optionally include a quick charge feature and a quick change feature, both of which can reduce non-operational time of utility vehicle 113 (FIG. 5). Battery 1163 (FIG. 5) can optionally include a locking feature that can lock battery 1163 (FIG. 5) to utility vehicle 113 (FIG. 5). The locking feature can include a security feature that can enable removal of battery 1163 (FIG. 5).

Referring now to FIGS. 1 and 7, utility vehicles 113 (FIG. 2) can optionally include at least one autonomous vehicle 119A and/or at least one semi-autonomous vehicle 119B. Autonomous vehicles 119A of the present teachings can include vehicles that can navigate with little to no human intervention. Semi-autonomous vehicles 119B of the present teachings can collect information from an operator while traversing terrain either autonomously or under human control or under shared control between the human and an autonomous processor. Autonomous vehicles 119A and semi-autonomous vehicles 119B can operate on, for example, but not limited to, sidewalks 602 (FIG. 1) and other pedestrian pathways that can include, for example, but not limited to, cross-walks 604 (FIG. 1), curbs 612 (FIG. 1), stairs 614 (FIG. 1), and elevators. System collectors 119 (FIG. 2) can optionally include at least one beacon 119C positioned along the updated proposed path. System collectors 119 (FIG. 2) can optionally include beacons 119C positioned along the updated proposed path. Beacons 119C can sense, for example, but not limited to, obstacles, weather, and fiducials and can provide those data to other system collectors 119 (FIG. 2), one or more of which can include utility vehicles 113 (FIG. 2). Beacons 119C can enable communication among system collectors 119 (FIG. 2), and can enable data protection during the exchange of data between beacons 119C and other system collectors 119 (FIG. 2). Beacons 119C, along with all other system collectors 119 (FIG. 2), can receive and transmit data over communications network 115 (FIG. 2), and can provide those data to utility vehicles 113 (FIG. 2), among other recipients. Members of communications network 115 (FIG. 2) can optionally receive GPS navigation information 145 (FIG. 7) and information from wireless devices using, for example, but not limited to, wireless access points (WAP) 147 (FIG. 7). At least one WAP 147 (FIG. 7) can optionally enable fleet communications when communications network 115 (FIG. 2) is inadequate, and location information when GPS 145 (FIG. 7) is inadequate.

Referring now to FIG. 8, utility vehicle 113 can optionally include seat feature 157 that can accommodate an operator. The operator can control utility vehicle 113, or can partially control utility vehicle 113. In some configurations, semi-autonomous utility vehicle 119B (FIG. 1) can include seat feature 157. In some configurations, semi-autonomous utility vehicle 119B (FIG. 1) can include a wheelchair. In some configurations, semi-autonomous utility vehicle 119B (FIG. 1) can be remotely controlled, with no seating feature 157 and no operator.

Referring now to FIG. 9, utility vehicle 113 (FIG. 2) can optionally include at least one localization subsystem 141 that can localize utility vehicle 113 (FIG. 2) based at least on historic data 137, and/or real time data 127, and/or local data 143, where localization can include, but is not limited to, determining the current location and orientation of utility vehicle 113 (FIG. 2).

Referring now to FIGS. 10 and 11, utility vehicle 113 (FIG. 2) can optionally include obstacle subsystem 146 that can locate at least one obstacle in the update proposed path. Obstacle subsystem 146 can update the updated proposed path when obstacle data 144 are discovered. Obstacle subsystem 146 can rely upon training subsystem 1159 (FIG. 11) to provide obstacle recognition means. Training subsystem 1159 (FIG. 11) can provide continuous learning of situations encountered by members of the fleet, and can provide those data to obstacle subsystem 146 to improve route planning and execution. Obstacle subsystem 146 can be pre-trained. Training subsystem 1159 (FIG. 11) can include and/or can be based on neural network technology, for example. Training subsystem 1159 (FIG. 11) can operate remotely from processor A 114A. Utility vehicle 113 (FIG. 2) can optionally include rules compliance subsystem 1157 (FIG. 11) that can access navigation rule information from at least one of historic data 137, real time data 127, and sensor data 135. Rules compliance subsystem 1157 (FIG. 11) can command utility vehicle 113 (FIG. 2) to navigate at least according to the navigation rule information.

Referring now to FIG. 12, utility vehicle 113 (FIG. 2) can optionally include preferred route subsystem 147 that can determine at least one preferred route 149 between at least one starting point and at least one destination 128 (FIG. 1). Utility vehicle 113 (FIG. 2) can select at least one preferred route 149 based at least on historic data 137 and real time data 127. Preferred route subsystem 147 can optionally determine at least one path between at least one starting point and at least one destination 128 (FIG. 1) that utility vehicle 113 (FIG. 2) should avoid based at least on the number of obstacles in the updated proposed path.

Referring now to FIG. 13, utility vehicle 113 (FIG. 2) can optionally include road obstacle-climbing subsystem 1149 that can detect road obstacles. Road obstacle-climbing subsystem 1149 can send road obstacle data 1151 to delivery path subsystem 117, and command utility vehicle 113 (FIG. 2) to crest the road obstacles, and to maintain balance and stability while traversing the road obstacles. Road obstacles can optionally include curbs 612 (FIG. 1) and steps 614 (FIG. 1).

Referring now to FIG. 14, utility vehicle 113 (FIG. 2) can optionally include stair-climbing subsystem 1153 that can detect stairs 614 (FIG. 1), send stair data 1155 to delivery path subsystem 117, and command utility vehicle 113 (FIG. 2) to encounter and traverse stairs 614 (FIG. 1), and command utility vehicle 113 (FIG. 2) to achieve stabilized operation while traversing stairs 614 (FIG. 1).

Referring now to FIGS. 15A-15K, balanced and safe autonomous stair-climbing can be accomplished by vehicle wheels clustered together to provide coordinated ascent and descent, in combination with a supporting arm deployed as the vehicle wheels encounter the stairs. Stair-climbing can begin with autonomous movement of autonomous vehicle 1500A from floor 618A towards stairs 614 (FIG. 15A). As autonomous vehicle 1500A approaches stairs 614, supporting arm 1505 is in storage position with arm wheels 1501A adjacent to vehicle storage 101, and segment 1501 folded towards arm 1504. As autonomous vehicle 1500A encounters riser 618 (FIG. 15B), front wheels 2815 sense contact from sensors (not shown), and sensor data can be provided to a powerbase (not shown). The powerbase can initiate the active rotation at pivot point 1506 of arm 1504 by a servo (not shown) based at least one the sensor data. Such active rotation can enable segment 1501 to move towards the ground surface, for example, but not limited to, under the weight of gravity. Stabilizing wheels 1503, which can optionally be powered, operably coupled with segment 1501, can land on the ground, extending supporting arm 1505 and providing support to autonomous vehicle 1500A. Stabilizing wheels 1503 can optionally be replaced by a skid-like feature. The powerbase can issue commands to a cluster motor (not shown) to rotate a cluster, and thus move rear wheels 2817 onto landing 628 (FIG. 15C). As utility vehicle 1500A climbs stairs 614, arm wheel cluster 1501A rotates at axle 1508 as supporting arm 1505 maintains balance and stability of autonomous vehicle 1500A. As rear wheel 2817 encounters riser 616 (FIG. 15C), the cluster can rotate front wheel 2815 to arrive at landing 632 (FIG. 15D), while supporting arm 1505 rolls towards stairs 614 on wheel cluster 1501A to provide balance and support to autonomous vehicle 1500A. As front wheel 2815 encounters riser 622 (FIG. 15D), the cluster can rotate rear wheel 2817 to arrive at landing 624 (FIG. 15E), while supporting arm 1505 rolls onto landing 628 (FIG. 15E) as wheel cluster 1501A reaches riser 616, providing balance and support to autonomous vehicle 1500A. As rear wheel 2817 reaches landing 624 (FIG. 15F), the cluster can rotate front wheel 2815 to arrive at landing 624 (FIG. 15F), while supporting arm 1505 rolls onto landing 634 (FIG. 15F) as wheel cluster 1501A reaches riser 622, providing balance and support to autonomous vehicle 1500A. With no further risers to meet, the cluster can rotate front wheel 2815 to rest on landing 624 (FIG. 15G), as wheel cluster 1501A reaches riser 626 and landing 624 (FIG. 15G), and the servo rotates pivot point 1506 (FIG. 15H) to raise supporting arm 1505 (FIG. 15G) in preparation for either forward motion or descending stairs 614. To descend stair 614, the cluster can rotate front wheel 2815 above rear wheel 2817 as supporting arm 1505 reaches towards stairs 614 to stabilize the downward trip. Wheels 2815/2817 can travel down stairs 614 (FIG. 15I) alternating as described for the upward climb, while arm wheels 1501A roll down stairs 614 from landing to landing. Eventually supporting wheels 1501A (FIG. 15J) make ground contact before the final rotation of the cluster. Rear wheels 2817 (or front wheels 2815, depending on how many risers there are in stairs 614) are rotated to the ground adjacent to riser 618 (FIG. 15J), balanced by supporting arm 1505. One further rotation by the cluster places all of front wheels 2815, rear wheels 2817, and supporting wheels 1501A (FIG. 15K) on the ground. In some configurations, supporting wheels 1501A can be pressure-activated. In some configurations, pivot point 1506 (FIG. 15A) and optionally wheels 1501A (FIG. 15A) can be actuated by motors in power base 531 (FIG. 14). The motors can be connected to pivot point 1506 (FIG. 15A) and optionally wheels 1501A (FIG. 15A) by wires that can run through structures such as tubes that support 1501A (FIG. 15A). In some configurations, one or more of supporting wheels 1501A can be omitted from supporting arm 1505.

Referring now to FIG. 16, utility vehicle 113 (FIG. 2) can optionally include grouping subsystem 161 that can command one utility vehicle 113 (FIG. 2) to follow another utility vehicle 113 (FIG. 2). Grouping subsystem 161 can maintain a coupling between utility vehicles 113 (FIG. 2). In some configurations, grouping subsystem 161 can enable coupling electronic coupling among utility vehicles 113 (FIG. 2). In some configurations, the coupling can include a physical coupling. In some configurations, grouping subsystem 161 can group several of utility vehicles 113 (FIG. 2) together, and can enable one or more of utility vehicles 113 (FIG. 2) to collect navigational path data and provide the data to the utility network. In some configurations, grouping subsystem 161 can enable groups of utility vehicles (FIG. 2) to travel together until one of more of utility vehicles 113 (FIG. 2) achieves a destination and moves out of the group to perform services.

Referring now to FIG. 17, system 500 for moving utility vehicle 113 from at least one first location to at least one second location, another configuration of system 100 (FIG. 2), can include, but is not limited to including, at least one processor, including, but not limited to, processor 1512 and processor 2513. Processor 1512 is also referred to herein as receiving processor 512. Processor 2513 is also referred to herein as executing processor 513. Receiving processor 512 can receive at least one request from the at least one first location to perform services at the at least one second location. Receiving processor 512 can choose at least one optimum utility vehicle from utility vehicles 113 (FIG. 4), and the choice can be based at least on the status of at least one utility vehicle 113 (FIG. 4). Receiving processor 512 can direct executing processor 513 associated with the at least one optimum utility vehicle to command the optimum utility vehicle to the at least one first location to receive the goods. Executing processor 513 can associate at least one security means with the goods as the goods are stored in the at least one optimum utility vehicle. The at least one security means can require security information before the services are executed. Executing processor 513 can determine a proposed path between the at least one first location and the at least one second location based at least on historic information 137 received from the network of system collectors 119 (FIG. 4) and map database 505. Executing processor 513 can enable the at least one optimum utility vehicle to proceed along the proposed path, and can proceed until the at least one optimum utility vehicle reaches the at least one second location. Executing processor 513 can verify the security information and release the goods at the location of utility vehicle 113 (FIG. 2). Executing processor 513 can optionally (a) update the proposed path based at least on information received in real time from the network of system collectors 119 (FIG. 2), (b) enable the at least one optimum utility vehicle to proceed along the updated proposed path, and can (c) repeat (a) and (b) until the at least one optimum utility vehicle reaches the at least one second location. Truck 2001 (FIG. 4) can optionally transport utility vehicle 113 (FIG. 4) to the at least one first location, then on to the vicinity of the at least one second location. System 500 can include dispatch mechanism 501 that can coordinate activities among members of the network. In some configurations, dispatch mechanism 501 can couple trucks 2001 (FIG. 4) with utility vehicles 113 (FIG. 4). In some configurations, dispatch mechanism 501 can track battery life in utility vehicles 113 (FIG. 4). In some configurations, dispatch mechanism 501 can enable utility vehicle 113 (FIG. 4) to respond to a summons. Dispatch mechanism 501 can enable utility vehicle 113 (FIG. 4) to respond to a summons by receiving the summons from system collectors 119 (FIG. 2) and transmitting the summons to utility vehicle 113 (FIG. 4). Processor 2513 can communicate movement control commands 551 that can include path data 549 to power base 531 through CANbus 527. Powerbase 2531 can communicate user update information 553 through communications interface 551 to processor 2513. In some configurations, packages can be delivered from one location to another using utility vehicle 113 (FIG. 4). Optional package subsystem 545 can interface with physical storage 541 through package interface 539 to receive and discharge the contents of optional physical storage 541. Optional physical storage 541 can provide and receive package information 543 concerning the status of the contents of optional physical storage 541.

Referring now to FIG. 18A, system 600 for moving utility vehicle 113 from at least one first location to at least one second location, another configuration of system 100 (FIG. 2), can include, but is not limited to including, at least one layer. In some configurations, the at least one layer can include autonomous layer 701, supervisory autonomous layer 703, and human autonomous layer 705. Autonomous layer 701 can enable autonomous control of utility vehicle 113, whether or not utility vehicle 113 is manned or unmanned. In some configurations, utility vehicle 113 can send, for example, video signals to fleet manager 601, and fleet manager 601 and respond with commands to utility vehicle that can travel on a message bus to power base 531. In some configurations, the commands can be made to mimic joystick commands. Utility vehicle 113 can measure the latency of the connection between utility vehicle 113 and fleet manager 601, and can adjust the speed of utility vehicle 113 accordingly. If the latency is greater than a pre-selected threshold, utility vehicle 113 can be placed in a semi-autonomous mode. Supervisory autonomous layer 703 can enable remote control of utility vehicle 113. Remote control of utility vehicle 113 can occur as a result of, for example, but not limited to, an unexpected event, pre-selected sensor and processor configurations, and delivery optimization concerns. Human autonomous layer 705 can enable remote event management requiring some form of human intervention. Connections between elements of system 600 indicate functionality groups such as, for example, but not limited to:

Function group
Line format

Remote control
−−−−−−

Mapping/routing
−..−..−..−

Autonomous driving
−.−.−.−.−.

Outputs from autonomous driving
----------------

Continuing to refer to FIG. 18A, in some configurations, autonomous layer 701 can include, but is not limited to including, utility vehicle 113, sensors 118, powerbase 531, and user interface and storage 615. Utility vehicle 113 can create a path based on sensor data and the proposed path, and provide commands to various parts of utility vehicle 113 that enable autonomous behavior of utility vehicle 113. Utility vehicle 113 can follow the created path to a destination, securely execute the services, and securely accept payment for the services. Utility vehicle 113 can respond to sensor data by insuring the safety of pedestrians and other obstacles in and near the created path. For example, if sensors 118 detect an obstacle, utility vehicle 113 can automatically stop and/or change course. Utility vehicle 113 can communicate with sensors 118, user interface/storage 615, motors, signals, and powerbase 531, all of which can be integral parts of utility vehicle 113. Utility vehicle 113 can communicate with remote members of the fleet network through vehicle network interface 623 and communications network 115. Utility vehicle 113 can include robot path processing 621 that can receive a proposed route from infrastructure 6128 through the communications route, and can create a travel path based on the proposed route and data received from sensors 118 through sensor interface 547. Sensors 118 can include, but are not limited to including, close range robust sensors that can enable emergency stop detection by emergency stop subsystem 525 that can direct motor controller 629 to stop utility vehicle 113 through safety subsystem 537 (FIG. 17), and long range sensors. Close range sensors can include features such as, for example, but not limited to, (a) detecting obstacles while traveling at up to a certain pre-selected speed, (b) identifying an obstacles envelope location to within a pre-selected distance, (c) detecting small obstacles on and holes in driving surfaces at least a pre-selected distance away, a pre-selected width, and a pre-selected width, (d) detecting large obstacles on and holes in driving surfaces at least a pre-selected distance away, a pre-selected depth, a pre-selected distance perpendicular to the direction of travel, and a pre-selected length, (e) detecting obstacles at least a pre-selected distance away where the obstacles are a pre-selected height/depth, width (as measured perpendicular to the direction of travel of utility vehicle 113), and length (as measured parallel to the direction of travel of utility vehicle 113), and (f) detecting obstacles no less than a pre-selected distance away under environmental conditions such as, for example, but not limited to, indoors, outdoors, direct sunlight, at night without external illumination, in rain, in snow, and during times of reduced visibility due to fog, smog, and dust. Long range sensors can include characteristics such as, for example, but not limited to, (a) detecting obstacles when utility vehicle 113 is traveling at up to a pre-selected speed, (b) locating obstacles that are moving at up to a pre-selected speed for within a pre-selected distance, (c) estimating the velocity of obstacles moving at up to a pre-selected speed to within a pre-selected tolerance, (d) estimating the direction of obstacles that are moving up to a pre-selected speed to within a pre-selected tolerance and faster than a pre-selected speed to within a pre-selected tolerance, (e) identifying obstacles that move faster than a pre-selected speed, (f) detecting obstacles under pre-selected environmental conditions such as, for example, indoors, outdoors, direct sunlight, and at night without external illumination, and (g) estimating a sensing range in compromised environmental conditions with a pre-selected accuracy, where the environmental conditions can include, but are not limited to including rain up to a pre-selected rate, snow up to a pre-selected rate, reduced visibility due to pre-selected conditions to no less than a pre-selected distance. Long range sensors can detect large obstacles such as, for example, but not limited to, cars, motorcycles, bicycles, fast-moving animals, and pedestrians. Robot path processing 621 can access robot map database 619, that can include local storage for fleet map database 609, and use those data to create a new proposed route, if robot path processing 621 determines that the proposed route is suboptimal. Robot path processing 621 can control, through master controller 627, the direction, based on the created travel path, and speed of utility vehicle 113 through motor controller 629, and can control signaling, through signal controller 631, that can indicate to nearby pedestrians the travel path and speed of utility vehicle 113. Remote control 625 can augment sensor data with data received from infrastructure 6128. Utility vehicle 113 can receive requests to execute services from UI 615 through UI interface 617.

Referring now to FIG. 18B, robot path processing 621 can use sensor information and map data to dynamically plan a path for utility vehicle 113 (FIG. 18A). The goal of robot path processing 621 is to create a substantially obstacle-free path for utility vehicle 113 (FIG. 18A). Map data can include drivable surfaces that can meet certain criteria such as, for example, but not limited to, the surface is within a pre-selected number of degrees of horizontal, at least a pre selected width and length, reachable by driving over curbs no higher than a pre-selected height, and reachable by traversing stairs. Driving surfaces can be classified by type. Types can include, but are not limited to including, road lanes on carriage ways, concrete/asphalt side walks, dirt/grass side walks, bike lanes, road crossings, stair landings, corridors, and interior rooms. Map data can include the location, orientation, and height of the curbs. Map data can include the location, orientation, and intent of traffic signs and signals along the drivable surfaces. Map data can include the relationships between the traffic signs and signals and the drivable surfaces. Map data can include any required activation mechanism for the traffic signals. Map data can include the location, orientation, and activation mechanism for gates, doors, and other pedestrian traffic barriers, as well as the location, orientation, and number of stairs in stair cases. Map data can include the location, orientation, and activation mechanisms for elevators. Map data can include localization features for the drivable surfaces, and can include LIDAR and image data to facilitate localization of utility vehicle 113 (FIG. 18A). Map data can include an association between street addresses and the entrances to premises. Map data can include elevation expressed, but is not limited to being expressed, as floors above ground and height, for example, in meters.

Continuing to refer to FIG. 18B, robot path processing 621 can begin with a proposed route that can be locally-determined or provided by, for example, but not limited to, route planning 503 (FIG. 18A), between the starting location and the destination. Robot path processing 621 can include, but is not limited to including, perception subsystem 536, path planning subsystem 517, and path following subsystem 523. Perception subsystem 536 can include, but is not limited to including, processes such as localization process 653 that can determine the location and orientation of utility vehicle 113 (FIG. 18A). Perception subsystem 536 can include object detection process 655 that can detect objects and obstacles based at least on sensor data, and object identification process 657 that can identify the detected objects and obstacles based at least on systems trained to identify objects. Perception subsystem 536 can include object parameter estimator process 659 that can estimate a measurement of parameters that can be associated with the identified objects, for example, but not limited to size, shape, speed, and acceleration based at least on systems trained to associate the measurements with the identified objects. Perception subsystem 536 can include object modeling process 661 that can, based at least on object identification and object parameter estimation, create a model, based at least on training system data, of how the object or obstacle will behave, and propagate the behavior of the object or obstacle into the future to determine possible interaction, if any, between the object or obstacle and utility vehicle 113 (FIG. 18A). Perception subsystem 536 can include dynamic map cross check 521 that can perform an estimate of the free space available for utility vehicle 113 (FIG. 18A) to navigate, and will use that estimate to cross-check the route map that is created by route planning 503 (FIG. 18A). The estimate is based at least on, for example, but not limited to, the data derived from image segmentation or point cloud segmentation. Free space is the obstacle-free drivable space around utility vehicle 113 (FIG. 18A). Map cross check 521 can access data along the proposed route from robot map database 619 and check the planned travel path against map updates and further sensor data. Robot map database 619 can receive updates from fleet map database 609 through the communications route. Fleet map database 609 can be updated under conditions such as, for example, but not limited it, if an obstacle has been detected for a pre-selected period of time. The combination of perception subsystem 536 and map cross check process 521 can produce a travel path, checked map 515, for utility vehicle 113 (FIG. 18A) that can be provided to path planning subsystem 517. Path planning subsystem 517 can include, but is not limited to including, path planning control process 667 and path cross check process 519. Path planning control process 667 can translate the travel path into commands that can be understood by master controller 627. The commands can direct utility vehicle 113 (FIG. 18A) to the starting location, and then to the destination where the services are executed. Path cross check process 519 can update the travel path based on sensor data, if necessary. Path planning subsystem 517 can provide the updated (if necessary) travel path to path following process 523. Path following process 523 can provide the commands to master controller 627. Master controller 627 can use the commands to control utility vehicle 113 (FIG. 18A) and signaling that can alert pedestrians of the movement of utility vehicle 113 (FIG. 18A). Close range robust sensors 116 can enable master controller 627 to stop utility vehicle 113 (FIG. 18A).

Referring now to FIG. 19, perception subsystem 536 (FIG. 18B) can include localization process 653 that can locate utility vehicle 113 (FIG. 1) on map 751 (FIG. 30) and determine the orientation of utility vehicle 113 (FIG. 1). Sensors 118 (FIG. 18B) can include cameras that can provide visual odometry 801 at high frequency and low fidelity. The cameras can estimate the motion of objects 757 (FIG. 29), and can recognize previously seen corners. The cameras can update, at high frequency, data about utility vehicle 113 (FIG. 1) according to the corners. Sensors 118 (FIG. 18B) can include LIDAR devices that can provide LIDAR odometry 803 at low frequency. The LIDAR data can refine the motion estimate and remove distortion from point clouds. The LIDAR data can be used to recognize previously-seen surfaces and lines, to triangulate from them, and to update data about utility vehicle 113 (FIG. 1) according to the surfaces and lines.

Referring now to FIGS. 20 and 21, perception subsystem 536 (FIG. 18B) and map management pipeline process 611 (FIG. 18A) can include object detection process 655 and object detection/classification process 6551 (FIG. 26) that can access image information 805 (FIG. 20) and/or depth information 807 (FIG. 20), and can classify objects. In some configurations, images can be inspected to find/classify objects, objects can be correlated to depth data, and bounding boxes can be drawn around objects in the depth data with classification. In some configurations, depth data can be inspected for objects, an image region of interest can be created to classify the objects, and bounding boxes can be drawn around objects in the depth data with classification. In some configurations, region-based convolutional neural networks can be used for visual object detection. In some configurations, stereo matching with stixel representation can be used to segment a scene into static background/infrastructure and moving objects. Object detection process 655 (FIG. 18B) and object detection/classification process 6551 (FIG. 26) can generate 2d bounding boxes 809 (FIG. 20) around the classified objects using conventional convolution neural networks. For example, vehicle 2d bounding box 811B (FIG. 21) can surround vehicle 811C in image 811. Pedestrian 2d bounding box 811A (FIG. 21) can surround pedestrian 811D in image 811. Object detection process 655 (FIG. 18B) and object detection/classification process 6551 (FIG. 26) can lift 2d bounding boxes to frustra, creating 3d bounding boxes. For example, vehicle 3d bounding box 813B (FIG. 21) can include vehicle 811C (FIG. 21), and pedestrian 3d bounding box 813A (FIG. 21) can include pedestrian 811D (FIG. 21). The front and back ends of 3d bounding boxes can be detected from a database of point cloud depth data. Raw point cloud data can also be used to provide data to a feature learning network, the feature learning network can partition the space into voxels, and can transform the points within each voxel to a vector representation characterizing the shape information.

Referring now to FIG. 22, object detection process 655 (FIG. 18B) and object detection/classification process 6551 (FIG. 26) can extract points from a bounding box that are associated with the object that has been identified within the bounding box. An associated 2d object classification can be used along with the extracted points to improve the 3d bounding box, i.e. modify the 3d bounding box so that it follows more closely the contours of the object within the 3d bounding box. For example, vehicle 811C (FIG. 21) within vehicle 3d bounding box 813B can be represented by vehicle points 815B, and pedestrian 811D (FIG. 21) within pedestrian 3d bounding box 813A can be represented by pedestrian points 815A. Object parameter estimation process 659 (FIG. 18B) can track bounding boxes in subsequent frames and combine these data with sensor data, such as, for example, but not limited to, radar data, to estimate parameters associated with objects. The parameters can include, but are not limited to including, velocity and acceleration. For example, when pedestrian 811D (FIG. 21) moves, pedestrian points 815A bounded by pedestrian 3d bounding box 813A can move to updated pedestrian 3d bounding box 817A, and can be associated with updated pedestrian points 817B.

Referring now to FIG. 23, object parameter estimation process 659 can combine the updated bounding box and point data with 2d classification information to produce a dynamic map scene. Object model/propagation process 661 can predict the movement of the objects in the dynamic map scene according to models associated with the classified objects. For example, pedestrians and moving vehicles can generally follow movement patterns that can enable the prediction of the future locations of these objects. Pedestrian 811D, for example, beginning movement at pedestrian starting location 825, can move at a speed and in a direction that can be estimated based on sensor data, and can be used by object model/propagation process 661 pedestrian model(s) to predict the location of pedestrian 811D at location 829. A measure of uncertainty can be factored into the location prediction based on any number of possible reasons a pedestrian would not follow a standard model. Pedestrian 811D can end up at ending location 829 or anywhere within uncertainty area 821. Utility vehicle 113 can begin travel at starting location 827 and can travel to ending location 823 in an amount of time that can be predicted by models of utility vehicles 113 executed by object model/propagation process 661. Object model/propagation process 661 (FIG. 18B) can estimate whether utility vehicle 113 will encounter obstacles based on the predicted starting and ending locations of utility vehicle 113 and any obstacles that could end up in its path. The proposed route can be modified depending on expected obstacles.

Referring now to FIG. 24, path planning subsystem 517 can include, but is not limited to including, path planning control process 667 can include guided policy search that can use differential dynamic programming to generate guiding samples to assist in the policy search by exploring high reward regions. In some configurations, features 824 such as, for example, but not limited to, acceleration, deceleration, turn left, and turn right, and labels 826 such as, for example, state and action can be used to create models for path planning. The relationships between feature values 828 can be used to create the model. In some configurations, when features include actions, feature values 828 can be based at least on the reward for performing the action, the learning rate of the neural network, and the best reward obtainable from the state where the action places the actor. For example, when pedestrian 811D and utility vehicle 113 move, the model executed by path planning control process 667 can determine if/when the path of pedestrian 811D intersects with the path of utility vehicle 113 by predicting the movement of both pedestrian 811D and utility vehicle 113 using the model.

Referring now to FIG. 25, confidence values 832 can indicate the likelihood that the model predictions accurately predict the path convergence between obstacles. Confidence values 832 can be determined as the model is developed by executing the model in under test conditions. According to the model executed by path planning process 667, the likelihood of path convergence is highest in region 832, lowest in region 836, and moderately high in region 834.

Continuing to still further refer to FIG. 18A, when supervisory autonomous layer 703 is activated, remote control interface 603 can automatically control utility vehicle 113. Remote control interface 603 can, for example, receive data from system collectors 119 such as, for example, but not limited to, beacon 119C (FIG. 1) that can supplement and/or replace data that can be locally received by sensors 118 associated with utility vehicle 113. Beacon 119C (FIG. 1) can, for example, include overhead sensors whose data can be used to automatically update the delivery route being executed by utility vehicle 113. In some configurations, supervisory autonomous layer 703 can include, but is not limited to including, autonomous layer 701, remote control interface 603, fleet network interface 613, route planning 503, fleet map database 609, and map management pipeline 611. Route planning 503 can access fleet map database 609 and can prepare a proposed route between the location of the goods and the goods' destination. Route planning 503 can provide the proposed route to utility vehicle 113 through fleet network interface 613, communications network 115, and vehicle network interface 623 (also referred to herein as the communications route). Remote control interface 603 can automatically control the direction and speed of utility vehicle 113 as utility vehicle 113 travels along the updated delivery route based at least in part on data from system collectors 119. Supervisory autonomous layer 703 can take over control when, for example, but not limited to, utility vehicle 113 recognizes that sensors 118 could be returning faulty or no data. When faulty or no sensor data are available to utility vehicle 113 to continually update its travel route, utility vehicle 113 may request assistance from remote control interface 603.

Referring now primarily to FIG. 26, map management pipeline process 611 can provide maps to route planning process 503 (FIG. 18A), and those maps can be provided to utility vehicle 113 through the communications route. To provide maps, map management pipeline process 611 can access current map data 751 (FIG. 29), localize the data, detect and classify objects and surfaces, remove unwanted objects, and update current map data. Map management pipeline process 611 can include, but is not limited to including data collection process 652, route localization process 6531, object detection/classification process 6551, surface detection/classification process 658, object removal process 656, and map update process 662. Data collection process 652 can receive sensor data 753 (FIG. 29) from system collectors 119C and provide the data to localization process 6531. Localization process 6531 can receive sensor data 753 (FIG. 29) and current map data 751 (FIG. 29) from robot map database 619A. Robot map database 619A can include, but is not limited to including, map data as described herein. Other data that can optionally be included are pedestrian traffic densities, pedestrian crossing requirements, traffic signs, sidewalk locations, sidewalk conditions, and non-sidewalk drivable area. Current map data 751 (FIG. 29) can include information about the route between a starting location and a destination. Localization process 653 can create localized data 755 (FIG. 29) from current map data 751 (FIG. 29) and sensor data 753 (FIG. 29). Object detection process 6551 can detect and classify localized objects in current map data 751 (FIG. 29) and sensor data 753 (FIG. 29), and object removal process 656 can remove objects that meet pre-selected criteria from localized data 755 (FIG. 29). Surface detection process 658 can detect and classify localized surfaces in the current map data and the system collector data. Surface detection process 658 can detect solid surfaces such as, for example, but not limited to, brick walls, building corners, and jersey barriers. Surface detection process 658 can locate approximately horizontal surfaces, for example, but not limited to, surfaces that rise no more than a pre-selected number of degrees from horizontal. Surface detection process 658 can create a polygon in point cloud data associated with the delivery area, and match the polygon to an image that is temporally coincident with the point cloud data. The polygon can be projected onto the image, and the image within the polygon can be identified. Once identified, the image can be used to teach surface detection process 548 to identify the image automatically. Object removal process 656 and surface detection process 658 can provide the detected and classified objects 757 (FIG. 29) and surfaces, for example, but not limited to, driving surfaces 759 (FIG. 29), to map update process 662 which can update current map data 751 (FIG. 29) and provide the updated current map data to robot map database 619A.

Referring again to FIG. 18A, supervisory autonomous layer 703 can include remote control interface 603 that can take control of utility vehicle 113 under at least one pre-selected circumstance. Remote control interface 603 can receive sensor data and plan a path for utility vehicle 113 in real time. Remote control interface 603 can include, but is not limited to including, real time multi-robot path planning 503A, object identification and tracking 655A, robot tracking 603C, data receivers from utility vehicle 113, and data receivers for sensor data. Remote control interface 603 can execute in, for example, but not limited to, beacon 120C or any system collector 119C (FIG. 2) in the vicinity of utility vehicle 113. Real time multi-robot path planning 503A can receive data from any source that is in the vicinity of remote control interface 603 and utility vehicle 113. In some configurations, real time multi-robot path planning 503A can receive sensor data from traffic light interface 7122. Traffic light interface 7122 can receive sensor data from sensors mounted on traffic lights and other stationary features. In some configurations, real time multi-robot path planning 503A can receive sensor data from object identification and tracking process 655A. In some configurations, object identification and tracking process 655A can receive and process LIDAR 7124 and camera 7126 data. In some configurations, real time multi-robot path planning 503A can receive telemetry data 603A from vehicle tracking process 603C. Vehicle tracking process 603A can process data from utility vehicle 113 to create telemetry data 603A. Real time multi-robot path planning 503A can use the received data to prepare an obstacle-free path for utility vehicle according to conventional path planning methods. Real time multi-robot path planning 503A can provide the path to vehicle command 603B that can generate movement commands for utility vehicle 113.

Referring again to FIG. 18A, in some configurations, fleet manager 601 can manage dispatcher 501 by insuring that utility vehicles 113 are efficiently allocated, and can monitor deployment. Fleet manager 601 can receive requests for deliveries and decide which utility vehicles 113 are available and/or which utility vehicles 113 can most efficiently perform the requested deliveries. Fleet manager 601 can direct dispatcher 501 to begin the process of providing utility vehicle(s) 113 for a requested delivery. Dispatcher 501 can provide the location of the goods and the destination to which the goods are to be provided to route planning 503.

Referring now primarily to FIGS. 27 and 28, supervisory autonomous layer 703 (FIG. 18A) can rescue utility vehicle 113 (FIG. 28) automatically under some circumstances. In other circumstances, utility vehicle 113 (FIG. 28) can encounter situations that could require a non-automated response. Human autonomous layer 705 (FIG. 18A) can provide such support. One way to determine to which layer to hand over control of utility vehicle 113 (FIG. 28) is to determine if sensors that are providing path-relevant information to utility vehicle 113 (FIG. 28) are providing accurate data. Localization process 653 (FIG. 18B) can include a fall-over sequence, that can include method 700 (FIG. 27). Method 700 (FIG. 27) can manage transfer of control of utility vehicle 113 (FIG. 28) when assistance is required. Method 700 (FIG. 27) can include, but is not limited to including, calculating 702 (FIG. 27), by localization process 653 (FIG. 18B), a confidence interval in the sensors whose data are used by localization process 653 (FIG. 18B), for example, but not limited to, sensors 118 (FIG. 18B), that can provide local perception. The confidence interval is calculated based at least on whether the signal to noise ratio in the sensor data is low, for example, if image contrast is within a pre-selected range. If 704 (FIG. 27) the confidence interval is greater than or equal to a pre-selected percentage, method 700 (FIG. 27) can include transferring 707 (FIG. 27) control, by localization process 653 (FIG. 18B), to object detection process 655 (FIG. 18B). After completing execution of perception process 536 (FIG. 18B) and path planning process 517 (FIG. 18B), method 700 (FIG. 27) can include following 719 (FIG. 27), by path following process 523 (FIG. 18B), the planned path. If 704 (FIG. 27) the confidence interval is less than a pre-selected percentage, method 700 (FIG. 27) can include locating 706 (FIG. 27), by localization process 653 (FIG. 18B), at least one of system collectors 119 (FIG. 2) that can satisfy pre-selected threshold criteria or a single threshold criterion. The threshold criteria can include, but are not limited to including, geographic location relative to utility vehicle 113 (FIG. 28), height of system collector 119 (FIG. 2), processing capability of system collector 119 (FIG. 2), and status of system collector 119 (FIG. 2). If 709 (FIG. 27) utility vehicle 113 (FIG. 28) and the located of system collectors 119 (FIG. 2) can communicate electronically, method 700 (FIG. 27) can include requesting 711 (FIG. 27), by localization process 653 (FIG. 18B), path planning instructions from the located of system collectors 119 (FIG. 2). If 713 (FIG. 27) the located of system collectors 119 (FIG. 2) can prepare a planned path for utility vehicle 113 (FIG. 28), method 700 (FIG. 27) can include receiving 715 (FIG. 27), by path following process 523 (FIG. 18B), the planned path, and following 719 (FIG. 27), by path following 523 (FIG. 18B), the planned path. If 713 (FIG. 27) the located of system collectors 119 (FIG. 2) cannot prepare a planned path for utility vehicle 113 (FIG. 28), or if 709 (FIG. 27) utility vehicle 113 (FIG. 28) and the located of system collectors 119 (FIG. 2) cannot communicate electronically, method 700 (FIG. 27) can include requesting 717 (FIG. 27), by localization system 653 (FIG. 18B) assistance from infrastructure 6128 (FIG. 18A).

Referring now to FIGS. 29 and 30, route planning process 503 (FIG. 30) can create a route map that can be used to create a path for utility vehicle 113 (FIG. 2) to follow. Route planning process 503 (FIG. 30) can form a series of connected nodes 781 (FIG. 30) based on map 751 (FIG. 29), start location 783 (FIG. 30), and destination location 785 (FIG. 30). Route planning process 503 (FIG. 30) can assign costs 787 (FIG. 30) to each segment at each of nodes 781 (FIG. 30). Assigned costs 787 (FIG. 30) can be based at least on detected and classified objects 757 (FIG. 29), identified driving surfaces 759 (FIG. 29), and localized sensor data 755 (FIG. 29), and conventional route planning algorithms that take into account the distance between nodes 781 (FIG. 30), the road surface, and the complexity of the travel route. Route planning process 503 (FIG. 30) can traverse graph 793 (FIG. 30) of costs 787 (FIG. 30) to create lowest cost path 791 (FIG. 30), and can overlay lowest cost path 791 (FIG. 30) on route map 789 (FIG. 30), a subset of map 751 (FIG. 30) corresponding to the geographic location of lowest cost path 791 (FIG. 30).

Referring now to FIG. 31, method 650 for providing maps to route planning process 503 (FIG. 18A) can include, but is not limited to including, identifying 651 at least one map associated with a delivery area, where the map includes the path between a starting location and a destination. Method 650 can include localizing 653 data associated with the map and with data collected by at least one sensor associated with the delivery area. Method 650 can include detecting 655 at least one localized object in the at least one delivery area, classifying 657 the at least one localized object, and removing 659, when necessary, at least one of the localized objects based at least on removal criteria. Method 650 can include detecting 661 at least one localized surface in the at least one delivery area, and classifying 663 the at least one localized object. Method 650 can include updating 665 the map with the localized objects and surfaces, and planning 667 a utility vehicle route based at least on the updated map.

Referring now to FIG. 32A, human autonomous layer 705 can include, but is not limited to including, autonomous layer 701 (FIG. 18A) and infrastructure 6128 (FIG. 18A). Infrastructure 6128 can include, but is not limited to including, fleet manager 601 that can insure communication and coordination among the fleet members, among which are utility vehicles 113. Fleet manager 601 can, for example, execute in any appropriately-configured processor that is in electronic communications with fleet members. Utility vehicles 113 can send alerts to fleet manager 601, and fleet manager 601 can triage the alerts according to pre-selected criteria. In some configurations, fleet manager 601 can provide a first set of responses to alerts that are generated by utility vehicles 113 that are in a pre-selected geography relative to fleet assets, for example, trucks 2001. Fleet manager 601 can provide a second set of responses, that could be the same or different from the first set of responses, depending on the capabilities of the fleet assets. Fleet manager 601 can provide a third set of responses if utility vehicle 113 reports a malfunction. For example, as part of the third set of responses, fleet manager 601 can locate the fleet asset that is closest to utility vehicle 113 and that also includes repair capabilities appropriate for the malfunction. Fleet manager 601 can provide a fourth set of responses if utility vehicle 113 needs to deliver to an unreachable location, for example, a location that includes unnavigable terrain. For example, as part of the fourth set of responses, fleet manager 601 can request help from a human asset who has been trained to assist utility vehicles 113. Fleet manager 601 can include response sets for any number of use cases.

Referring now to FIG. 32B, fleet manager 601 can manage security screening of any entity that can have access to utility vehicles 113. Fleet manager 601 can include, but is not limited to including, authentication and roll-based access control. For entities that are known to fleet manager 601, credentials can be proved by something that the entity has, or something that the entity knows. For example, when utility vehicle 113 needs help, a remote operator 2002 can authenticate to take control of utility vehicle 113. In some configurations, a local employee known to fleet manager 601 who has a specific capability can take control of utility vehicle 113 and can authenticate to fleet manager 601 to perform a task. In some configurations, the entity can have a portable device, for example, that can be used as a credential. In some configurations, characteristics of the entity that are known to fleet manager 601 can be used for security purposes, for example, the entity's employment shift and employment location. Fleet manager 601 can manage authentication by remote entities or local entities. Authentication can be achieved by, for example, entering a password into a portable device when pre-selected criteria are met. Fleet manager 601 can identify utility vehicles 113 through, for example, but not limited to, cryptographic keys managed by fleet manager 601. Fleet manager 601 can combine the provable identity of utility vehicle 113 with the provable identity of worker 2003, can check the access controls of worker 2003, and can signal utility vehicle 113 to allow access to utility vehicle 113 by worker 2003. Access can include physical or remote access. If utility vehicle 113 is not able to communicate with fleet manager 601, fleet manager 601 can deploy assistants to rescue utility vehicle 113.

While the present teachings have been described in terms of specific configurations, it is to be understood that they are not limited to these disclosed configurations. Many modifications and other configurations will come to mind to those skilled in the art to which this pertains, and which are intended to be and are covered by both this disclosure and the appended claims. It is intended that the scope of the present teachings should be determined by proper interpretation and construction of the appended claims and their legal equivalents, as understood by those of skill in the art relying upon the disclosure in this specification and the attached drawings.

Claims

1. A utility execution system delivering goods from at least one starting point to at least one utility execution point, the utility execution system comprising: a plurality of system collectors, the system collectors forming a communications network, the system collectors accessing historic data associated with a proposed path between the at least one starting point and the at least one utility execution point, the plurality of system collectors including at least one utility vehicle, the at least one utility vehicle including at least one sensor and at least one storage container, the at least one storage container housing the goods, the historic data including vehicle data previously collected along the proposed path, the plurality of system collectors collecting real time data about the proposed path before and while the at least one utility vehicle navigates the proposed path, at least one of the plurality of system collectors updating the proposed path based at least on the vehicle data, the historic data, and the real time data; anda processor continually updating, based at least on the historic data, the real time data, and the at least one sensor, the updated proposed path as the at least one utility vehicle navigates the updated proposed path from the at least one starting point to the at least one utility execution point, the processor configured to locate at least one obstacle in the updated proposed path, the processor updating the updated proposed path when the at least one obstacle is discovered, the processor configured to access image information, the processor configured to classify the at least one obstacle in the image information, the processor configured to draw bounding boxes around the classified at least one obstacle, the processor configured to segment the classified at least one obstacle into static obstacles and moving obstacles, the processor configured to track the bounding boxes of the moving obstacles over time, the processor configured to combine the tracked bounding boxes with sensor data from the at least one sensor to estimate parameters associated with the moving obstacles, the processor configured to predict movement of the moving obstacles based at least on the parameters and models associated with the moving obstacles forming an obstacle movement prediction, the processor configured to estimate uncertainty associated with the obstacle movement prediction, the processor configured to predict if the moving obstacles will end up in the updated proposed path of the at least one utility vehicle based at least on the obstacle movement prediction forming a collision prediction, the processor configured to modify the updated proposed path based at least on the collision prediction.
2. The utility execution system as in claim 1 wherein the processor executes in the at least one utility vehicle.
3. The utility execution system as in claim 1 wherein the processor executes in a server.
4. The utility execution system as in claim 1 wherein the plurality of system collectors comprises: at least one autonomous vehicle.
5. The utility execution system as in claim 1 wherein the plurality of system collectors comprises: at least one beacon positioned along the updated proposed path, the at least one beacon receiving and transmitting data over the communications network.
6. The utility execution system as in claim 1 wherein the plurality of system collectors comprises: at least one beacon positioned along the updated proposed path, the at least one beacon providing fiducial information to the utility execution system.
7. The utility execution system as in claim 1 wherein the plurality of system collectors comprises: at least one vehicle operating on a city sidewalk.
8. The utility execution system as in claim 1 wherein the plurality of system collectors comprises: at least one vehicle operating on a rural street.
9. The utility execution system as in claim 1 wherein the at least one utility vehicle comprises: at least one localization subsystem detecting, based at least on the historic data and the real time data, the current location and situation of the at least one utility vehicle.
10. The utility execution system as in claim 1 wherein the at least one utility vehicle comprises: at least one localization subsystem detecting, based at least on the historic data, the current location and situation of the at least one utility vehicle.
11. The utility execution system as in claim 1 wherein the plurality of system collectors comprises: at least one wireless access point.
12. The utility execution system as in claim 1 wherein the at least one utility vehicle comprises: a preferred route subsystem determining at least one preferred path between the at least one starting point and the at least one utility execution point based at least on the historic data and the real time data, the preferred route subsystem determining at least one avoidable path between the at least one starting point and the at least one utility execution point based at least on the number of the at least one obstacle in the updated proposed path.
13. The utility execution system as in claim 12 further comprising: a dispatching mechanism coupling the at least one delivery truck with the at least one utility vehicle, the dispatching mechanism tracking battery life in the at least one utility vehicle, the dispatching mechanism enabling the at least one utility vehicle to respond to a summons.
14. The utility execution system as in claim 1 wherein the at least one utility vehicle comprises: a road obstacle-climbing subsystem detecting at least one road obstacle, the road obstacle-climbing subsystem commanding the at least one utility vehicle to crest the at least one road obstacle, the road obstacle-climbing subsystem commanding the at least one utility vehicle to maintain balance and stability while traversing the at least one road obstacle.
15. The utility execution system as in claim 14 wherein the at least one road obstacle comprises a curb.
16. The utility execution system as in claim 14 wherein the at least one road obstacle comprises a step.
17. The utility execution system as in claim 1 wherein the at least one utility vehicle comprises: a stair-climbing subsystem detecting at least one stair, the stair-climbing subsystem commanding the at least one utility vehicle to encounter and traverse the at least one stair, the stair-climbing subsystem commanding the at least one utility vehicle to achieve stabilized operation while traversing the at least one stair.
18. The utility execution system as in claim 1 wherein the processor comprises: a rules compliance subsystem accessing navigation rule information from at least one of the historic data, the real time data, and the at least one sensor, the rules compliance subsystem commanding the at least one utility vehicle to navigate at least according to the navigation rule information, the system collectors learning the navigation rule information as the system collectors operate and interact with the updated proposed navigation path.
19. The utility execution system as in claim 1 further comprising a training subsystem including a neural network.
20. The utility execution system as in claim 1 wherein the at least one utility vehicle comprises: a grouping subsystem commanding at least one second of the at least one utility vehicle to follow a first of the at least one utility vehicle, the grouping subsystem maintaining a coupling between the first utility vehicle and the at least one second utility vehicle.
21. The utility execution system as in claim 20 wherein the coupling comprises an electronic coupling.
22. The utility execution system as in claim 1 wherein the at least one utility vehicle comprises: at least one battery, the battery including a quick charge feature, the quick charge feature accommodating a minimum amount of non-operational time of the at least one utility vehicle.
23. The utility execution system as in claim 22 wherein the at least one battery comprises: a locking feature locking the at least one battery to the at least one utility vehicle, the locking feature including a security feature to enable removal of the at least one battery.
24. The utility execution system as in claim 1 wherein the at least one utility vehicle comprises: at least one battery, the battery including a quick change feature, the quick change feature accommodating a minimum amount of non-operational time of the at least one utility vehicle.
25. The utility execution system as in claim 1 further comprising: a sensor subsystem processing data from the at least one sensor, the at least one sensor including: at least one heat sensor sensing live objects;at least one camera sensing moving objects;at least one laser sensor providing a point cloud representation of an object, the laser sensor sensing distance to an obstacle;at least one ultrasonic sensor sensing distance to the obstacle; andat least one radar sensor sensing speed of the obstacle, and weather and traffic proximate to the at least one utility vehicle;a sensor fusion subsystem fusing data from a plurality of the at least one sensor, the sensor fusion subsystem classifying the at least one obstacle; anda behavior model subsystem predicting a future position of the at least one obstacle.
26. The utility execution system as in claim 1 further comprising: a sensor subsystem processing data from the at least one sensor, the at least one sensor including at least two of: at least one heat sensor sensing dynamic objects;at least one camera sensing moving objects;at least one laser sensor providing a point cloud representation of an object, the laser sensor sensing distance to an obstacle;at least one ultrasonic sensor sensing distance to the obstacle; andat least one radar sensor sending speed of the obstacle, and weather and traffic proximate to the at least one utility vehicle;a sensor fusion subsystem fusing data from a plurality of the at least one sensor, the sensor fusion subsystem classifying the at least one obstacle; anda behavior model subsystem predicting a future position of the at least one obstacle.
27. The utility execution system as in claim 1 wherein the plurality of system collectors comprises: at least one delivery truck transporting the goods to the at least one utility vehicle, the at least one delivery truck transporting the at least one utility vehicle to at least one delivery location.
28. The utility execution system as in claim 27 wherein the at least one delivery truck enabling exchanging of a spent at least one battery with a charged at least one battery in the at least one utility vehicle.
29. The utility execution system as in claim 27 wherein the at least one delivery truck comprises at least one battery charging feature.
30. The utility execution system as in claim 27 wherein the at least one delivery truck comprising: at least one lift mechanism enabling ingress and egress of the at least one utility vehicle.
31. The utility execution system as in claim 27 wherein the at least one delivery truck comprising: at least one in-lift feature enabling ingress of the at least one utility vehicle; andat least one out-lift feature enabling egress of the at least one utility vehicle,wherein the delivery truck is capable of moving during the ingress and the egress.
32. The utility execution system as in claim 27 wherein the plurality of system collectors comprises: at least one airborne vehicle transporting the goods to the at least one delivery truck.
33. The utility execution system as in claim 1 wherein the plurality of system collectors comprises: at least one beacon sensing the at least one obstacle, the at least one beacon enabling communication among the plurality of system collectors, the at least one beacon protecting the sensor data exchanged between the at least one beacon and the plurality of system collectors from tampering, the at least one beacon being physically separate from the at least one utility vehicle, the at least one beacon collecting the sensor data from the updated proposed path.
34. The utility execution system as in claim 1 wherein at least one of the static obstacles comprises a stair.

CROSS REFERENCE TO RELATED APPLICATIONS

This application claims the benefit of U.S. Provisional Application Ser. No. 62/682,129, filed Jun. 7, 2018, entitled SYSTEM AND METHOD FOR DISTRIBUTED UTILITY SERVICE EXECUTION, which is incorporated herein by reference in its entirety.

US Referenced Citations (822)

Number	Name	Date	Kind
584127	Draullette et al.	Jun 1897	A
849270	Schafer	Apr 1907	A
880823	Redfield	Mar 1908	A
2224411	Smith	Dec 1940	A
2415056	Wheeler	Jan 1947	A
2618447	Lecarme	Nov 1952	A
2742973	Johannesen	Apr 1956	A
2966223	Gleasman	Dec 1960	A
3017199	Sechrist	Jan 1962	A
3145797	Taylor	Aug 1964	A
3179355	Pickering	Apr 1965	A
3260324	Suarez	Jul 1966	A
3283398	Andren	Nov 1966	A
3288234	Feliz	Nov 1966	A
3306626	Kawada	Feb 1967	A
3313365	Jackson	Apr 1967	A
3338328	Cataldo	Aug 1967	A
3348518	Forsyth	Oct 1967	A
3374845	Selwyn	Mar 1968	A
3399742	Malick	Sep 1968	A
3446304	Alimanestiand	May 1969	A
3450219	Fleming	Jun 1969	A
3515401	Gross	Jun 1970	A
3580344	Floyd	May 1971	A
3596298	Durst, Jr.	Aug 1971	A
3628624	Wesener	Dec 1971	A
3718342	Freed	Feb 1973	A
3787066	Hautier	Jan 1974	A
3790150	Lippert	Feb 1974	A
3860264	Douglas	Jan 1975	A
3872945	Hickman	Mar 1975	A
3893689	Verhoff	Jul 1975	A
3952822	Udden	Apr 1976	A
3965402	Mogle	Jun 1976	A
3993154	Simmons et al.	Nov 1976	A
4005907	Bonomo	Feb 1977	A
4018440	Deutsch	Apr 1977	A
4030753	Meiners	Jun 1977	A
4054319	Fogg et al.	Oct 1977	A
4062558	Wasserman	Dec 1977	A
4065145	Chambers	Dec 1977	A
4065146	Denzer	Dec 1977	A
4076270	Winchell	Feb 1978	A
4078627	Brown et al.	Mar 1978	A
4087107	Winchell	May 1978	A
4088199	Trautwein	May 1978	A
4094372	Notter	Jun 1978	A
4109741	Gabriel	Aug 1978	A
4111445	Haibeck	Sep 1978	A
4115445	Hearsey	Sep 1978	A
4140200	Tucek	Feb 1979	A
4151892	Francken	May 1979	A
D253234	Cooke	Oct 1979	S
4222449	Feliz	Sep 1980	A
4264082	Fouchey, Jr.	Apr 1981	A
4266627	Lauber	May 1981	A
4274503	Charles Mackintosh	Jun 1981	A
4281734	Johnston	Aug 1981	A
4293052	Daswick	Oct 1981	A
4307788	Shelton	Dec 1981	A
4325565	Winchell	Apr 1982	A
4354569	Eichholz	Oct 1982	A
D266758	Johannsen et al.	Nov 1982	S
4363493	Veneklasen	Dec 1982	A
4373600	Buschbom	Feb 1983	A
4375840	Campbell	Mar 1983	A
4413693	Derby	Nov 1983	A
4448455	Ellegaard	May 1984	A
4456086	Wier	Jun 1984	A
4484648	Jephcott	Nov 1984	A
4510956	King	Apr 1985	A
4512588	Cox	Apr 1985	A
4556997	Takamiya	Dec 1985	A
4560022	Kassai	Dec 1985	A
4566707	Nitzberg	Jan 1986	A
4570078	Yashima	Feb 1986	A
4571844	Komasaku	Feb 1986	A
4624469	Bourne, Jr.	Nov 1986	A
4648783	Tan	Mar 1987	A
4657271	Salmon	Apr 1987	A
4657272	Davenport	Apr 1987	A
4674584	Watkins	Jun 1987	A
4685693	Vadjunec	Aug 1987	A
4709772	Brunet	Dec 1987	A
4712806	Patrin	Dec 1987	A
4716980	Butler	Jan 1988	A
4722547	Kishi	Feb 1988	A
4732353	Studer	Mar 1988	A
4740001	Torleumke	Apr 1988	A
4746132	Eagan	May 1988	A
4750578	Brandenfels	Jun 1988	A
4754255	Sanders et al.	Jun 1988	A
4770410	Brown	Sep 1988	A
4778133	Sakurai	Oct 1988	A
4786069	Tang	Nov 1988	A
4787679	Arnold	Nov 1988	A
4790400	Sheeter	Dec 1988	A
4790548	Decelles	Dec 1988	A
4794730	Fischbach	Jan 1989	A
4794999	Hester	Jan 1989	A
4798255	Wu	Jan 1989	A
4802542	Houston	Feb 1989	A
4809804	Houston	Mar 1989	A
4834200	Kajita	May 1989	A
4837694	Narita et al.	Jun 1989	A
4863182	Chern	Sep 1989	A
4867188	Reid	Sep 1989	A
4869279	Hedges	Sep 1989	A
4874055	Beer	Oct 1989	A
4890853	Olson	Jan 1990	A
4897070	Wagstaff	Jan 1990	A
4913252	Bartley et al.	Apr 1990	A
4919225	Sturges	Apr 1990	A
D308364	Beasley, Jr. et al.	Jun 1990	S
4941854	Takahashi et al.	Jul 1990	A
4944360	Sturges	Jul 1990	A
4953851	Sherlock	Sep 1990	A
4964679	Rath	Oct 1990	A
4967862	Pong et al.	Nov 1990	A
4973071	Ishizaki	Nov 1990	A
4984754	Yarrington	Jan 1991	A
4985947	Ethridge	Jan 1991	A
4998596	Miksitz	Mar 1991	A
5001636	Shiraishi et al.	Mar 1991	A
5002295	Lin	Mar 1991	A
5011171	Cook	Apr 1991	A
5012176	LaForge	Apr 1991	A
RE33675	Young	Aug 1991	E
5044457	Aikman	Sep 1991	A
5052237	Reimann	Oct 1991	A
5076390	Haskins	Dec 1991	A
5087103	Pompier	Feb 1992	A
5088761	Takehara et al.	Feb 1992	A
5098041	Uetrecht	Mar 1992	A
5111899	Reimann	May 1992	A
5123972	Ostrander	Jun 1992	A
5124938	Algrain	Jun 1992	A
5125468	Coker	Jun 1992	A
5127709	Rubinstein	Jul 1992	A
5136219	Takahashi	Aug 1992	A
5158493	Morgrey	Oct 1992	A
5161820	Vollmer	Nov 1992	A
5165711	Tsai	Nov 1992	A
5168947	Rodenborn	Dec 1992	A
5171173	Henderson	Dec 1992	A
5186270	West	Feb 1993	A
5208521	Aoyama	May 1993	A
5217246	Williams	Jun 1993	A
5221883	Takenaka	Jun 1993	A
5229068	Johansson et al.	Jul 1993	A
5241875	Kochanneck	Sep 1993	A
5248007	Watkins	Sep 1993	A
5261503	Yasui	Nov 1993	A
5274576	Williams	Dec 1993	A
5276588	Repplinger	Jan 1994	A
5276624	Ito	Jan 1994	A
5297646	Yamamura	Mar 1994	A
5307888	Urvoy	May 1994	A
5307892	Philips	May 1994	A
5314034	Chittal	May 1994	A
5350033	Kraft	Sep 1994	A
5364165	Okamoto	Nov 1994	A
5366036	Perry	Nov 1994	A
5369580	Monji	Nov 1994	A
5376868	Toyoda	Dec 1994	A
D355148	Orsolini	Feb 1995	S
5388658	Ando et al.	Feb 1995	A
5397890	Schueler	Mar 1995	A
5408411	Nakamura	Apr 1995	A
5408811	Satake	Apr 1995	A
5417298	Shibahata	May 1995	A
5419624	Adler	May 1995	A
5450919	Shitani	Sep 1995	A
5465806	Higasa	Nov 1995	A
5482125	Pagett	Jan 1996	A
D373121	Deiuliis et al.	Aug 1996	S
5551756	Gurasich et al.	Sep 1996	A
5575348	Goertzen	Nov 1996	A
5576959	Hrovat	Nov 1996	A
D376585	Wathen et al.	Dec 1996	S
5615116	Gudat	Mar 1997	A
D381325	McMahan et al.	Jul 1997	S
5646845	Gudat	Jul 1997	A
5649605	Ronne et al.	Jul 1997	A
5657828	Nagamachi	Aug 1997	A
5695021	Schaffner	Dec 1997	A
5701965	Kamen et al.	Dec 1997	A
5701968	Wright-Ott	Dec 1997	A
5705746	Trost	Jan 1998	A
5732379	Eckert	Mar 1998	A
5743347	Gingerich	Apr 1998	A
5746282	Fujiwara	May 1998	A
5769441	Namngani	Jun 1998	A
5774819	Yamamoto et al.	Jun 1998	A
5775452	Patmont	Jul 1998	A
5791425	Kamen	Aug 1998	A
5794730	Kamen	Aug 1998	A
5799745	Fukatani	Sep 1998	A
5799914	Chivallier et al.	Sep 1998	A
5826209	Matsuno	Oct 1998	A
D402645	Garguilo	Dec 1998	S
5848660	McGreen	Dec 1998	A
5850136	Kaneko	Dec 1998	A
5869943	Nakashima et al.	Feb 1999	A
5869946	Carobolante	Feb 1999	A
5893896	Imamura et al.	Apr 1999	A
5927414	Kan et al.	Jul 1999	A
5928309	Korver	Jul 1999	A
5931421	Surauer et al.	Aug 1999	A
5939864	Lenhart et al.	Aug 1999	A
5965991	Koike	Oct 1999	A
5971091	Kamen	Oct 1999	A
5973463	Okuda	Oct 1999	A
5975225	Kamen	Nov 1999	A
5986221	Stanley	Nov 1999	A
6002975	Schiffmann	Dec 1999	A
6003624	Jorgensen et al.	Dec 1999	A
6024182	Hamada et al.	Feb 2000	A
6036619	Tashiro	Mar 2000	A
6039142	Eckstein	Mar 2000	A
6050357	Staelin	Apr 2000	A
6052647	Parkinson	Apr 2000	A
6053579	Nelson et al.	Apr 2000	A
6059062	Staelin	May 2000	A
6062600	Kamen et al.	May 2000	A
6062651	Schaad	May 2000	A
6065558	Wielenga	May 2000	A
6073951	Jindra et al.	Jun 2000	A
6076033	Hamada	Jun 2000	A
6089680	Yoshioka et al.	Jul 2000	A
6092249	Kamen et al.	Jul 2000	A
D428936	Serfaty et al.	Aug 2000	S
6105704	Hamada	Aug 2000	A
6108592	Kurtzberg et al.	Aug 2000	A
6123398	Arai	Sep 2000	A
6125953	Arai	Oct 2000	A
6125957	Kauffmann	Oct 2000	A
6131057	Tamaki	Oct 2000	A
6141613	Fan	Oct 2000	A
6148939	Brookhart	Nov 2000	A
6154692	Cielaszyk	Nov 2000	A
D434762	Ikenaga	Dec 2000	S
6169946	Griessbach	Jan 2001	B1
6189643	Takahashi	Feb 2001	B1
6192305	Schiffmann	Feb 2001	B1
6208734	Ortscheid et al.	Mar 2001	B1
6208929	Matsuno et al.	Mar 2001	B1
6212276	Inoue	Apr 2001	B1
6223104	Kamen	Apr 2001	B1
6223114	Boros	Apr 2001	B1
6225977	Li	May 2001	B1
D444184	Kettler	Jun 2001	S
6247548	Hayashi	Jun 2001	B1
6260646	Fernandez et al.	Jul 2001	B1
6263261	Brown	Jul 2001	B1
6264218	Slagerman	Jul 2001	B1
6270105	Friedrich	Aug 2001	B1
6273212	Husted et al.	Aug 2001	B1
6276471	Kratzenberg et al.	Aug 2001	B1
6285778	Nakajima	Sep 2001	B1
6288505	Heinzmann	Sep 2001	B1
6292722	Holmes et al.	Sep 2001	B1
6302230	Kamen	Oct 2001	B1
6311794	Morrell et al.	Nov 2001	B1
6320336	Eguchi	Nov 2001	B1
6324446	Brown et al.	Nov 2001	B1
6325736	Hamada	Dec 2001	B1
6328125	Van Den Brink	Dec 2001	B1
6332103	Steenson, Jr.	Dec 2001	B1
6332104	Brown	Dec 2001	B1
D452692	Fukuda	Jan 2002	S
6343664	Morrell et al.	Feb 2002	B2
6356188	Meyers	Mar 2002	B1
6357544	Kamen	Mar 2002	B1
6360996	Bockman et al.	Mar 2002	B1
6367817	Kamen	Apr 2002	B1
6371228	Husted et al.	Apr 2002	B1
6375209	Schlangen	Apr 2002	B1
6377906	Rowe	Apr 2002	B1
6386576	Kamen et al.	May 2002	B1
6388580	Graham	May 2002	B1
6397046	Kfoury	May 2002	B1
6405816	Kamen et al.	Jun 2002	B1
6408240	Morrell et al.	Jun 2002	B1
6415215	Nishizaki	Jul 2002	B1
6415879	Kamen et al.	Jul 2002	B2
6416272	Suehiro	Jul 2002	B1
6435535	Field	Aug 2002	B1
6435538	Ellis	Aug 2002	B2
6443250	Kamen et al.	Sep 2002	B1
6443251	Morrell et al.	Sep 2002	B1
6446320	Kilgore	Sep 2002	B2
6463369	Sadano	Oct 2002	B2
D466122	Moody	Nov 2002	S
6484829	Cox	Nov 2002	B1
D466516	Peiker	Dec 2002	S
6502011	Haag	Dec 2002	B2
6508319	Langenfeld et al.	Jan 2003	B1
6538411	Field et al.	Mar 2003	B1
6543564	Kamen	Apr 2003	B1
6543848	Yasuo Suga et al.	Apr 2003	B1
6543858	Melton	Apr 2003	B1
6547026	Kamen et al.	Apr 2003	B2
6553271	Morrell	Apr 2003	B1
6554250	Alves et al.	Apr 2003	B2
6556909	Matsumoto	Apr 2003	B2
6561294	Kamen	May 2003	B1
6562511	Daroux	May 2003	B2
6571176	Shinmura	May 2003	B1
6571892	Kamen	Jun 2003	B2
6575539	Reich	Jun 2003	B2
6581714	Kamen	Jun 2003	B1
6582181	Suehiro et al.	Jun 2003	B2
6586901	Singer et al.	Jul 2003	B1
6593849	Chubb	Jul 2003	B2
6598941	Field et al.	Jul 2003	B2
6614343	Fennel	Sep 2003	B1
6615938	Morrell et al.	Sep 2003	B2
6634451	Sakakiyama	Oct 2003	B2
6643451	Tokura et al.	Nov 2003	B1
6647248	Ortscheid et al.	Nov 2003	B1
6651763	Kamen et al.	Nov 2003	B1
6651766	Kamen	Nov 2003	B2
6654674	Lu	Nov 2003	B2
6654675	Pedersen et al.	Nov 2003	B2
6659211	Esposito	Dec 2003	B2
6659570	Nakamura	Dec 2003	B2
D485279	DeCombe	Jan 2004	S
6694225	Aga	Feb 2004	B2
6704622	Tinskey	Mar 2004	B2
6713693	Sadowski et al.	Mar 2004	B1
D489027	Waters	Apr 2004	S
D489029	Waters	Apr 2004	S
D489300	Chang	May 2004	S
6752231	Hume	Jun 2004	B2
D493127	Waters	Jul 2004	S
D493128	Waters	Jul 2004	S
D493801	Byun	Aug 2004	S
D494099	Maurer	Aug 2004	S
6779621	Kamen et al.	Aug 2004	B2
6781960	Charas	Aug 2004	B1
6789640	Arling	Sep 2004	B1
6793258	Gray	Sep 2004	B2
6796396	Kamen	Sep 2004	B2
6799649	Kamen et al.	Oct 2004	B2
6827163	Amsbury et al.	Dec 2004	B2
6856326	Zhai	Feb 2005	B1
D503402	Su et al.	Mar 2005	S
6866107	Heinzmann et al.	Mar 2005	B2
6868931	Morrell	Mar 2005	B2
D503928	Obata	Apr 2005	S
6874591	Morrell et al.	Apr 2005	B2
6889784	Troll	May 2005	B2
6907949	Wang	Jun 2005	B1
D507206	Wang	Jul 2005	S
6920947	Kamen et al.	Jul 2005	B2
6938923	Mulhern et al.	Sep 2005	B2
6962383	Takenoshita et al.	Nov 2005	B2
6965206	Kamen et al.	Nov 2005	B2
6969079	Kamen et al.	Nov 2005	B2
7000933	Arling et al.	Feb 2006	B2
7004271	Kamen et al.	Feb 2006	B1
7006901	Wang	Feb 2006	B2
D517086	Siebel	Mar 2006	S
7017686	Kamen et al.	Mar 2006	B2
D521017	Jewitt	May 2006	S
7040713	Rudolf	May 2006	B2
D524315	Reusing	Jul 2006	S
7090040	Kamen et al.	Aug 2006	B2
D528468	Arling	Sep 2006	S
D529005	Hong	Sep 2006	S
7102328	Long et al.	Sep 2006	B2
7130702	Morrell	Oct 2006	B2
7174976	Kamen et al.	Feb 2007	B2
7178611	Zupanick	Feb 2007	B2
7178614	Ishii	Feb 2007	B2
7182166	Gray et al.	Feb 2007	B2
D539810	Cummins	Apr 2007	S
7198223	Phelps, III et al.	Apr 2007	B2
7210544	Kamen et al.	May 2007	B2
7219912	Meyer	May 2007	B2
D544486	Hussaini	Jun 2007	S
7234779	Bedford et al.	Jun 2007	B2
D546782	Poulet et al.	Jul 2007	S
D549721	Ito	Aug 2007	S
D549722	Ito et al.	Aug 2007	S
D551592	Chang et al.	Sep 2007	S
D551722	Chang et al.	Sep 2007	S
7272681	Davies	Sep 2007	B2
7273116	Kamen et al.	Sep 2007	B2
D552609	Kornblum	Oct 2007	S
7275607	Kamen et al.	Oct 2007	B2
D556149	Kaufhold et al.	Nov 2007	S
D557220	Ewringmann	Dec 2007	S
D557221	Ewringmann	Dec 2007	S
7303032	Kahlert et al.	Dec 2007	B2
7316441	Iwatani et al.	Jan 2008	B2
D564033	Itskov et al.	Mar 2008	S
7363993	Ishii	Apr 2008	B2
7370713	Kamen	May 2008	B1
7399035	Kusanagi et al.	Jul 2008	B2
7481291	Nishikawa	Jan 2009	B2
D585906	Berg et al.	Feb 2009	S
D587660	Lin	Mar 2009	S
7539557	Yamauchi	May 2009	B2
7546889	Kamen et al.	Jun 2009	B2
D598927	Hirsch	Aug 2009	S
7589643	Dagci	Sep 2009	B2
7592900	Kamen et al.	Sep 2009	B2
D601922	Imai et al.	Oct 2009	S
7640086	Nakashima et al.	Dec 2009	B2
7688191	Lu	Mar 2010	B2
7690447	Kamen et al.	Apr 2010	B2
7690452	Kamen et al.	Apr 2010	B2
7703568	Ishii	Apr 2010	B2
D614998	Fujita	May 2010	S
7740099	Field et al.	Jun 2010	B2
D619945	Sadanowicz et al.	Jul 2010	S
7757794	Heinzmann et al.	Jul 2010	B2
7789174	Kamen	Sep 2010	B2
7823676	Yamada et al.	Nov 2010	B2
7856248	Fujisaki	Dec 2010	B1
7857088	Field	Dec 2010	B2
D632229	Kruse	Feb 2011	S
7896440	Tsai	Mar 2011	B2
7900725	Heinzmann et al.	Mar 2011	B2
7917097	Hawkins et al.	Mar 2011	B2
7938207	Kamen et al.	May 2011	B2
7958956	Kakinuma et al.	Jun 2011	B2
D644654	Maitlen et al.	Sep 2011	S
8011459	Serai	Sep 2011	B2
8014923	Ishii	Sep 2011	B2
8025325	Carrier et al.	Sep 2011	B1
8028777	Kakinuma	Oct 2011	B2
8050820	Yanaka	Nov 2011	B2
8050837	Yamada	Nov 2011	B2
8074388	Trainer	Dec 2011	B2
8091672	Gutsch	Jan 2012	B2
8113244	Kamen et al.	Feb 2012	B2
8151912	Koide et al.	Apr 2012	B2
8155828	Fuwa et al.	Apr 2012	B2
8160794	Fuwa	Apr 2012	B2
8162089	Shaw	Apr 2012	B2
8170780	Field	May 2012	B2
8170781	Fuwa	May 2012	B2
8172016	Goertzen et al.	May 2012	B2
8186462	Kamen	May 2012	B2
8224524	Nakashima	Jul 2012	B2
8225891	Takenaka	Jul 2012	B2
8239992	Schnittman	Aug 2012	B2
8248222	Kamen	Aug 2012	B2
8249773	Kawada	Aug 2012	B2
8255105	Weissert	Aug 2012	B2
8265774	Senba	Sep 2012	B2
8269130	Mangan et al.	Sep 2012	B2
8285474	Doi	Oct 2012	B2
8312017	Martin et al.	Nov 2012	B2
8326469	Phillips	Dec 2012	B2
8346441	Miki et al.	Jan 2013	B2
8371410	Fuwa	Feb 2013	B2
D678217	Helm	Mar 2013	S
D678320	Kanalakis, Jr.	Mar 2013	S
8396611	Phillips	Mar 2013	B2
8417404	Yen	Apr 2013	B2
8418705	Ota et al.	Apr 2013	B2
8453768	Kamen	Jun 2013	B2
8467941	Field	Jun 2013	B2
D686200	Huang et al.	Jul 2013	S
8490723	Heinzmann	Jul 2013	B2
8504248	Taira	Aug 2013	B2
8564444	Ota et al.	Oct 2013	B2
8572822	Hasegawa	Nov 2013	B2
8584782	Chen	Nov 2013	B2
8587583	Newcombe	Nov 2013	B2
8621684	Okumatsu	Jan 2014	B2
8636451	Yamashita et al.	Jan 2014	B2
8639416	Jones	Jan 2014	B2
8640807	Takenaka	Feb 2014	B2
8672339	Raike, III	Mar 2014	B2
8672356	Inaguma	Mar 2014	B2
8684123	Chen	Apr 2014	B2
8690265	Noblanc	Apr 2014	B2
D704621	Taylor	May 2014	S
D705799	Funabashi et al.	May 2014	S
8738238	Rekow	May 2014	B2
8738278	Chen	May 2014	B2
D706807	Harre	Jun 2014	S
D707701	d'Amore	Jun 2014	S
8744720	Fujisaki	Jun 2014	B1
8753208	Jaouen et al.	Jun 2014	B2
D708203	Johnson	Jul 2014	S
8775001	Phillips	Jul 2014	B2
8807250	Chen	Aug 2014	B2
8830048	Kamen et al.	Sep 2014	B2
8832875	Odashima et al.	Sep 2014	B2
8843244	Phillips	Sep 2014	B2
D716325	Brudnicki	Oct 2014	S
8860551	Carraher	Oct 2014	B2
D716818	Alegiani	Nov 2014	S
8925563	Ota et al.	Jan 2015	B2
8958976	Kajima	Feb 2015	B2
D723558	Downs	Mar 2015	S
8978791	Ha	Mar 2015	B2
9002535	Powers	Apr 2015	B2
9016410	Trowell et al.	Apr 2015	B2
D729270	Clare	May 2015	S
D729833	Clare	May 2015	S
9038212	Yamaguchi et al.	May 2015	B2
D732062	Kwon	Jun 2015	S
9045190	Chen	Jun 2015	B2
9056629	Kamo	Jun 2015	B2
9079039	Carlson	Jul 2015	B2
9096281	Li	Aug 2015	B1
D738907	Cabrera-Cordon et al.	Sep 2015	S
D738913	Cabrera-Cordon et al.	Sep 2015	S
9126497	Heinzmann	Sep 2015	B2
9156516	Kahlert	Oct 2015	B2
D742300	Fontaeus	Nov 2015	S
D742407	Park	Nov 2015	S
D742795	Siao	Nov 2015	S
9187071	Vinck et al.	Nov 2015	B2
9193066	Ohm	Nov 2015	B2
9218003	Fong	Dec 2015	B2
D747352	Lee et al.	Jan 2016	S
D750179	Foulkes et al.	Feb 2016	S
D752572	Kohler et al.	Mar 2016	S
9278036	Lee	Mar 2016	B2
9309692	Westwinkel	Apr 2016	B2
D755785	Sirotich	May 2016	S
D757732	Galanti	May 2016	S
D758284	Ringer et al.	Jun 2016	S
D762179	Wong	Jul 2016	S
9400044	Wadhva et al.	Jul 2016	B2
D763359	Kwong	Aug 2016	S
D764520	Lee et al.	Aug 2016	S
9403566	Jacobsen	Aug 2016	B2
9404756	Fong	Aug 2016	B2
D765718	Vinna	Sep 2016	S
D766312	Hedges	Sep 2016	S
9455104	Leusenkamp et al.	Sep 2016	B1
D769314	Piroddi	Oct 2016	S
D770514	Bae et al.	Nov 2016	S
D772255	Taylor et al.	Nov 2016	S
D772924	Begin et al.	Nov 2016	S
D772930	Vazquez et al.	Nov 2016	S
D775148	Anzures	Dec 2016	S
9527213	Luo	Dec 2016	B2
D778312	Goodwin et al.	Feb 2017	S
9567021	Mailey	Feb 2017	B2
D784405	Kim et al.	Apr 2017	S
D786278	Motamedi	May 2017	S
D786770	Smallhorn	May 2017	S
D787420	Smallhorn	May 2017	S
D787996	Rode et al.	May 2017	S
9636265	Furuta	May 2017	B2
9656704	Couture	May 2017	B2
9662438	Kamen et al.	May 2017	B2
D791174	Hart et al.	Jul 2017	S
D792444	Cho et al.	Jul 2017	S
D794674	Brush	Aug 2017	S
9730029	Choudhury	Aug 2017	B2
9744879	Drako	Aug 2017	B2
D797772	Mizono et al.	Sep 2017	S
D798318	Ferguson	Sep 2017	S
9750896	Kamen et al.	Sep 2017	B2
9770825	Goldenberg	Sep 2017	B2
D801996	Yang	Nov 2017	S
D802002	Howard et al.	Nov 2017	S
D804393	Yoo et al.	Dec 2017	S
D805972	Lee et al.	Dec 2017	S
D805973	Mullaney	Dec 2017	S
D807235	Collins	Jan 2018	S
D807236	Collins	Jan 2018	S
D807277	Lee et al.	Jan 2018	S
D812533	Lee et al.	Mar 2018	S
D814370	Kim et al.	Apr 2018	S
D816090	Stonecipher et al.	Apr 2018	S
9974467	Blahnik et al.	May 2018	B2
D821410	Vinna et al.	Jun 2018	S
9989970	Morey	Jun 2018	B1
9996157	Chaudhri et al.	Jun 2018	B2
10007391	Sabatelli et al.	Jun 2018	B2
10025472	Sabatelli	Jul 2018	B2
D825437	Hilton et al.	Aug 2018	S
D825493	Chen	Aug 2018	S
D826244	Yampolskaya	Aug 2018	S
D826255	Andrizzi et al.	Aug 2018	S
10055108	Bates	Aug 2018	B2
10055184	Ferrell et al.	Aug 2018	B1
D829740	Lepine et al.	Oct 2018	S
D830384	Lepine et al.	Oct 2018	S
D830385	Lepine et al.	Oct 2018	S
D830386	Lepine et al.	Oct 2018	S
D831046	Hashimoto et al.	Oct 2018	S
D832289	Chen et al.	Oct 2018	S
10088993	Hall	Oct 2018	B2
10127250	Dingman et al.	Nov 2018	B2
10130534	Mattes	Nov 2018	B2
D835118	Lee et al.	Dec 2018	S
D835139	Li	Dec 2018	S
D835141	Li et al.	Dec 2018	S
D835632	Liu et al.	Dec 2018	S
10149589	Lindhe	Dec 2018	B2
D838731	Pillalamarri et al.	Jan 2019	S
10172752	Goffer	Jan 2019	B2
D840413	Leach et al.	Feb 2019	S
D841021	Klar et al.	Feb 2019	S
D841022	Klar et al.	Feb 2019	S
D841676	Zhang	Feb 2019	S
D841687	Muller et al.	Feb 2019	S
10203211	Mishra	Feb 2019	B1
10216188	Brady et al.	Feb 2019	B2
D842897	Kumar	Mar 2019	S
10220843	Coulter	Mar 2019	B2
10222798	Brady et al.	Mar 2019	B1
10229245	Laurance	Mar 2019	B2
10230538	Killian et al.	Mar 2019	B2
10233021	Brady	Mar 2019	B1
10235014	Yang	Mar 2019	B2
10241516	Brady et al.	Mar 2019	B1
D847161	Chaudhri	Apr 2019	S
10266097	Takahata	Apr 2019	B2
10272294	Williams et al.	Apr 2019	B2
D847836	Thoreson	May 2019	S
10296167	Liu	May 2019	B2
10296194	McLean	May 2019	B2
10308430	Brady et al.	Jun 2019	B1
10310499	Brady et al.	Jun 2019	B1
10318589	Sharp	Jun 2019	B2
10338776	Andersson	Jul 2019	B2
D855634	Kim	Aug 2019	S
10372304	Jaramillo, III	Aug 2019	B2
10379695	Carlos	Aug 2019	B2
10386942	Kim	Aug 2019	B2
10423283	Ikeda	Sep 2019	B2
10474737	Girsova et al.	Nov 2019	B1
10532885	Brady et al.	Jan 2020	B1
10628790	Aggarwal	Apr 2020	B1
10901418	Brady et al.	Jan 2021	B2
20010006125	Richey	Jul 2001	A1
20010037163	Allard	Nov 2001	A1
20020007239	Matsumoto	Jan 2002	A1
20020011361	Richey	Jan 2002	A1
20020056582	Chubb	May 2002	A1
20020063006	Kamen	May 2002	A1
20020074189	Hester	Jun 2002	A1
20020082749	Meyers	Jun 2002	A1
20020121394	Kamen	Sep 2002	A1
20020121572	Jacobson	Sep 2002	A1
20020189870	Kamen	Dec 2002	A1
20030014167	Pedersen	Jan 2003	A1
20030128840	Luginbill	Jul 2003	A1
20030226698	Kamen	Dec 2003	A1
20040005958	Kamen	Jan 2004	A1
20040007121	Graves	Jan 2004	A1
20040007399	Heinzmann et al.	Jan 2004	A1
20040007644	Phelps, III et al.	Jan 2004	A1
20040055796	Kamen	Mar 2004	A1
20040069543	Kamen	Apr 2004	A1
20040124655	Takenoshita et al.	Jul 2004	A1
20040135434	Honda	Jul 2004	A1
20040201271	Kakinuma	Oct 2004	A1
20040256886	Wu	Dec 2004	A1
20040262871	Schreuder	Dec 2004	A1
20050029023	Takami	Feb 2005	A1
20050121866	Kamen	Jun 2005	A1
20050134014	Xie	Jun 2005	A1
20050211477	Gray	Sep 2005	A1
20050236208	Runkles	Oct 2005	A1
20050236894	Lu et al.	Oct 2005	A1
20050251292	Casey	Nov 2005	A1
20060108956	Clark	May 2006	A1
20060163437	Lin	Jul 2006	A1
20060187646	Belson et al.	Aug 2006	A1
20060202439	Kahlert	Sep 2006	A1
20060231313	Ishii	Oct 2006	A1
20060279554	Shin	Dec 2006	A1
20060293850	Ahn	Dec 2006	A1
20070001830	Dagci	Jan 2007	A1
20070055424	Peters et al.	Mar 2007	A1
20070085300	Loewenthal	Apr 2007	A1
20070100511	Koerlin	May 2007	A1
20070156286	Yamauchi	Jul 2007	A1
20070198175	Williams	Aug 2007	A1
20070208483	Rabin	Sep 2007	A1
20070213900	Raab	Sep 2007	A1
20070216205	Davis	Sep 2007	A1
20070221423	Chang	Sep 2007	A1
20070296170	Field	Dec 2007	A1
20080029985	Chen	Feb 2008	A1
20080042379	Amran	Feb 2008	A1
20080066974	Pearlman	Mar 2008	A1
20080086241	Phillips	Apr 2008	A1
20080147281	Ishii	Jun 2008	A1
20080149798	Tinoco	Jun 2008	A1
20080174415	Tanida	Jul 2008	A1
20080197599	Comstock	Aug 2008	A1
20080238005	James	Oct 2008	A1
20080294288	Yamauchi	Nov 2008	A1
20080302938	Goodwin et al.	Dec 2008	A1
20090009984	Mangiardi	Jan 2009	A1
20090032323	Kakinuma	Feb 2009	A1
20090037033	Phillips	Feb 2009	A1
20090045025	Bassett	Feb 2009	A1
20090078485	Gutsch	Mar 2009	A1
20090105908	Casey	Apr 2009	A1
20090115149	Wallis	May 2009	A1
20090224524	Rathsack	Sep 2009	A1
20100025139	Kosaka	Feb 2010	A1
20100107076	Laurance	Apr 2010	A1
20100114468	Field	May 2010	A1
20100121538	Ishii	May 2010	A1
20100126787	Kawada	May 2010	A1
20100138128	Strothmann	Jun 2010	A1
20100222994	Field	Sep 2010	A1
20100230919	Kawada	Sep 2010	A1
20100235028	Ishii	Sep 2010	A1
20100237645	Trainer	Sep 2010	A1
20100250040	Yamano	Sep 2010	A1
20110035101	Kawada et al.	Feb 2011	A1
20110054717	Yamauchi	Mar 2011	A1
20110106339	Phillips et al.	May 2011	A1
20110123286	Van Roosmalen	May 2011	A1
20110175329	Gingras	Jul 2011	A1
20110209929	Heinzmann	Sep 2011	A1
20110215540	Hunziker et al.	Sep 2011	A1
20110220427	Chen	Sep 2011	A1
20110221160	Shaw	Sep 2011	A1
20110225417	Maharajh et al.	Sep 2011	A1
20110238247	Yen	Sep 2011	A1
20110285195	Ratgen	Nov 2011	A1
20120019554	Narimatu et al.	Jan 2012	A1
20120046821	Pettersson	Feb 2012	A1
20120072052	Powers	Mar 2012	A1
20120168240	Wilson	Jul 2012	A1
20120174037	Relyea et al.	Jul 2012	A1
20120185091	Field	Jul 2012	A1
20120185094	Rosenstein	Jul 2012	A1
20120197470	Inui	Aug 2012	A1
20120205176	Ha	Aug 2012	A1
20120215355	Bewley	Aug 2012	A1
20120219395	Inaguma et al.	Aug 2012	A1
20120239284	Field	Sep 2012	A1
20120290162	Stevens	Nov 2012	A1
20120313335	Zanderlehn	Dec 2012	A1
20130032422	Chen	Feb 2013	A1
20130032423	Chen	Feb 2013	A1
20130080015	Strothmann	Mar 2013	A1
20130081885	Connor	Apr 2013	A1
20130105239	Fung	May 2013	A1
20130146409	Boyle	Jun 2013	A1
20130188809	Jones	Jul 2013	A1
20130218380	Phillips et al.	Aug 2013	A1
20130228385	Chen	Sep 2013	A1
20130231814	Sarokhan	Sep 2013	A1
20130253769	Kamo et al.	Sep 2013	A1
20130332064	Funk	Dec 2013	A1
20140005933	Fong	Jan 2014	A1
20140018994	Panzarella	Jan 2014	A1
20140034400	Underwood	Feb 2014	A1
20140058600	Hoffmann	Feb 2014	A1
20140083225	Downs	Mar 2014	A1
20140088761	Shamlian	Mar 2014	A1
20140187237	Li	Jul 2014	A1
20140202777	Lee	Jul 2014	A1
20140246257	Jacobsen	Sep 2014	A1
20140246258	Wyrobek	Sep 2014	A1
20140277888	Dastoor et al.	Sep 2014	A1
20140371979	Drew	Dec 2014	A1
20150006005	Yu et al.	Jan 2015	A1
20150060162	Goffer	Mar 2015	A1
20150112264	Kamen et al.	Apr 2015	A1
20150119289	Chen	Apr 2015	A1
20150123453	Benoit, Jr.	May 2015	A1
20150012057	Carlson et al.	Jul 2015	A1
20150197247	Ichinokawa	Jul 2015	A1
20150198440	Pearlman et al.	Jul 2015	A1
20150231891	Yashiro et al.	Aug 2015	A1
20150245962	Furuta	Sep 2015	A1
20150246703	Oishi et al.	Sep 2015	A1
20150289653	Hector et al.	Oct 2015	A1
20150342517	Rabischong	Dec 2015	A1
20160014252	Biderman et al.	Jan 2016	A1
20160031497	Luo	Feb 2016	A1
20160035161	Friedli et al.	Feb 2016	A1
20160069691	Fong	Mar 2016	A1
20160075535	Ooms	Mar 2016	A1
20160101685	Darpino et al.	Apr 2016	A1
20160144505	Fong	May 2016	A1
20160170411	Wei	Jun 2016	A1
20160264019	Drako	Sep 2016	A1
20160291848	Hall	Oct 2016	A1
20160362147	Mailey	Dec 2016	A1
20170052033	Fong	Feb 2017	A1
20170080967	Atkins	Mar 2017	A1
20170176188	Georgy et al.	Jun 2017	A1
20170225321	Oeyle	Aug 2017	A1
20170240169	Coulter et al.	Aug 2017	A1
20170243365	Nuijten	Aug 2017	A1
20170259811	Coulter et al.	Sep 2017	A1
20170300058	Peret et al.	Oct 2017	A1
20180024553	Kong et al.	Jan 2018	A1
20180056985	Coulter	Mar 2018	A1
20180102227	Poon	Apr 2018	A1
20180143801	Stucker et al.	May 2018	A1
20180146757	Singh Johar	May 2018	A1
20180164829	Oshima et al.	Jun 2018	A1
20180185212	Lucas	Jul 2018	A1
20180203522	Stucki et al.	Jul 2018	A1
20180253220	Tuhami	Sep 2018	A1
20180329418	Baalke	Nov 2018	A1
20190025853	Julian	Jan 2019	A1
20190033868	Ferguson	Jan 2019	A1
20190038487	Cherny	Feb 2019	A1
20190041219	Schubert	Feb 2019	A1
20190046373	Coulter	Feb 2019	A1
20190087778	Evans, Jr.	Mar 2019	A1
20190114564	Ferguson	Apr 2019	A1
20190224057	Jordan	Jul 2019	A1
20190231617	Cazali	Aug 2019	A1
20190269567	Kao	Sep 2019	A1
20210141377	Brady et al.	May 2021	A1

Foreign Referenced Citations (101)

Number	Date	Country
2822729	Mar 2006	CA
2897221	Mar 2006	CA
2897542	Jan 2016	CA
101056680	Oct 2007	CN
104071275	Oct 2014	CN
2048593	May 1971	DE
3103961	Sep 1982	DE
3128112	Feb 1983	DE
3242880	Jun 1983	DE
3411489	Oct 1984	DE
4110905	Oct 1991	DE
4404594	Aug 1995	DE
19625498	Nov 1997	DE
29808091	Aug 1998	DE
29808096	Aug 1998	DE
10209093	Sep 2003	DE
0109927	May 1984	EP
0193473	Sep 1986	EP
0537698	Apr 1993	EP
0551986	Jul 1993	EP
0663313	Jul 1995	EP
0746089	Dec 1996	EP
0958978	Nov 1999	EP
1063530	Dec 2000	EP
1791609	Sep 2005	EP
1791609	Mar 2006	EP
1759973	Mar 2007	EP
1805071	Jul 2007	EP
980237	May 1951	FR
2502090	Sep 1982	FR
152664	Jan 1922	GB
1213930	Nov 1970	GB
2139576	Nov 1984	GB
2388579	Nov 2003	GB
52-44933	Apr 1977	JP
57-87766	Jan 1982	JP
57-110569	Jul 1982	JP
59-73372	Apr 1984	JP
60-255580	Dec 1985	JP
62-12810	Jan 1987	JP
63-305082	Dec 1988	JP
H01-316810	Dec 1989	JP
2-190277	Jul 1990	JP
4-201793	Jul 1992	JP
5-213240	Aug 1993	JP
6-171562	Dec 1994	JP
61-05415	Dec 1994	JP
7255780	Oct 1995	JP
09-010375	Jan 1997	JP
9-248320	Sep 1997	JP
10-023613	Jan 1998	JP
2000-070308	Jul 2000	JP
2000-288032	Oct 2000	JP
2005-022631	Jan 2005	JP
4572594	Jan 2006	JP
2007-069688	Mar 2007	JP
D1314974	Nov 2007	JP
D1323922	Mar 2008	JP
4687784	Jul 2010	JP
2010-240011	Oct 2010	JP
2010-274759	Dec 2010	JP
2011-246124	Dec 2011	JP
5243795	Jul 2013	JP
2014-019212	Feb 2014	JP
2014-174275	Sep 2014	JP
2014-195403	Oct 2014	JP
2014-204544	Oct 2014	JP
2014-218247	Nov 2014	JP
2015-070897	Apr 2015	JP
2015-171895	Oct 2015	JP
2016-084135	May 2016	JP
2018-062344	Apr 2018	JP
D124943	Jun 2006	TW
WO 198605752	Oct 1986	WO
WO 198906117	Jul 1989	WO
WO 199623478	Aug 1996	WO
WO 199846474	Oct 1998	WO
WO 199911488	Mar 1999	WO
WO 2000023315	Apr 2000	WO
WO 2000054719	Sep 2000	WO
WO 2000054721	Sep 2000	WO
WO 2000075001	Dec 2000	WO
WO 2001002920	Jan 2001	WO
WO 200230730	Apr 2002	WO
WO 2002072383	Sep 2002	WO
WO 2003068342	Aug 2003	WO
WO2003103559	Dec 2003	WO
WO 2003106250	Dec 2003	WO
WO 2004007264	Jan 2004	WO
WO2004078603	Sep 2004	WO
WO 2006031917	Mar 2006	WO
WO 2006042302	Apr 2006	WO
WO 2009052471	Apr 2009	WO
WO 2010084421	Jul 2010	WO
WO 2012090248	Jul 2012	WO
WO 2013096789	Jun 2013	WO
WO 2015167411	Nov 2015	WO
WO 2017147347	Aug 2017	WO
WO 2017156586	Sep 2017	WO
WO 2017180868	Oct 2017	WO
WO 2017201513	Nov 2017	WO

Non-Patent Literature Citations (87)

Entry
U.S. Appl. No. 15/600,703 (U22), B1-B100, B102-B103.
U.S. Appl. No. 15/600,703 (U22), C8-C40, C42-050, C52-C65, C67-C74.
U.S. Appl. No. 15/787,613 (W10) C1-C7, C41, C51, C66.
Derry et al., Automated Doorway Detection for Assistive Shared-Control Wheelchairs, 2013 IEEE International Conference on Robotics and Automation, May 6-10, 2013, https://cpb-us-e1.wpmucdn.com/sites.northwestern.edu/dist/5/1812/files/2016/05/13icra_derry.pdf.
U.S. Appl. No. 16/035,205, filed Jul. 13, 2018.
U.S. Appl. No. 15/600,703, filed May 20, 2017.
PCT/US17/33705, May 20, 2017.
PCT/2018/042114, Jul. 13, 2018.
U.S. Appl. No. 15/486,980, filed Apr. 13, 2017.
PCT/US17/27410, Apr. 13, 2017.
U.S. Appl. No. 15/441,190, filed Feb. 23, 2017.
PCT/US17/19214, Feb. 23, 2017.
U.S. Appl. No. 16/200,088, filed Nov. 26, 2018.
Adhikari, B., A Single Subject Participatory Action Design Method for Powered Wheelchairs Providing Automated Back-in Parking Assistance to Cognitively Impaired Older Adults: A pilot study, Department of Computer Science, The University of British Columbia, Vancouver, Canada, Jan. 5, 2015, slide deck.
Adhikari, B., A Single Subject Participatory Action Design Method for Powered Wheelchairs Providing Automated Back-in Parking Assistance to Cognitively Impaired Older Adults: A pilot study, Master's Thesis, Department of Comptuer Science, The University of British Columbia, Vancouver, Canada, Dec. 2014.
Brown, Jr. et al., “A Single-Wheel, Gyroscopically Stabilized Robot,” IEEE Robotics & Automation Magazine, Sep. 1997.
“BTCR9 Fansyn Bluetooth . . . ” Fanimation, published Feb. 4, 2017 (Retrieved from the Internet Sep. 27, 2019). Internet URL: https://web.archive.org/web/20170204193258/https://www.fanimation.com/products/index.php/controls-remotes/fansync-bluetooth-receiver-transmitter-downlight.html(Year : 2017).
Cho et al, Sloped Terrain Segmentation for Autonomous Drive Using Sparse 3D Point Cloud, The Scientific World Journal, 2014, https://www.hindawi.com/journals/tswj/2014/582753/.
Cooper, Rory A., “Intelligent Control of Power Wheelchairs”, IEEE Engineering in Medicine and Biology Magazine, IEEE Service Center, Piscataway, NJ, US, vol. 14, No. 4, Jul. 1, 1995, pp. 423-431, XP11084628.
Dejun Yin and Yoichi Hori, “A Novel Traction Control for Electric Vehicle without Chassis Velocity, Motion Control”, Federico Casolo (Ed.), InTech, DOI: 10.5772/6962. Available from: https://mts.intechopen.com/books/motion-control/a-novel-traction-control-for-electric-vehide-without-chassis-velocity, 2010.
Derry et al., Automated Doorway Detection for Assistive Shared-Control Wheelchairs, 2013 IEEE International Conference on Robotics and Automation, May 6-10, 2013, https://cpb-us-el.wpmacdn.com/sites.northwestern.edu/dist/5/1812/files/2016/05/13icra_derry.pdf.
Elnagar, A., “Prediction of Moving Objects in Dynamic Environments Using Kalman Filters,” Proceedings of 2001 IEEE International Symposium on Computational Intelligence in Robotics and Automation, Jul. 29-Aug. 1, 2001.
Fresk, et al., “Full Quaternion Based Attitude Control for a Quadrator”, 2013 European Control Conference (EDD), Jul. 17-19, 2013, Zurich, Switzerland, pp. 3864-3869.
Grasser, F. et al., “JOE: A Mobile, Inverted Pendulum,” IEEE Transactions on Industrial Electronics, vol. 49, No. 1, Feb. 2002.
Ha, et al. “Trajectory Tracking Control for Navigation of Self-Contained Mobile Inverse Pendulum” Intelligent Robots and Systems '94. ‘Advanced Robotic Systems and the Real World’, IROS '94. Proceedings of the IEEE/RSJ/GI International Conference on, vol. 3, no., pp. 1875-1882, Sep. 12-16, 1994.
Ha, et al., “Trajectory Tracking Control for Navigation of the Inverse Pendulum Type Self-Contained Mobile Robot” Robotics and Autonomous Systems 17, 65-80 (1996).
Helgesson, L., “Pitch and roll estimating Kalman filter for stabilizing quadrocopters”, http:/lhelge.se/2012/04/pitch-and-roll-estimating-kalman-filter-for-stabilizing-quadrocopters/, Oct. 15, 2012.
How et al., “Clinical Evaluation of the Intelligent Wheelchair System”, Proceedings of Festival of international Conference on Caregiving, Disability, Aging and Technology, Toronto, Canada, 2011.
I-Real, Personal Mobility Device, https://www.youtube.com/warch?v=WAGpxIUpdWw, Published on Jan. 15, 2013, appeared first in Apr. 2012, D1 Grand Prix event, Odaiba, JP.
Ishida and Miyamoto, “Collision-Detecting Device for Omnidirectional Electric Wheelchair”, Research Article, ISRN Robotics, vol. 2013, Article ID 672826, Nov. 1, 2012.
I-swing, Single Person Vehicle, https://www.youtube.com/watch?feature=player_embedded&v=1QSybf7sLtg, Published on Sep. 14, 2006, Featured on Hacked Gadgets, http://hackedgadgets.com.
I-Unit, Wheelchair, https://www.youtube.com/watch?v=RbrrIrh3GBE, Published on June 6, 2006, Filmed at Megaweb Center at Tokyo.
Johnson, R.C., “Unicycles and Bifurcations”, American J. of Physics, vol. 66, No. 7, 589-92 (Oct. 22, 2002).
Kanoh, “Applied Control of Inverted Pendulum”, Computrol, vol. 2, (1983), pp. 69-75.
Kawaji, S., “Stabilization of Unicycle Using Spinning Motion”, Denki Gakkai Ronbushi, D, vol. 107, Issue 1, Japan (1987), pp. 21-28.
Koyanagi et al., “A Wheeled Inverse Pendulum Type Self-Contained Mobile Robot”, The Society of Instrument and Control Engineers, Special issue of the 31st SICE Annual Conference, Japan 1992, pp. 51-56.
Koyanagi et al., “A Wheeled Inverse Pendulum Type Self-Contained Mobile Robot and its Two Dimensional Trajectory Control”, Proceeding of the Second International Symposium on Measurement and Control in Robotics, Japan 1992, pp. 891-897.
Koyanagi et al., “A Wheeled Inverse Pendulum Type Self-Contained Mobile Robot and its Posture Control and Vehicle Control”, The Society of Instrument and Control Engineers, Special issue of the 31st SICE Annual Conference, Japan, 1992, pp. 13-16.
Lam, H. K. et al., “Fuzzy Model Reference Control of Wheeled Mobile Robots,” The 27th Annual Conference of the IEEE Industrial Electronics Society (2001).
Liu, H.S. et al., “Accelerometer for Mobile Robot Positioning,” IEEE Transactions on Industry Applications, vol. No. 3, Oct. 1999.
Meeussen et al., Autonomous Door Opening and Plugging In with a Personal Robot, Willow Garage, USA, IEEE International Conference on Robotics and Automation, May 3-7, 2010, http://www.willowgarage.com/sites/default/files/m2.pdf.
Momoi & Yamafuji, “Motion Control of the Parallel Bicycle-Type Mobile Robot Composed of a Triple Inverted Pendulum”, Paper Read at Meeting of Japan Society of Mechanical Engineering (Series C), vol. 57, No. 541, (Sep. 1991), pp. 154-159.
Montella, C., et al., “To the Bookstore! Autonomous Wheelchair Navigation in an Urban Environment”, Lehigh University, published in FSR, 2012, Part of the Springer Tracts in Advanced Robotics book series (STAR, vol. 92), first online Dec. 31, 2013.
News article, “Amazing Wheelchair Goes Up and Down Stairs”.
Oishi et al., “Building A Smart Wheelchair On A Flexible Software Platform”, RESNA International Conference on Technology and Aging, 2011.
Osaka et al., “Stabilization of unicycle”, Systems and Control, vol. 25, No. 3, Japan Mar. 1981, pp. 159-166.
PCT/US2017/019214, Written Opinion of the International Search Authority, dated Aug. 31, 2017.
PCT/US2017/027410, Written Opinion of the International Search Authority, dated Dec. 4, 2017.
PCT/US2017/033705, Written Opinion of the International Search Authority, dated Nov. 23, 2017.
PCT/US2017/033705, Invitation to pay additional fees and partial search report, Int. App. #PCT/US2017/033705, Intl. filing date May 20, 2017.
Roy et al., “Five-Wheel Unicycle System”, Medical & Biological Engineering & Computing, vol. 23, No. 6, United Kingdom Nov. 1985, pp. 593-596. Entire document can be purchased via: https://link.springer.com/article/10.1007%2F8F02455316.
Sabatini, A, “Quaternion-based Extended Kalman Filter for Determining Orientation by Inertial and Magnetic Sensing”, IEEE Transactions on Biomedical Engineering, vol. 53:7, Jul. 2006, pp. 1346-1356.
Schoonwinkel, A., “Design and Test of a Computer-Stabilized Unicycle”, Stanford University (1988), UMI Dissertation Services, Dissertation Abstracts International, vol. 49/03-B, Stanford University 1987, pp. 890-1294.
Bob_Schor. “Re: Cannot get latch mechanical action on Boolean button . . . ” NI Community, published Jun. 2, 2018 (Retrieved from the Internet Sep. 26, 2019). Internet URL: https://forums.ni.com/t5/LabVIEW/Cannot-get-latch-mechanical-action-on-boolean-button-inside-a/td-p/3799821?profile.language=en (Year: 2018).
Sheng et al., “Postural Stability of a Human Riding a Unicycle and Its Emulation by a Robot,” IEEE Transactions on Robotics and Automation, vol. 13:5, Oct. 1997.
Sheng, Zaiquan; Yamafuji, Kazuo: “Realization of a Human Riding a Unicycle by a Robot”. Proceedings of the 1995 IEEE International Conference on Robotics and Automation, vol. 2, 1995, pp. 1319-1326.
Stew's Hovercraft Page, http://www.stewcam.com/hover-craft.html.
Takahashi et al., “Back and Forward Moving Scheme of Front Wheel Raising for Inverse Pendulum Control Wheel Chair Robot”, Proceedings of the 2001 IEEE International Conference of Robotics & Automation, Seoul, Korea, May 21-26, 2001, pp. 3189-3194.
Takahashi et al., “Front Wheel Raising and Inverse Pendulum Control of Power Assist Wheel Chair Robot”, IEEE, 1999, pp. 668-673.
Tanaka et al., “A Mobile Robot for Service Use: Behaviour Simulation System and Intelligent Control,” Proceedings of the 1997 IEEE/RSJ International Conference on Intelligent Robots and Systems, 1997.
TECKNICO'S Home Page, “Those Amazing Flying Machines”, http://www.swiftsite.com/technico, May 24, 1999.
Ulyanov et al., “Fuzzy Intelligent Emotion and Instinct Control of a Robotic Unicycle,” Proceedings of the 1996 4th International Workshop on Advanced Motion Control, Mar. 18-21, 1996.
Ulyanov et al., “Soft computing for the intelligent robust control of a robotic unicycle with a new physical measure for mechanical controllability”. Soft Computing vol. 2:2, Jun. 1998, pp. 73-88.
Umpad, Leomar. “How Do I Use My Samsung Galaxy Device as a TV Remote Control?” Tech Recipes,published Nov. 27, 2014 (Retrieved from the Internet Sep. 27, 2019). Internet URL: <https://www. Tech-recipes.com/rx/51556/how-do-i-use-my-samsung-galaxy-device-as-a-tv-remote-control/> (Year: 2014).
Viswanathan et al., “Navigation Assistance for Intelligent Wheelchairs”, 3rd International Conference on Technology and Aging/RESNA, Toronto, 2011.
Vos et al., “Dynamics and Nonlinear Adaptive Control of an Autonomous Unicycle—Theory and Experiment”, American Institute of Aeronautics and Astronautics, A90-26772 10-39, Washington, D.C. 1990, Abstract only.
Vos, D., Dynamics and Nonlinear Adaptive Control of an Autonomous Unicycle, Massachusetts Institute of Technology, Jun. 7, 1989.
Vos, D., “Nonlinear Control of an Autonomous Unicycle Robot: Practical Issues”, Massachusetts Institute of Technology, Jun. 5, 1992.
Wang et al., “Real-time Model-based Electrical Powered Wheelchair Control”, Med Eng Phys. Dec. 2009: 31(10): 1244-1254.
Watson Industries, Inc., “Single Axis Vertical Reference System Owner's Manual ADS-C132-1A”, Apr. 20, 2015, pp. 3-4.
Welch et al., “An Introduction to the Kalman Filter,” SIGGRAPH 2001, Department of Computer Science University of North Carolina at Chapel Hill, http://www.cs.unc.edu/˜l_welch.gbl, 2001.
WO 2000/073101, IPER of the International Search Authority, filing date Mar. 14, 2000.
WO 2000/075001, IPER of the International Search Authority, filing date Jun. 1, 2000.
WO2002/030730, IPER of the International Search Authority, filing date Oct. 11, 2001.
WO2004/007264, Initial Publication with ISR, International Publication Date Jan. 22, 2004.
WO 2017/147347 Written Opinion of the International Search Authority, Int. App. #PCT/US2017/019214, priority date Feb. 23, 2016.
WO 2017/201513, Invitation to pay additional fees and partial search report, Int. App. #PCT/US2017/033705, Inti, filing date May 20, 2017.
WO 2017/201513, Written Opinion of the International Searching Authority, Int. App. #PCT/US2017/033705, Intl. filing date May 20, 2017.
Wolstenholme, Kevin. “Updating Glide—The Full Breawkdown.” RisingHigh Academy, published Aug. 26, 2017 (Retrieved from the Internet Sep. 26, 2019). Internet URL: https://risinghighacademy.com/category/games/(Year:2017).
Yamafuji & Kawamura, “Study on the Postural and Driving Control of Coaxial Bicycle”, Paper Read at Meeting of Japan Society of Mechanical Engineering (Series C), vol. 54, No. 501, (May 1988), pp. 1114-1121, Abstract in English.
Yamafuji & Kawamura, “Study of Postural and Driving Control of Coaxial Bicycle”, Papers Read at Meeting of Japan Society of Mechanical Engineering (vol. C), vol. 54, No. 501 (May 1988), Paper No. 87-0901A.
Yamafuji et al., “Synchronization and Steering Control of Parallel Bicycle”, Paper Read at Meeting of Japan Society of Mechanical Engineering (Series C), vol. 55, No. 513, (May 1989), pp. 1229-1234.
Yamafuji, “A Proposal for Modular-Structured Mobile Robots for Work that Principally Involve a Vehicle with Two Parallel Wheels”, Automation Technology, vol. 20, pp. 113-118 (1988).
Yun et al., “Implementation and Experimental Results of a Quarternion-Based Kalman Filter for Human Body Motion Tracking”, Proceedings of the 2005 IEEE International Conference on Robotics and Automation, Barcelona, Spain, Apr. 2005, pp. 317-322.
Yun et al., “Design, Implementation and Experimental Results of a Quarternion-Based Kalman Filter for Human Body Motion Tracking”, IEEE Transactions on Robotics, vol. 22, No. 6, Dec. 2006, pp. 1216-1227.
Zenkov, DV, AM Bloch, and JE Marsden [2001] “The Lyapunov-Malkin Theorem and Stabilization of the Unicycle with Rider”. Systems and Control Letters, vol. 45, No. 4, Apr. 5, 2002, pp. 293-302(10).
Zenkov, DV, AM Bloch, NE Leonard and JE Marsden, “Matching and Stabilization of Low-Dimensional Nonholonomic Systems”. Proc. CPC, 39, (2000), 1289-1295.

Related Publications (1)

	Number	Date	Country
	20190377349 A1	Dec 2019	US

Provisional Applications (1)

	Number	Date	Country
	62682129	Jun 2018	US

System and method for distributed utility service execution

Information

Patent Number

Date Filed

Date Issued

Inventors

Original Assignees

Examiners

Agents

CPC

Field of Search

CPC

International Classifications

Term Extension

Abstract