Probabilistic decision support for obstacle detection and classification in a working area

Information

  • Patent Grant
  • 11320830
  • Patent Number
    11,320,830
  • Date Filed
    Monday, October 28, 2019
    4 years ago
  • Date Issued
    Tuesday, May 3, 2022
    2 years ago
Abstract
Systems and methods disclosed herein provide probabilistic decision support regarding detected obstacles in a working area. Real-time data sets are collected from obstacle sensors associated with at least one self-propelled work vehicle, corresponding to detected presence/absence of obstacles at given locations within the working area. The received real-time data sets are integrated in data storage comprising a priori data sets corresponding to the working area, to generate one or more new a priori data sets. Probabilities are determined for the detected presence or absence of the obstacle, and for each of one or more obstacle categories, based on the received real-time data set and at least an a priori data set corresponding to the work vehicle's location. An output corresponding to at least a most likely of the determined probabilities is generated as feedback to a user interface, and/or relevant machine control units.
Description
FIELD OF THE DISCLOSURE

The present disclosure relates to self-propelled vehicles such as working machines having onboard obstacle detection systems, and more particularly systems and methods for supplementing conventional obstacle detection with probabilistic decision support based on a priori knowledge development and estimation.


BACKGROUND

Working machines as discussed herein may for example include excavator machines, loaders, and other equipment which modify the terrain or equivalent working environment in some way. These machines may have tracked or wheeled ground engaging units supporting the undercarriage from the ground surface, and may further include one or more working implements which are used to modify the terrain in coordination with movement of the machine. Various situations arise with such machines where the human operator needs to control the movements of the machine position and/or the relative positions of the working implements based on detected obstacles in the working area, which may or may not be visible to the operator.


Obstacle detection systems are conventionally known to provide feedback to the operator regarding detected obstacles in the working area. When creating an obstacle detection system, it is advantageous to have a system that can distinguish between different types of objects (e.g., moving or transient obstacles such as people or vehicles, or fixed obstacles such as structures or material), since the system may want to react differently to different types of objects in the environment. Accordingly, it is also desirable to accurately detect these different object classes, such that the system does not detect an object which is in fact not present, or fail to detect a present object, or still further assign an incorrect classification to an object which results in an improper response by the human operator or automated machine control system.


SUMMARY OF THE DISCLOSURE

The current disclosure provides improvements in the collection and implementation of data corresponding to working machine actions over time to improve obstacle detection and classification results.


In an embodiment, a method of probabilistic decision support regarding detected obstacles in a working area comprises: providing data storage comprising one or more a priori data sets corresponding to a working area; receiving real-time data sets from one or more obstacle sensors associated with at least one self-propelled work vehicle, the real-time data sets corresponding to a detected presence or absence of an obstacle at a given location within the working area; determining probabilities for the detected presence or absence of the obstacle, and for each of one or more obstacle categories, based on the received real-time data set and an a priori data set corresponding to the given location as a subset of the one or more a priori data sets corresponding to the working area, and generating an output corresponding to at least a most likely of the determined probabilities.


Another embodiment of the method may further comprise integrating the received real-time data sets in data storage comprising a priori data sets corresponding to the working area, to generate one or more new a priori data sets.


Another embodiment of the method may further comprise monitoring movements of one or more moveable work vehicle components comprising at least one of the work implements, and updating one or more of the one or more a priori data sets with data categorizing an obstacle at a location within the working area based on a predetermined sequence of movement for the one or more work vehicle components.


In another exemplary embodiment of the method, for updating the one or more of the one or more a priori data sets, a first obstacle is categorized based on a first predetermined sequence of movement, a second obstacle is categorized based on a second predetermined sequence of movement, and a work vehicle route is further categorized between the first and second obstacles. Based on a specified work task, one or more of the first obstacle and the second obstacle may be categorized as either a transient obstacle or a permanent obstacle.


Another embodiment of the method may further comprise predicting a work vehicle route based on monitored vehicle movement. An obstacle detection area may further be defined as a subset of the work area corresponding to the predicted work vehicle route. A priori data retrieval and processing may be dynamically limited with respect to a subset of the data storage based on the defined obstacle detection area.


Another embodiment of the method may further comprise, for predicted and previously traversed work vehicle routes lacking a categorized permanent obstacle, restricting subsequent determination of probabilities of the detected presence or absence of the obstacle only to transient obstacles.


In an embodiment, the output may be provided to a user interface located in or on the work vehicle; and/or a user computing device. One or more predicted obstacles and associated categories on the user interface may in some embodiments be highlighted based on the output. One or more decision support tools or user-manipulable controls may in some embodiments be highlighted based on the output, for prompting or facilitating manual regulation of work vehicle components as for example may comprise or otherwise be integrated with an obstacle collision avoidance or negotiation system.


In an embodiment, the output may include at least one control signal provided to a vehicle steering control unit, a vehicle implement positioning control unit, and/or a vehicle drive control unit, for automated regulation of associated work vehicle components as for example may comprise or otherwise be integrated with an obstacle collision avoidance or negotiation system.


Manual and automatic working machine control based on detected obstacles as discussed above is known in the art, and conventional control structure and methodology may be implemented but is not necessarily further described in the context of a present invention. A system and method as disclosed herein supplements, complements, or otherwise provides additional and valuable feedback as part of an obstacle detection system, whether for operator decision support or automated obstacle negotiation, or in other words regardless of the control response unless otherwise stated. The a priori data generation and implementation enables an obstacle detection system to more confidently determine probabilities corresponding for example to whether an obstacle is actually present, and/or what type of obstacle is most likely to be present where applicable, such that a response may be tailored appropriately and not entirely subjected as in the prior art to the uncertainties inherent to the particular obstacle sensing apparatus and operating conditions.


In an embodiment, any one of the aforementioned methods may be performed via a self-propelled work vehicle having one or more work implements for working a terrain, one or more obstacle sensors, data storage comprising one or more a priori data sets corresponding to a working area, and a controller communicatively linked to the one or more work implements, the one or more obstacle sensors, and the data storage.


In an embodiment, any one of the aforementioned methods may be performed via a system comprising a server network communicatively linked to a plurality of self-propelled vehicles, each of the self-propelled vehicles comprising at least one work implement for working a terrain and one or more obstacle sensors, and data storage comprising one or more a priori data sets corresponding to a working area, wherein for example the respective method is executed with respect to each of the self-propelled vehicles located in the working area.


Numerous objects, features and advantages of the embodiments set forth herein will be readily apparent to those skilled in the art upon reading of the following disclosure when taken in conjunction with the accompanying drawings.





BRIEF DESCRIPTION OF THE DRAWINGS


FIG. 1 is a side elevation view of a tracked excavator as a working machine incorporating an embodiment of a system as disclosed herein.



FIG. 2 is a block diagram representing an exemplary control system for the working machine of FIG. 1.



FIG. 3 is an overhead schematic view of an exemplary working area including obstacles and a working machine.



FIG. 4 is a graphical diagram representing an embodiment of a distributed system as disclosed herein.



FIG. 5 is a flowchart representing an exemplary embodiment of a method as disclosed herein.





DETAILED DESCRIPTION

Referring now to the drawings and particularly to FIG. 1, a representative work vehicle is shown and generally designated by the number 100. FIG. 1 shows a tracked excavator machine 100. The systems disclosed herein are applicable to similar or otherwise equivalent vehicles, excavator machines, loaders, and other working machines of the type having one or more working implements 102 for modifying the proximate terrain. In certain embodiments, systems and methods as disclosed herein may also be applicable to vehicles lacking explicit work implements. Work vehicles as discussed herein may typically have tracked or wheeled ground engaging units supporting the undercarriage from the ground surface.


The illustrated and exemplary working machine 100 includes an undercarriage 122 including one or more tracked or wheeled ground engaging units 124 and at least one travel motor for driving the respective ground engaging units. A main frame 132 is supported by the undercarriage 122 via a swing bearing 134 such that the main frame 132 is pivotable about a pivot axis 136 relative to the undercarriage. The pivot axis 136 is substantially vertical when a ground surface 138 engaged by the ground engaging units 124 is substantially horizontal. A swing motor is configured to pivot the main frame 132 on the swing bearing 134 about the pivot axis 136 relative to the undercarriage 122.


A working machine as disclosed herein typically includes one or more working implements 102 which as illustrated for example in FIG. 1 may collectively define a boom assembly including a boom 144, an arm 146 pivotally connected to the boom 144, and a working tool 148. The boom 144 in the present example is pivotally attached to the main frame 132 to pivot about a generally horizontal axis relative to the main frame 132. The working tool in this embodiment is an excavator shovel or bucket 148 which is pivotally connected to the arm 146. The boom assembly 102 extends from the main frame 132 along a working direction of the boom assembly. The working direction can also be described as a working direction of the boom.


An operator's cab may be located on the main frame 132. The operator's cab and the one or more working implements 102 may both be mounted on the main frame so that the operator's cab faces in the working direction of the working implements. A control station including a user interface 114 may be located in the operator's cab.


As schematically illustrated in FIG. 2, the working machine 100 includes a control system including a controller 112. The controller may be part of the machine control system of the working machine, or it may be a separate control module. The controller may include the user interface 114 and optionally be mounted in the operators cab at a control panel.


The controller 112 is configured to receive input signals from some or all of various sensors collectively defining a sensor system 104. Certain of these sensors may be provided to detect machine operating conditions or positioning, including for example an orientation sensor, global positioning system (GPS) sensors, vehicle speed sensors, vehicle implement positioning sensors, and the like, and whereas one or more of these sensors may be discrete in nature the sensor system may further refer to signals provided from the machine control system.


Other sensors in the sensor system 104 more particularly refer to a group of sensors referred to herein as obstacle sensors 104. Various examples of obstacle sensors are conventionally known and may include laser scanners, thermal sensors, imaging devices, structured light sensors, ultrasonic sensors, and other optical sensors. The types and combinations of obstacle sensors may vary for a type of working machine, work area, and/or application, but generally are provided and configured to optimize recognition of obstacles in a working path of the vehicle.


The controller 112 may typically produce an output regarding obstacle detection and classification, as further described below, to the user interface 114 for display to the human operator. The controller may further or in the alternative generate control signals for controlling the operation of respective actuators, or signals for indirect control via intermediate control units, associated with a machine steering control system 126, a machine implement control system 128, and a machine drive control system 130.


The controller 112 includes or may be associated with a processor 150, a computer readable medium 152, a communication unit 154, data storage 156 such as for example a database network, and the aforementioned user interface 114 or control panel 114 having a display 118. An input/output device 116, such as a keyboard, joystick or other user interface tool, is provided so that the human operator may input instructions to the controller. It is understood that the controller described herein may be a single controller having all of the described functionality, or it may include multiple controllers wherein the described functionality is distributed among the multiple controllers.


Various operations, steps or algorithms as described in connection with the controller 112 can be embodied directly in hardware, in a computer program product such as a software module executed by the processor 150, or in a combination of the two. The computer program product can reside in RAM memory, flash memory, ROM memory, EPROM memory, EEPROM memory, registers, hard disk, a removable disk, or any other form of computer-readable medium 152 known in the art. An exemplary computer-readable medium can be coupled to the processor such that the processor can read information from, and write information to, the memory/storage medium. In the alternative, the medium can be integral to the processor. The processor and the medium can reside in an application specific integrated circuit (ASIC). The ASIC can reside in a user terminal. In the alternative, the processor and the medium can reside as discrete components in a user terminal.


The term “processor” as used herein may refer to at least general-purpose or specific-purpose processing devices and/or logic as may be understood by one of skill in the art, including but not limited to a microprocessor, a microcontroller, a state machine, and the like. A processor can also be implemented as a combination of computing devices, e.g., a combination of a DSP and a microprocessor, a plurality of microprocessors, one or more microprocessors in conjunction with a DSP core, or any other such configuration.


The communication unit 154 may support or provide communications between the controller and external systems or devices, and/or support or provide communication interface with respect to internal components of the working machine. The communications unit may include wireless communication system components (e.g., via cellular modem, WiFi, Bluetooth or the like) and/or may include one or more wired communications terminals such as universal serial bus ports.


The data storage 156 as further described below may be configured to at least receive and store real-time data sets 158 and a prior data sets 160 in selectively retrievable form. Data storage as discussed herein may, unless otherwise stated, generally encompass hardware such as volatile or non-volatile storage devices, drives, memory, or other storage media, as well as one or more databases residing thereon.


Referring generally to FIGS. 1-4, exemplary modes of operation may further be described with respect to the aforementioned working machine. As previously noted, obstacle detection sensors, algorithms, and controls are conventionally known, but it can often be difficult to reliably distinguish an appropriate obstacle detection signal from the noise characteristic of typical working operations. An optimal obstacle detection system would not only be able to accurately detect the presence or absence of an object in the (actual or projected) working path of the machine, but also to distinguish between different types of objects (e.g., people, vehicles, structures) and provide appropriate feedback.


Embodiments of a decision support system and method as disclosed herein may more accurately detect the presence of an object and accurately assign it a category or classification through the use of “priors” or a priori knowledge development. An a priori data set in this context may refer generally to information the system has collected, generated, and/or derived, prior to performing the classification, regarding the probability that an object of a certain class is present in a certain location. It may be appreciated that improving the accuracy of these priors over time can also significantly improve the performance of the classification.


A priori data sets may further be generated as corresponding to a working area, or operating environment, for the working machines. A working area in one example may be predetermined, such as for example via defined boundaries, configurations, and the like. Alternatively, a working area may be defined generally with respect to a specific working machine location, such as an area surrounding and within a certain distance from the working machine. In certain embodiments, the system may accordingly generate a priori data sets with respect to a specific work area that may be traversed by multiple working machines, or a priori data sets may be generated with respect to a given fixed location that may be later classified as a work area for a given working machine based on its detected position.


As noted above, working machines may typically modify their environment in some way, for example in that loaders may move piles of material around or excavators may dig a hole in one location and pile dirt in another. An obstacle detection system as disclosed herein is directed in part to detect and classify these objects as holes or piles, not only for safety reasons but also for improved job automation in that it is important to provide feedback regarding where the machine is to load and unload material.


In certain embodiments as disclosed herein, a system and method may accordingly involve analyzing the motions of the working machine (e.g., the working implements) and inferring therefrom the presence of holes and piles that the machine is creating. This information is further implemented to establish or integrate with previously established a priori knowledge and create better priors for the classifier to detect and classify these elements correctly in the environment.


As one example, the system may be configured to remember places where a wheel loader boom is raised and the bucket drops, further inferring that the operator is dumping material in this location, and accordingly increasing the prior for presence of a “Material Pile” in that location. The pile may be designated as a certain type of location or fixed “destination” associated with later iterations of a predicted working machine route. Alternatively, the system may have prior knowledge of a machine project or job which includes filling a dump truck, wherein the system is configured to increase the prior for “Vehicle” in this particular location. The vehicle may be assigned a certain type of location or transient “destination” in this context with respect to later iterations of a predicted working machine route.


As another example associated with a wheeled or tracked excavator as the working machine, the system may be configured to monitor and identify positions where the boom is down and the bucket is lower than the tracks, and accordingly to increase the prior for “Hole” in that location. The hole may be designated as for example another certain type of location or fixed “origin” associated with later iterations of a predicted working machine route.


One of skill in the art may appreciate that various additional or alternative combinations of machine and/or machine implement movements may be predetermined in association with relevant designations and implemented by the system for improving priors and associated classifications, depending for example on the type of working machine, the type of working area, the type of project, etc.


In addition to monitoring and analyzing machine implement motions to improve priors, the system may further (or in the alternative) consider the driving paths taken by the working machine to establish or further develop a priori knowledge for a “Driveability” class. In one example associated with a loader, the working machine may perform a series of movements such as driving into a pile, picking up material, backing up and then, while turning, driving forward towards a dump truck, unloading, reversing while turning, and then repeating the same. The system as disclosed herein can develop a priori knowledge regarding the working area by monitoring the motions of the working machine for a short period of time.


For example, the system may determine that the areas the working machine has driven over repeatedly are of the class “Driveable” and that the system may permissibly restrict subsequent analysis to only those objects that could move into those spaces (e.g., “transient” obstacles such as people and vehicles, rather than “fixed” obstacles such as buildings or equivalent structures).


Further, the system may be configured to accurately predict a working machine route or path, based on the knowledge that the working machine recently and successfully drove along that path or otherwise through a series of locations. The system may accordingly determine a most likely portion (subset) of the working area to be traversed by the working machine and restrict the focus of the obstacle detection system generally with respect to that subset of the working area, or even more particularly and sequentially with respect to specific locations along the predicted route, rather than the larger working area as a whole. This may in many cases substantially reduce the search space and associated data processing requirements of the system, thereby allowing the system to run much more quickly and accurately.


Referring for example to FIG. 3, a system as disclosed herein may be configured to detect a first obstacle 302 in a working area 300 and classify the obstacle based on a first combination of one or more monitored working machine movements as a material loading “origin,” and further detect a second obstacle 304 in the working area and classify the obstacle based on a second combination of one or more monitored working machine movements as a material unloading “destination.” The system may further define a working machine route 306 between the first and second obstacles based on a monitored and uneventful traverse of said route. Accordingly, in an embodiment of a method as disclosed herein, the system may (depending for example on a particular job or other condition) predict that the working machine will proceed along the defined route and restrict obstacle detection data processing, and the associated a priori data retrieval from storage, to only a portion of the overall work area that corresponds to the predicted route. This may for example exempt previously detected and classified “permanent” (or otherwise e.g. “highly likely” to be permanent) obstacles 308 that do not correspond to or overlap with the predicted route. This may further assist the system in classifying an object 310 that is detected along the predicted route as being a “transient” (or “highly likely” to be transient) obstacle (e.g., human or vehicle), and further enabling optimized feedback and/or control response based thereon.


In some embodiments an obstacle detection method including a prior data set generation, selective data retrieval and processing, and feedback or control response may be implemented from a controller or machine control system associated with an individual working machine.


In other embodiments as illustrated for example in FIG. 4, a distributed obstacle detection system 400 may include one or more remote servers 402 linked to working machines 100a, 100b, 100c within or associated with a work area 300. The server may be configured for bidirectional communications with the working machines via a communications network 406 to receive obstacle detection data sets from the machines, and further to communicate with data storage 404 for storing and selectively retrieving a priori data based at least in part on the raw data from the machines.


The term “communications network” as used herein with respect to data communication between two or more entities (e.g., server, controller, user computing device) or otherwise between communications network interfaces associated with two or more entities may refer to any one of, or a combination of any two or more of, telecommunications networks (whether wired, wireless, cellular or the like), a global network such as the Internet, local networks, network links, Internet Service Providers (ISP's), and intermediate communication interfaces.


The distributed system as disclosed herein may thereby generate a priori data sets for collective data processing and feedback based on inputs from a number of working machines assigned to the same work area. In alternative embodiments of this structure, various data processing steps may optionally be designed for execution at the server level, at the local working machine controller level, or a combination of the two.



FIG. 5 represents an exemplary embodiment of a method of operation 500 as disclosed herein. Unless otherwise expressly noted, various steps of the method may be performed at the level of a local working machine controller, at the level of a computing device associated with an operator of a working machine or other user, and/or at the level of one or more remote servers communicatively linked thereto.


The method 500 may begin in step 502 with the collection of real-time data sets from one or more obstacle sensors associated with a working machine. The collected data sets may be stored in data storage, for example in predetermined database locations for selective data retrieval and processing in later steps.


As previously noted, the “obstacle sensors” available for system implementation are not inherently limited in scope and may encompass various conventionally known sensors or combinations of sensors for their intended purpose, or as otherwise may be understood by one of skill in the art to be applicable for implementation within the present disclosure.


The “real-time” data sets as referred to herein are also not inherently limited in scope unless otherwise stated, and may encompass raw data signals that are transmitted directly from the sensors in real-time for generation of data sets at the controller level, sets of raw data signals that are collectively transmitted from the sensors to the controller after predetermined periods of time rather than instantaneously, event-driven data signals that are transmitted from the sensors to the controller based on threshold or non-threshold detection events, and the like.


The method 500 may continue in step 504 by integrating the collected obstacle detection data sets to improve upon any known priors, thereby generating new a priori data sets in data storage.


The method 500 may optionally continue in steps 506 and 508 by monitoring movements of a working machine over time, predicting a path or route to be traversed by the working machine, and further defining an obstacle detection area for corresponding a priori data retrieval and processing. As previously noted, certain embodiments including such steps may substantially decrease the processing time required by the method, by requiring only a limited subset of the overall work area to be substantively reviewed at a given time.


The method 500 may continue in step 510 by determining a probability of an actual presence (or absence) of an obstacle at a given location, and in step 512 by determining the respective probabilities for categories of said obstacle. As one example, data signals received from a conventional obstacle sensor (or combination of obstacle sensors) may indicate that an obstacle has been detected. A priori data retrieval and processing, further involving for example integration of the new data signals, may be used to determine a likelihood that an obstacle is actually present. In contexts where the raw data signals may be relatively noisy in nature, implementation of priors may confirm that an obstacle is indeed present, or instead may prevent (or at least dampen or delay) inappropriate feedback or control response by determining that such an obstacle is relatively unlikely.


Such steps may for example include generating state vectors corresponding to a condition or category (e.g., the presence or absence of an obstacle), wherein transition functions or the like may be used to generate and update over time the probability values associated with the respective vectors. With repeated traverse of a given working machine route, the system and method may strengthen the priors associated with no fixed obstacle along the given route, without expressly eliminating for example the possibility of a transient obstacle entering the route. Accordingly, the probability values associated with the state vectors may reflect these relative distinctions, and if an obstacle is later indicated via signals from the obstacle detection sensors the system may infer a low probability that a fixed obstacle is present, but that there may be a possibility of a transient obstacle along the route.


Generally speaking, a priori knowledge development and implementation as discussed above may be primarily based on detected movements of the working vehicle and/or associated work implements, and therefore not relying on information from the obstacle sensors themselves to classify the probabilities that for example an obstacle of a certain type is present. However, data from the obstacle detection sensors may still be integrated in various embodiments and/or machine learning contexts, such as for example where the system may recognize false positive indications over time from a particular obstacle sensor at a particular location. The system may further learn over time to treat signals from certain obstacle detection sensors as being more or less reliable in context, such as for example at particular locations in the work area, or in proximity to a previously traversed vehicle route, etc.


In certain embodiments, the a priori data sets corresponding to a work area may be generated solely or primarily based on learned information over time with respect to detected movements of the work vehicles and/or implements.


Alternatively, the a priori data sets may integrate predetermined information regarding the work area such as for example mapped topographical information, or such predetermined information may be separately provided in data storage for integration by the machine controller or server with respect to the obstacle detection sensor data and the a priori data.


The method 500 may continue in step 514 by generating one or more outputs to a user interface (e.g., control panel within the operator cab, mobile computing device, remote display) and/or to control units associated with the working machine. The outputs may be associated with a highest probability with respect to state vectors in a given relationship, such as for example whether an obstacle is present or absent (or if the system is uncertain regarding the same) in close proximity to the working machine, or whether a detected obstacle is fixed or transient, human or machine, etc.


In one context, the outputs are for display purposes only, and are presented to the human operator as decision support for a manual control response based on a more accurate representation of potential obstacles in the path of the working machine.


In another context, the outputs may be provided for automated control response, and more particularly to enable, disable, dampen, etc., a control response that may otherwise be automatically provided based on the raw obstacle detection signals. For example, control signals or commands may be generated for regulating obstacle avoidance to a vehicle steering system, a vehicle drive system (e.g., for braking and/or propulsion), or the like.


Uncertainty in the probability values generated by the system may be reflected in the outputs as well, such as for example where the outputs are weighted as part of a separate algorithm for control purposes, or where the operator may be presented with displayed results that indicate a potential obstacle but reflect the relative uncertainty or lack of confidence. For example, a certain output may be provided to regulate control of the work vehicle if the system determines that an obstacle is highly likely to be present in the vehicle's path, whereas a separate output may be provided if the system is uncertain regarding the state of a detected obstacle, such that a different control response is prudent or otherwise required by the work vehicle.


Thus it is seen that the apparatus and methods of the present disclosure readily achieve the ends and advantages mentioned as well as those inherent therein. While certain preferred embodiments of the disclosure have been illustrated and described for present purposes, numerous changes in the arrangement and construction of parts and steps may be made by those skilled in the art, which changes are encompassed within the scope and spirit of the present disclosure as defined by the appended claims. Each disclosed feature or embodiment may be combined with any of the other disclosed features or embodiments.

Claims
  • 1. A self-propelled work vehicle having one or more work implements for working a terrain, the work vehicle comprising: one or more obstacle sensors;data storage comprising one or more a priori data sets corresponding to a working area; anda controller communicatively linked to the one or more work implements, the one or more obstacle sensors, and the data storage, and configured to monitor movements of one or more moveable work vehicle components comprising at least one of the work implements,update one or more of the one or more a priori data sets with data categorizing an obstacle at a location within the working area based on a predetermined sequence of movement for the one or more work vehicle components,receive a real-time data set from the one or more obstacle sensors corresponding to a detected presence or absence of an obstacle at a given location within the working area,determine probabilities of an actual presence or absence of the obstacle, and for each of one or more obstacle categories, based on the received real-time data set and an a priori data set corresponding to the given location as a subset of the one or more a priori data sets corresponding to the working area, andgenerate an output corresponding to at least a most likely of the determined probabilities.
  • 2. The work vehicle of claim 1, wherein the controller is further configured to integrate the real-time data set or data derived therefrom into the data storage to generate a new one or more a priori data sets.
  • 3. The work vehicle of claim 1, wherein for updating the one or more of the one or more a priori data sets: a first obstacle is categorized based on a first predetermined sequence of movement,a second obstacle is categorized based on a second predetermined sequence of movement, anda work vehicle route is further categorized between the first and second obstacles.
  • 4. The work vehicle of claim 3, wherein the controller is further configured based on a specified work task to categorize one or more of the first obstacle and the second obstacle as either a transient obstacle or a permanent obstacle.
  • 5. The work vehicle of claim 4, wherein the controller is configured to predict a work vehicle route based on monitored vehicle movement.
  • 6. The work vehicle of claim 5, wherein the controller is configured to define an obstacle detection area as a subset of the work area corresponding to the predicted work vehicle route, and to dynamically limit a priori data retrieval and processing with respect to a subset of the data storage based on the defined obstacle detection area.
  • 7. The work vehicle of claim 5, wherein the controller is configured, for predicted and previously traversed work vehicle routes lacking a categorized permanent obstacle, to restrict subsequent determination of probabilities of the detected presence or absence of the obstacle only to transient obstacles.
  • 8. The work vehicle of claim 1, wherein the output is provided to a user interface located in or on the work vehicle.
  • 9. The work vehicle of claim 8, wherein the user interface is configured to highlight one or more predicted obstacles and associated categories based on the output from the controller.
  • 10. The work vehicle of claim 1, wherein the output includes at least one control signal provided to one or more of a vehicle steering control unit, a vehicle implement positioning control unit, and a vehicle drive control unit, for regulation of associated work vehicle components, and the one or more of the vehicle steering control unit, the vehicle implement positioning control unit, and the vehicle drive control unit are configured to regulate the associated work vehicle components based at least in part on the at least one control signal.
  • 11. A system comprising: a server network communicatively linked to a plurality of self-propelled vehicles, each of the self-propelled vehicles comprising at least one work implement for working a terrain and one or more obstacle sensors; anddata storage comprising one or more a priori data sets corresponding to a working area;wherein the server is configured to monitor movements, for each of the work vehicles, of one or more moveable work vehicle components comprising at least one of the work implements,update one or more of the one or more a priori data sets with data categorizing an obstacle at a location within the working area based on a predetermined sequence of movement for the one or more work vehicle components,receive real-time data sets from the plurality of self-propelled work vehicles, the real-time data sets corresponding to a detected presence or absence of an obstacle at a given location within the working area,determine probabilities for an actual presence or absence of the obstacle, and for each of one or more obstacle categories, based on the received real-time data set and an a priori data set corresponding to the given location as a subset of the one or more a priori data sets corresponding to the working area, andgenerate an output to the respective work vehicle corresponding to at least a most likely of the determined probabilities.
  • 12. The system of claim 11, wherein for updating the one or more of the one or more a priori data sets: a first obstacle is categorized based on a first predetermined sequence of movement,a second obstacle is categorized based on a second predetermined sequence of movement, anda work vehicle route is further categorized between the first and second obstacles, andwherein the server is further configured based on a specified work task for the respective work vehicle to categorize one or more of the first obstacle and the second obstacle as either a transient obstacle or a permanent obstacle.
  • 13. The system of claim 12, wherein the server is configured to predict a work vehicle route based on monitored vehicle movement, anddefine an obstacle detection area as a subset of the work area corresponding to the predicted work vehicle route, and to dynamically limit a priori data retrieval and processing with respect to a subset of the data storage based on the defined obstacle detection area.
  • 14. The system of claim 13, wherein the server is configured, for predicted and previously traversed work vehicle routes lacking a categorized permanent obstacle, to restrict subsequent determination of probabilities of the detected presence or absence of the obstacle only to transient obstacles.
  • 15. A method of probabilistic decision support regarding detected obstacles in a working area, the method comprising: receiving real-time data sets from one or more obstacle sensors associated with at least one self-propelled work vehicle, the real-time data sets corresponding to a detected presence or absence of an obstacle at a given location within the working area;integrating the received real-time data sets in data storage comprising a priori data sets corresponding to the working area, to generate one or more new a priori data sets;determining probabilities for the detected presence or absence of the obstacle, and for each of one or more obstacle categories, based on the received real-time data set and an a priori data set corresponding to the given location as a subset of the one or more a priori data sets corresponding to the working area;generating an output corresponding to at least a most likely of the determined probabilities;monitoring movements of one or more moveable work vehicle components comprising at least one work implement of the work vehicle; andupdating one or more of the one or more a priori data sets with data categorizing an obstacle at a location within the working area based on a predetermined sequence of movement for the one or more work vehicle components.
  • 16. The method of claim 15, wherein for updating the one or more of the one or more a priori data sets: a first obstacle is categorized based on a first predetermined sequence of movement,a second obstacle is categorized based on a second predetermined sequence of movement, anda work vehicle route is further categorized between the first and second obstacles; andwherein based on a specified work task one or more of the first obstacle and the second obstacle are categorized as either a transient obstacle or a permanent obstacle.
  • 17. The method of claim 16, further comprising: predicting a work vehicle route based on monitored vehicle movement;defining an obstacle detection area as a subset of the work area corresponding to the predicted work vehicle route; anddynamically limiting a priori data retrieval and processing with respect to a subset of the data storage based on the defined obstacle detection area.
  • 18. The method of claim 17, further comprising: for predicted and previously traversed work vehicle routes lacking a categorized permanent obstacle, restricting subsequent determination of probabilities of the detected presence or absence of the obstacle only to transient obstacles.
  • 19. The method of claim 15, wherein the output is provided to a user interface located in or on the work vehicle, the method further comprising: highlighting one or more predicted obstacles and associated categories on the user interface based on the output.
  • 20. The method of claim 15, wherein the output includes at least one control signal provided to one or more of a vehicle steering control unit, a vehicle implement positioning control unit, and a vehicle drive control unit, for regulation of associated work vehicle components, the method further comprising: regulating, via the one or more of the vehicle steering control unit, the vehicle implement positioning control unit, and the vehicle drive control unit, the associated work vehicle components based at least in part on the at least one control signal.
US Referenced Citations (153)
Number Name Date Kind
6247538 Takeda et al. Jun 2001 B1
6292725 Kageyama et al. Sep 2001 B1
6321147 Takeda et al. Nov 2001 B1
6324462 Kageyama Nov 2001 B1
6487303 Yamaguchi et al. Nov 2002 B1
6539294 Kageyama Mar 2003 B1
6546115 Ito et al. Apr 2003 B1
8331195 Schumer et al. Dec 2012 B1
9760806 Ning et al. Sep 2017 B1
10037033 Yako et al. Jul 2018 B2
10073462 Debreczeni Sep 2018 B2
10093289 Lee Oct 2018 B2
10108867 Vallespi-Gonzalez et al. Oct 2018 B1
10144353 Karner et al. Dec 2018 B2
10147320 Ellis Dec 2018 B1
10150414 Myers et al. Dec 2018 B2
10196058 Paris et al. Feb 2019 B2
10202115 Stenneth Feb 2019 B2
10205457 Josefsberg et al. Feb 2019 B1
10239446 May et al. Mar 2019 B2
10262528 Fields et al. Apr 2019 B2
20020044082 Woodington et al. Apr 2002 A1
20020152079 Furuta Oct 2002 A1
20040217851 Reinhart Nov 2004 A1
20040252862 Camus et al. Dec 2004 A1
20050154503 Jacobs Jul 2005 A1
20050182540 Sugiura et al. Aug 2005 A1
20050231339 Kudo Oct 2005 A1
20070112513 Mathevon et al. May 2007 A1
20070219720 Trepagnier et al. Sep 2007 A1
20080046150 Breed Feb 2008 A1
20090088979 Koch Apr 2009 A1
20090187321 Otsuka et al. Jul 2009 A1
20090229150 Pfohl et al. Sep 2009 A1
20090259399 Kotejoshyer et al. Oct 2009 A1
20090259400 Coats et al. Oct 2009 A1
20090259401 Edwards et al. Oct 2009 A1
20090268946 Zhang et al. Oct 2009 A1
20100100324 Caminiti et al. Apr 2010 A1
20100253489 Cui et al. Oct 2010 A1
20100289632 Seder et al. Nov 2010 A1
20110115615 Luo et al. May 2011 A1
20110153169 Peterson Jun 2011 A1
20110196568 Nickolaou et al. Aug 2011 A1
20110254700 Gharsalli et al. Oct 2011 A1
20120050024 Taylor et al. Mar 2012 A1
20120130588 Hukkeri May 2012 A1
20120130598 Hukkeri et al. May 2012 A1
20120206336 Bruder et al. Aug 2012 A1
20120218412 Dellantoni et al. Aug 2012 A1
20120249342 Koehrsen et al. Oct 2012 A1
20120271504 Reiners et al. Oct 2012 A1
20120287277 Koehrsen et al. Nov 2012 A1
20120326917 Kiehne Dec 2012 A1
20120327261 Tafazoli Bilandi et al. Dec 2012 A1
20130141581 Mitsuta et al. Jun 2013 A1
20130147958 Mitsuta et al. Jun 2013 A1
20130222573 Onuma et al. Aug 2013 A1
20140000974 Nilsson et al. Jan 2014 A1
20140035775 Zeng Feb 2014 A1
20140118533 Chang May 2014 A1
20140203959 Kriel et al. Jul 2014 A1
20140205139 Kriel et al. Jul 2014 A1
20140236477 Chen et al. Aug 2014 A1
20140266859 Kurihara et al. Sep 2014 A1
20140347483 Nakanishi et al. Nov 2014 A1
20140362220 Izumikawa et al. Dec 2014 A1
20140375503 Kurihara et al. Dec 2014 A1
20150037762 Martin et al. Feb 2015 A1
20150049911 Doettling et al. Feb 2015 A1
20150070498 Kriel Mar 2015 A1
20150183370 Nakanishi et al. Jul 2015 A1
20150210213 Mitsuta et al. Jul 2015 A1
20150210279 Agnew et al. Jul 2015 A1
20150307024 Fukuda et al. Oct 2015 A1
20150326829 Kurihara et al. Nov 2015 A1
20150343976 Lim et al. Dec 2015 A1
20150346337 Jung Dec 2015 A1
20150355333 Ono et al. Dec 2015 A1
20160016512 Masutani et al. Jan 2016 A1
20160059854 Ohsugi et al. Mar 2016 A1
20160060826 Berning et al. Mar 2016 A1
20160063864 Ohsugi et al. Mar 2016 A1
20160114728 Tan et al. Apr 2016 A1
20160120095 Fujimoto May 2016 A1
20160137125 Petrany et al. May 2016 A1
20160170411 Wei et al. Jun 2016 A1
20160176338 Husted et al. Jun 2016 A1
20160185346 Awamori et al. Jun 2016 A1
20160202351 Uotsu Jul 2016 A1
20160236616 Kurihara et al. Aug 2016 A1
20160264133 Ohsugi et al. Sep 2016 A1
20160264134 Ohsugi et al. Sep 2016 A1
20160272172 Lee Sep 2016 A1
20160282874 Kurata et al. Sep 2016 A1
20160312446 Pettersson et al. Oct 2016 A1
20160375830 Karner et al. Dec 2016 A1
20170016210 Kowatari et al. Jan 2017 A1
20170018070 Onuma et al. Jan 2017 A1
20170057497 Laur et al. Mar 2017 A1
20170113608 Patel et al. Apr 2017 A1
20170135277 Hiramatsu et al. May 2017 A1
20170146990 Wang et al. May 2017 A1
20170180707 Hsu et al. Jun 2017 A1
20170191243 Sharp et al. Jul 2017 A1
20170205506 Voorheis et al. Jul 2017 A1
20170220044 Sakai et al. Aug 2017 A1
20170225680 Huang et al. Aug 2017 A1
20170229019 Ichinose et al. Aug 2017 A1
20170236422 Naka et al. Aug 2017 A1
20170242436 Creusot Aug 2017 A1
20170248439 Sakai et al. Aug 2017 A1
20170267172 Irby, Sr. Sep 2017 A1
20170285649 Debreczeni Oct 2017 A1
20170286763 Fukuda et al. Oct 2017 A1
20170301107 Sasaki Oct 2017 A1
20170349169 Deshpande Dec 2017 A1
20170356994 Wodrich et al. Dec 2017 A1
20170364072 Yako et al. Dec 2017 A1
20170364759 Creusot Dec 2017 A1
20180009378 Myers et al. Jan 2018 A1
20180015875 May et al. Jan 2018 A1
20180045812 Hess Feb 2018 A1
20180059236 Wodrich et al. Mar 2018 A1
20180072313 Stenneth Mar 2018 A1
20180074178 Cook et al. Mar 2018 A1
20180074201 Sakai et al. Mar 2018 A1
20180074506 Branson Mar 2018 A1
20180081368 Watanabe et al. Mar 2018 A1
20180105107 Hassan-Shafique et al. Apr 2018 A1
20180107871 Singh et al. Apr 2018 A1
20180204335 Agata et al. Jul 2018 A1
20180210454 Ready-Campbell et al. Jul 2018 A1
20180236985 Kim et al. Aug 2018 A1
20180236986 Kim et al. Aug 2018 A1
20180239991 Weller et al. Aug 2018 A1
20180252810 Uotsu Sep 2018 A1
20180273030 Weldon et al. Sep 2018 A1
20180284787 Naka et al. Oct 2018 A1
20180292840 Sakai Oct 2018 A1
20180307921 Vallespi-Gonzalez et al. Oct 2018 A1
20180322653 Tatarnikov et al. Nov 2018 A1
20180326982 Paris et al. Nov 2018 A1
20190011993 Ette et al. Jan 2019 A1
20190025057 Sakai et al. Jan 2019 A1
20190027028 Fields et al. Jan 2019 A1
20190047488 Lee et al. Feb 2019 A1
20190072666 Duque Biarge et al. Mar 2019 A1
20190072667 Duque Biarge et al. Mar 2019 A1
20190072668 Duque Biarge et al. Mar 2019 A1
20190072669 Duque Biarge et al. Mar 2019 A1
20190094857 Jertberg et al. Mar 2019 A1
20210254314 Takata Aug 2021 A1
Foreign Referenced Citations (72)
Number Date Country
2016376507 May 2018 AU
2017305113 Nov 2018 AU
2017408956 Apr 2019 AU
2744561 Dec 2012 CA
102955938 Mar 2013 CN
103057459 Apr 2013 CN
202873034 Apr 2013 CN
104655186 May 2015 CN
204346505 May 2015 CN
107244324 Oct 2017 CN
206657493 Nov 2017 CN
208460034 Feb 2019 CN
19961731 Jul 2001 DE
10144689 Mar 2003 DE
102011017707 Oct 2012 DE
102016012342 Apr 2017 DE
102016215102 Dec 2017 DE
102018003788 Nov 2018 DE
102017216000 Mar 2019 DE
717290 Jun 1996 EP
2354316 Nov 2012 EP
2524847 Nov 2012 EP
3482503 May 2019 EP
2663655 Apr 2018 ES
2507769 May 2014 GB
2567142 Apr 2019 GB
11353593 Dec 1999 JP
2002074595 Mar 2002 JP
2008163719 Jul 2008 JP
2008248613 Oct 2008 JP
2010112100 May 2010 JP
2010198519 Sep 2010 JP
2010204821 Sep 2010 JP
2016022804 Feb 2016 JP
2016119559 Jun 2016 JP
2017074870 Apr 2017 JP
2017074871 Apr 2017 JP
2017074873 Apr 2017 JP
2017085370 May 2017 JP
2017161467 Sep 2017 JP
2017172223 Sep 2017 JP
2018043599 Mar 2018 JP
20020093659 Dec 2002 KR
100531198 Nov 2005 KR
20120140545 Dec 2012 KR
20130021986 Mar 2013 KR
20130042963 Apr 2013 KR
101489836 Feb 2015 KR
101566963 Nov 2015 KR
20160091585 Aug 2016 KR
20160121026 Oct 2016 KR
101684095 Dec 2016 KR
20170052713 May 2017 KR
20170119059 Oct 2017 KR
20180041525 Apr 2018 KR
2016121688 Aug 2011 WO
2014072730 May 2014 WO
2015060218 Apr 2015 WO
2017022262 Feb 2017 WO
2017110159 Jun 2017 WO
2018007995 Jan 2018 WO
2018026021 Feb 2018 WO
2018050945 Mar 2018 WO
2018056125 Mar 2018 WO
2018078054 May 2018 WO
2018042567 Aug 2018 WO
2018158642 Sep 2018 WO
2018172849 Sep 2018 WO
2018189952 Oct 2018 WO
2018215833 Nov 2018 WO
2019043446 Mar 2019 WO
2019048011 Mar 2019 WO
Related Publications (1)
Number Date Country
20210124364 A1 Apr 2021 US