Patent Grant
9849588
This application is related to co-pending and co-owned U.S. patent application Ser. No. 14/489,368, entitled “APPARATUS AND METHODS FOR CONTEXT DETERMINATION USING REAL TIME SENSOR DATA”, filed contemporaneously herewith on Sep. 17, 2014; Ser. No. 14/489,353, entitled “HOME ANIMATION APPARATUS AND METHODS”, filed contemporaneously herewith on Sep. 17, 2014; Ser. No. 14/489,373, entitled “APPARATUS AND METHODS FOR REMOVAL OF LEARNED BEHAVIORS IN ROBOTS”, filed contemporaneously herewith on Sep. 17, 2014, Ser. No. 14/244,888, entitled “LEARNING APPARATUS AND METHODS FOR REMOTE CONTROL OF ROBOTIC DEVICES VIA SPOOFING”, filed Apr. 3, 2014, Ser. No. 14/244,892 entitled “SPOOFING REMOTE CONTROL APPARATUS AND METHODS”, filed Apr. 3, 2014, Ser. No. 14/244,890, entitled “APPARATUS AND METHODS FOR REMOTELY CONTROLLING ROBOTIC DEVICES”, filed Apr. 3, 2014, Ser. No. 14/208,709 entitled “TRAINABLE MODULAR ROBOTIC APPARATUS AND METHODS”, filed Mar. 13 2014, Ser. No. 13/953,595 entitled “APPARATUS AND METHODS FOR CONTROLLING OF ROBOTIC DEVICES”, filed Jul. 29, 2013, Ser. No. 13/918,338 entitled “ROBOTIC TRAINING APPARATUS AND METHODS”, filed Jun. 14, 2013, Ser. No. 13/918,298 entitled “HIERARCHICAL ROBOTIC CONTROLLER APPARATUS AND METHODS”, filed Jun. 14, 2013, Ser. No. 13/907,734 entitled “ADAPTIVE ROBOTIC INTERFACE APPARATUS AND METHODS”, filed May 31, 2013, Ser. No. 13/842,530 entitled “ADAPTIVE PREDICTOR APPARATUS AND METHODS”, filed Mar. 15, 2013, Ser. No. 13/842,562 entitled “ADAPTIVE PREDICTOR APPARATUS AND METHODS FOR ROBOTIC CONTROL”, filed Mar. 15, 2013, Ser. No. 13/842,616 entitled “ROBOTIC APPARATUS AND METHODS FOR DEVELOPING A HIERARCHY OF MOTOR PRIMITIVES”, filed Mar. 15, 2013, Ser. No. 13/842,647 entitled “MULTICHANNEL ROBOTIC CONTROLLER APPARATUS AND METHODS”, filed Mar. 15, 2013, Ser. No. 13/842,583 entitled “APPARATUS AND METHODS FOR TRAINING OF ROBOTIC DEVICES”, filed Mar. 15, 2013, Ser. No. 13/918,620 entitled “PREDICTIVE ROBOTIC CONTROLLER APPARATUS AND METHODS”, filed Jun. 14, 2013, and Ser. No. 13/465,918 entitled “SPIKING NEURAL NETWORK OBJECT RECOGNITION APPARATUS AND METHODS”, filed May 7, 2012, each of the foregoing being incorporated herein by reference in its entirety.
A portion of the disclosure of this patent document contains material that is subject to copyright protection. The copyright owner has no objection to the facsimile reproduction by anyone of the patent document or the patent disclosure, as it appears in the Patent and Trademark Office patent files or records, but otherwise reserves all copyright rights whatsoever.
Technological Field
The present disclosure relates to adaptive control and training of computerized devices.
Background
Computerized devices may be used in a variety of home automation applications such as autonomous vacuum cleaners, multifunction remote controllers, entertainment appliances, lighting systems and/or other appliances. Some existing devices (e.g., Roomba®) may be pre-programmed at a factory in order to perform desired functionality. Some existing devices (e.g., Harmony Remote®) may be programmed by a user in order to perform desired functionality. Some home automation devices (e.g., Savant audio/video distribution and/or lighting appliances) may interface with a portable device (e.g., a tablet, smartphone) running a specific application (App) in order to enable a user to operate the appliance. However, operation of robotic devices by such controllers may be characterized by a constrained form of the user interface, e.g., a requirement to carry around e.g., an iPhone, in order to be able to turn on/off an audiovisual equipment. Some such existing solutions may require some form of “programming” (e.g., button assignment, model selection) at minimum, arranging buttons on one or more page views of a smartphone app, and/or other operation.
One aspect of the disclosure relates to a robotic system. The system may comprise a first robotic device. The system may comprise a remote controller configured to provide a first wireless transmission conveying a first control instruction from a user to the first robotic device. The first control instruction may be configured to cause the first robotic device to execute a first action. The system may comprise a learning controller apparatus configured to, during first time interval: receive the first wireless transmission; determine a first context associated with execution of the first action; and determine an association between data related to the first control instruction and the first context. During a second time interval, subsequent to the first time interval, the learning controller apparatus may be configured to: determine another version of the first context related to execution of the first action; and provide, in lieu of the first control instruction, a first remote transmission conveying data related to the first control instruction, the data related to the first control instruction being configured to cause the first robotic device to execute the first action.
A method of providing transmissions to a remotely controlled apparatus by a learning controller is disclosed. In one embodiment, the method includes: configuring the learning controller to monitor one or more sensors configured to provide information related to a state of the remotely controlled apparatus disposed remotely from the learning controller; and enabling the learning controller to transmit a wireless command to the apparatus.
In one variant, at least one of the one or more sensors comprises a video camera. In some implementations, the transmission of the wireless command comprises an infrared signal transmission. In other implementations, the transmission of the wireless command comprises a radio frequency signal transmission.
In another variant the learning controller is configured to record user commands and the state of the remotely controlled apparatus; and the learning controller transmits the wireless command when the remotely controlled apparatus is in a recording state.
In still other variants, the learning controller comprises a computing cloud entity in data communication with the remotely controlled apparatus; the computing cloud entity is configured to record the state of the remotely controlled apparatus; and the computing cloud entity is configured to provide the wireless command based on a detection of the remotely controlled apparatus entering a recorded state.
In another variant, the learning controller is further configured to receive the wireless command from a computing cloud entity, the wireless command based on a profile configured to associate the state with the wireless command, where the learning controller is disposed remote from the cloud entity; and the profile is configured in accordance with a characteristic of the remotely controlled apparatus.
A method of remotely controlling a robotic device by a universal controller is disclosed. In one embodiment the method includes: loading a set of instructions within the universal controller, the set of instructions configured to control the robotic device; monitoring a state of the robotic device; determining a command according to the set of instructions; and transmitting the command to the robotic device, the transmission causing the robotic device to execute the command to perform a task.
In one variant, the state of the robotic device comprises a position of the robotic device in an environment of the robotic device.
In another variant, the state of the robotic device comprises a parameter characterizing an event that is external to the robotic device. For example, the parameter may convey one or more of a position, a gesture, or a movement of a person.
In one variant, the robotic device is configured to perform the task within an environment comprising one or more objects; and the task comprises the robotic device displacing at least one of the one or more objects within the environment.
In one variant, the robotic device is configured to perform the task within an environment comprising one or more objects characterized by corresponding object motion parameters; the command execution is characterized by a first motion parameter of the robotic device; and the command execution is configured based on a modification of the first motion parameter based on value associated with at least one of the corresponding object motion parameters.
In some variants the set of instructions was previously configured for another robotic device or another task.
In some implementations, the loading of the set of instructions is triggered automatically based on a detection of a remote transmission or a pre-defined context in a sensory input.
In yet another incarnation, monitoring the state of the robotic device comprises receiving an input from a video camera. In one such case, the robotic device is configured to perform the task within an environment comprising one or more objects; and the monitored state comprises a position of the robotic device relative to a position of the one or more objects determined based on the input from the video camera.
In one variation, the robotic device is configured to perform the task within an environment comprising one or more objects characterized by corresponding object motion parameters; and the task comprises removing one of the one or more objects from a subset of the environment. In another such example, the robotic device is configured to perform the task within an environment comprising a first area and a second area; the monitoring comprises monitoring the state of the robotic device by the first universal controller in the first area and where the state of the robotic device is monitored by another universal controller in the second area; and the transmission of the command is effectuated by one of the universal controller when the robotic device is within the first area.
In some embodiments, the robotic device is configured to receive the transmission of the command from the universal controller when the robotic device is within the first monitoring area; and the robotic device is configured to receive another transmission from the another universal controller when the robotic device is within the second monitoring area.
In one variant, the robotic device comprises a sensor component; the transmission comprises transmission of an infrared frequency signal; and the state of the robotic device is determined based on an output provided by the sensor component.
In still other variants, the robotic device comprises a household appliance; and the task comprises one or more of a cleaning task.
These and other objects, features, and characteristics of the present disclosure, as well as the methods of operation and functions of the related elements of structure and the combination of parts and economies of manufacture, will become more apparent upon consideration of the following description and the appended claims with reference to the accompanying drawings, all of which form a part of this specification, wherein like reference numerals designate corresponding parts in the various figures. It is to be expressly understood, however, that the drawings are for the purpose of illustration and description only and are not intended as a definition of the limits of the disclosure. As used in the specification and in the claims, the singular form of “a”, “an”, and “the” include plural referents unless the context clearly dictates otherwise.
All Figures disclosed herein are © Copyright 2014 Brain Corporation. All rights reserved.
Implementations of the present technology will now be described in detail with reference to the drawings, which are provided as illustrative examples so as to enable those skilled in the art to practice the technology. Notably, the figures and examples below are not meant to limit the scope of the present disclosure to a single implementation, but other implementations are possible by way of interchange of or combination with some or all of the described or illustrated elements. Wherever convenient, the same reference numbers will be used throughout the drawings to refer to same or like parts.
Where certain elements of these implementations can be partially or fully implemented using known components, only those portions of such known components that are necessary for an understanding of the present technology will be described, and detailed descriptions of other portions of such known components will be omitted so as not to obscure the disclosure.
In the present specification, an implementation showing a singular component should not be considered limiting; rather, the disclosure is intended to encompass other implementations including a plurality of the same component, and vice-versa, unless explicitly stated otherwise herein.
Further, the present disclosure encompasses present and future known equivalents to the components referred to herein by way of illustration.
As used herein, the term “bus” is meant generally to denote all types of interconnection or communication architecture that is used to access the synaptic and neuron memory. The “bus” may be optical, wireless, infrared, and/or another type of communication medium. The exact topology of the bus could be for example standard “bus”, hierarchical bus, network-on-chip, address-event-representation (AER) connection, and/or other type of communication topology used for accessing, e.g., different memories in pulse-based system.
As used herein, the terms “computer”, “computing device”, and “computerized device” may include one or more of personal computers (PCs) and/or minicomputers (e.g., desktop, laptop, and/or other PCs), mainframe computers, workstations, servers, personal digital assistants (PDAs), handheld computers, embedded computers, programmable logic devices, personal communicators, tablet computers, portable navigation aids, J2ME equipped devices, cellular telephones, smart phones, personal integrated communication and/or entertainment devices, and/or any other device capable of executing a set of instructions and processing an incoming data signal.
As used herein, the term “computer program” or “software” may include any sequence of human and/or machine cognizable steps which perform a function. Such program may be rendered in a programming language and/or environment including one or more of C/C++, C#, Fortran, COBOL, MATLAB™, PASCAL, Python, assembly language, markup languages (e.g., HTML, SGML, XML, VoXML), object-oriented environments (e.g., Common Object Request Broker Architecture (CORBA)), Java™ (e.g., J2ME, Java Beans), Binary Runtime Environment (e.g., BREW), and/or other programming languages and/or environments.
As used herein, the term “memory” may include an integrated circuit and/or other storage device adapted for storing digital data. By way of non-limiting example, memory may include one or more of ROM, PROM, EEPROM, DRAM, Mobile DRAM, SDRAM, DDR/2 SDRAM, EDO/FPMS, RLDRAM, SRAM, “flash” memory (e.g., NAND/NOR), memristor memory, PSRAM, and/or other types of memory.
As used herein, the terms “integrated circuit”, “chip”, and “IC” are meant to refer to an electronic circuit manufactured by the patterned diffusion of trace elements into the surface of a thin substrate of semiconductor material. By way of non-limiting example, integrated circuits may include field programmable gate arrays (e.g., FPGAs), a programmable logic device (PLD), reconfigurable computer fabrics (RCFs), application-specific integrated circuits (ASICs), and/or other types of integrated circuits.
As used herein, the terms “microprocessor” and “digital processor” are meant generally to include digital processing devices. By way of non-limiting example, digital processing devices may include one or more of digital signal processors (DSPs), reduced instruction set computers (RISC), general-purpose (CISC) processors, microprocessors, gate arrays (e.g., field programmable gate arrays (FPGAs)), PLDs, reconfigurable computer fabrics (RCFs), array processors, secure microprocessors, application-specific integrated circuits (ASICs), and/or other digital processing devices. Such digital processors may be contained on a single unitary IC die, or distributed across multiple components.
As used herein, the term “network interface” refers to any signal, data, and/or software interface with a component, network, and/or process. By way of non-limiting example, a network interface may include one or more of FireWire (e.g., FW400, FW800, etc.), USB (e.g., USB2), Ethernet (e.g., 10/100, 10/100/1000 (Gigabit Ethernet), 10-Gig-E, etc.), MoCA, Coaxsys (e.g., TVnet™), radio frequency tuner (e.g., in-band or OOB, cable modem, etc.), Wi-Fi (802.11), WiMAX (802.16), PAN (e.g., 802.15), cellular (e.g., 3G, LTE/LTE-A/TD-LTE, GSM, etc.), IrDA families, and/or other network interfaces.
As used herein, the term “Wi-Fi” includes one or more of IEEE-Std. 802.11, variants of IEEE-Std. 802.11, standards related to IEEE-Std. 802.11 (e.g., 802.11 a/b/g/n/s/v), and/or other wireless standards.
As used herein, the term “wireless” means any wireless signal, data, communication, and/or other wireless interface. By way of non-limiting example, a wireless interface may include one or more of Wi-Fi, Bluetooth, 3G (3GPP/3GPP2), HSDPA/HSUPA, TDMA, CDMA (e.g., IS-95A, WCDMA, etc.), FHSS, DSSS, GSM, PAN/802.15, WiMAX (802.16), 802.20, narrowband/FDMA, OFDM, PCS/DCS, LTE/LTE-A/TD-LTE, analog cellular, CDPD, satellite systems, millimeter wave or microwave systems, acoustic, infrared (i.e., IrDA), and/or other wireless interfaces.
The apparatus 100 may comprise an interface to a sensor component configured to provide sensory input 108 to the learning process. In some implementations, the sensor component of the may comprise a camera, a microphone, a radio wave sensor, an ultrasonic sensor, and/or other sensor capable of providing information related environment of the appliance device. The sensor information may be used for determining context associated with the environment using any applicable methodologies including these described below with respect to
In one or more implementations, such as object and/or gesture recognition, the sensory input 108 may comprise a stream of pixel values associated with one or more digital images. In one or more implementations of e.g., video, radar, sonography, x-ray, magnetic resonance imaging, and/or other types of sensing, the input may comprise electromagnetic waves (e.g., visible light, infrared (IR), ultraviolet (UV), and/or other types of electromagnetic waves) entering an imaging sensor array. In some implementations, the imaging sensor array may comprise one or more of artificial retinal ganglion cells (RGCs), a charge coupled device (CCD), an active-pixel sensor (APS), and/or other sensors. The input signal may comprise a sequence of images and/or image frames. The sequence of images and/or image frame may be received from a CCD camera via a receiver apparatus and/or downloaded from a file. The image may comprise a two-dimensional matrix of red green, blue (RGB) values refreshed at a 25 Hz frame rate (and/or any suitable frame rate). It will be appreciated by those skilled in the arts that the above image parameters are merely exemplary, and many other image representations (e.g., bitmap, CMYK, HSV, HSL, grayscale, and/or other representations) and/or frame rates are equally useful with the present disclosure. Pixels and/or groups of pixels associated with objects and/or features in the input frames may be encoded using, for example, latency encoding described in U.S. patent application Ser. No. 12/869,583, filed Aug. 26, 2010 and entitled “INVARIANT PULSE LATENCY CODING SYSTEMS AND METHODS”; U.S. Pat. No. 8,315,305, issued Nov. 20, 2012, entitled “SYSTEMS AND METHODS FOR INVARIANT PULSE LATENCY CODING”; U.S. patent application Ser. No. 13/152,084, filed Jun. 2, 2011, entitled “APPARATUS AND METHODS FOR PULSE-CODE INVARIANT OBJECT RECOGNITION”; and/or latency encoding comprising a temporal winner take all mechanism described U.S. patent application Ser. No. 13/757,607, filed Feb. 1, 2013 and entitled “TEMPORAL WINNER TAKES ALL SPIKING NEURON NETWORK SENSORY PROCESSING APPARATUS AND METHODS”, each of the foregoing being incorporated herein by reference in its entirety.
In one or more implementations, object recognition and/or classification may be implemented using spiking neuron classifier comprising conditionally independent subsets as described in co-owned U.S. patent application Ser. No. 13/756,372 filed Jan. 31, 2013, and entitled “SPIKING NEURON CLASSIFIER APPARATUS AND METHODS USING CONDITIONALLY INDEPENDENT SUBSETS” and/or co-owned U.S. patent application Ser. No. 13/756,382 filed Jan. 31, 2013, and entitled “REDUCED LATENCY SPIKING NEURON CLASSIFIER APPARATUS AND METHODS”, each of the foregoing being incorporated herein by reference in its entirety.
In one or more implementations, encoding may comprise adaptive adjustment of neuron parameters, such neuron excitability described in U.S. patent application Ser. No. 13/623,820 entitled “APPARATUS AND METHODS FOR ENCODING OF SENSORY DATA USING ARTIFICIAL SPIKING NEURONS”, filed Sep. 20, 2012, the foregoing being incorporated herein by reference in its entirety.
In some implementations, analog inputs may be converted into spikes using, for example, kernel expansion techniques described in U.S. patent application Ser. No. 13/623,842 filed Sep. 20, 2012, and entitled “SPIKING NEURON NETWORK ADAPTIVE CONTROL APPARATUS AND METHODS”, issued as U.S. Pat. No. 9,367,798 on Jun. 14, 2016, the foregoing being incorporated herein by reference in its entirety. In one or more implementations, analog and/or spiking inputs may be processed by mixed signal spiking neurons, such as U.S. patent application Ser. No. 13/313,826 entitled “APPARATUS AND METHODS FOR IMPLEMENTING LEARNING FOR ANALOG AND SPIKING SIGNALS IN ARTIFICIAL NEURAL NETWORKS”, filed Dec. 7, 2011, and/or U.S. patent application Ser. No. 13/761,090 entitled “APPARATUS AND METHODS FOR GATING ANALOG AND SPIKING SIGNALS IN ARTIFICIAL NEURAL NETWORKS”, filed Feb. 6, 2013 and issued as U.S. Pat. No. 9,213,937 on Dec. 15, 2015, each of the foregoing being incorporated herein by reference in its entirety.
Contemporaneous with the sensory input 108 the apparatus 100 may receive control input 106 comprising one or more commands for the appliance device. In one or more implementations, the appliance device may comprise an autonomous vacuum cleaner, a lighting device, an entertainment appliance, a heating, ventilation, and air conditioning (HVAC) device, a remotely operated fan, a lawnmower, a mechanized trash bin, a sprinkler system, and/or other home electrical and/or electromechanical apparatus.
In some implementations, e.g., such as described in U.S. patent application Ser. No. 14/244,888, entitled “LEARNING APPARATUS AND METHODS FOR CONTROL OF ROBOTIC DEVICES VIA SPOOFING”, filed Apr. 3, 2014, the contents of which were previously incorporated herein by reference in its entirety supra, the control input 106 may be provided by a user of the appliance using a remote control device. On one or more implementations, the input 106 may be provided by a computerized agent, e.g., home automation controller, a proximity sensor, a motion detection device. The input 106 may be communicated using via a wired (e.g., light switch, serial bus, Ethernet) interface and/or wireless (e.g., using radio frequency (RF), infrared (IR), pressure (sound), light, and/or other wave carrier transmissions).
By way of an illustration, a user may use a gesture (e.g., point a hand at a remotely operated lamp) accompanied by a command issued via a remote control device. The sensory input 108 may comprise a video camera output, the control input 106 may comprise one or more infrared light, RF pulses and/or other pulses associated with a button press of the remote controller.
The apparatus 100 may operate an adaptable predictor block configured to, inter alia, determine an association between the remote control instructions 106 and context determined from the sensory input 108. In some implementations, the context may comprise location of user body, head, and/or limb position, orientation, and/or configuration), presence, size, and/or location of objects (e.g., other users), and/or other information associated with environment of the user (e.g., daylight temperature, time of day, and/or other information). By way of an illustration, the same gesture issued in different locations by a user may comprise different sensory contexts.
Various methodologies may be utilized in order to determine the associations between the context and user instructions, including, e.g., these described in U.S. patent application Ser. No. 13/953,595 entitled “APPARATUS AND METHODS FOR CONTROLLING OF ROBOTIC DEVICES”, filed Jul. 29, 2013; U.S. patent application Ser. No. 13/918,338 entitled “ROBOTIC TRAINING APPARATUS AND METHODS”, filed Jun. 14, 2013; U.S. patent application Ser. No. 13/918,298 entitled “HIERARCHICAL ROBOTIC CONTROLLER APPARATUS AND METHODS”, filed Jun. 14, 2013; U.S. patent application Ser. No. 13/918,620 entitled “PREDICTIVE ROBOTIC CONTROLLER APPARATUS AND METHODS”, filed Jun. 14, 2013; U.S. patent application Ser. No. 13/907,734 entitled “ADAPTIVE ROBOTIC INTERFACE APPARATUS AND METHODS”, filed May 31, 2013; U.S. patent application Ser. No. 13/842,530 entitled “ADAPTIVE PREDICTOR APPARATUS AND METHODS”, filed Mar. 15, 2013; U.S. patent application Ser. No. 13/842,562 entitled “ADAPTIVE PREDICTOR APPARATUS AND METHODS FOR ROBOTIC CONTROL”, filed Mar. 15, 2013; U.S. patent application Ser. No. 13/842,616 entitled “ROBOTIC APPARATUS AND METHODS FOR DEVELOPING A HIERARCHY OF MOTOR PRIMITIVES”, filed Mar. 15, 2013; U.S. patent application Ser. No. 13/842,647 entitled “MULTICHANNEL ROBOTIC CONTROLLER APPARATUS AND METHODS”, filed Mar. 15, 2013; and U.S. patent application Ser. No. 13/842,583 entitled “APPARATUS AND METHODS FOR TRAINING OF ROBOTIC DEVICES”, filed Mar. 15, 2013; each of the foregoing being incorporated herein by reference in its entirety, and/or U.S. patent application Ser. No. 14/244,890, entitled “APPARATUS AND METHODS FOR REMOTELY CONTROLLING ROBOTIC DEVICES”, filed Apr. 3, 2014, previously incorporated herein by reference in its entirety supra.
An action may be assigned to a given context. In some implementations, e.g., such as described below with respect to
Developed associations between the sensory context and the control commands may be stored for further use, as shown by arrow 104 in
During learning, the apparatus 100 may be configured to provide feedback 102. In some implementations of training by the user, the feedback 102 may comprise one or more of audible, vibration, and/or visible indications to the user configured to communicate information related to learning progress (e.g., a success indication signifying that the control command was received and the association with the sensory context was determined and stored, a failure-to-learn indication signifying that the control command was received but the association with the sensory context was not successfully determined or stored; and/or a perceptual indication signifying that the sensory context was ambiguous, had no discernible signal, and/or substantially overlapped with the sensory context already associated with another control command.) In some implementations, lack of the indication may signify to the user that the control command was not received by the appliance (e.g., due to obstruction or interference). In some implementations of training by the computerized agent, the feedback 102 may comprise one or more of electrical, infrared, RF, signals and/or computer messages configured to communicate information related to learning progress.
Upon developing the associations between the sensory context and the remote control commands, the computerized apparatus (e.g., 100 in
The apparatus 110 of
The apparatus 110 of
The control instructions of the output 116 may be produced based on a determination of occurrence one or more previously occurring context(s) within the sensory input 118. By way of an illustration, the apparatus 110 may determine an occurrence of a gesture (e.g., a user pointing a hand at a TV) using a camera, a voice command “ROOMBA VACUUM DINGING ROOM NOW” using a microphone, and/or other context. During learning, contemporaneously with the gesture, the user may issue an “ON” control command to the TV, e.g., using an RF or IR remote controller as described above with respect to
During operation, the apparatus 110 may be configured to provide feedback 112. In some implementations of training by the user, the feedback 112 may comprise one or more of audible, vibration, and/or visible indications to the user configured to communicate information related to operation progress (e.g., such as described above with respect to
In some implementations, wherein the output 116 may comprise RF signal transmission, a bound connection (e.g. a radio link session) may be established between a controller 110 and the appliance. Some link examples may include a Bluetooth session, a Digital Spectrum Modulation (DSM) session, and/or other links. The link may be established based on identity of the appliance device and/or the controller.
In one or more implementations, wherein information for establishment of the link (e.g., communication protocol, hopping sequence, device identification codes, and/or other), may not be readily available, a multi-carrier command delivery approach may be employed, e.g., such as described below with respect to
The system 200 may comprise a learning controller apparatus 210, e.g., the apparatus 100, 110 described above with respect to
Various implementations of the data communication 216 between the controller component 214 and the appliance 220 may be employed. In some implementations, a Direct Sequence Spread Spectrum (DSSS), and/or frequency hopping spread spectrum” (FHSS) technology may be utilized. DSSS communication technology may employ carrier phase-modulation using a string of pseudorandom (PR) code symbols called “chips”, each of which may have duration that is shorter than an information bit. That is, each information bit is modulated by a sequence of much faster chips. Therefore, the chip rate is much higher than the information signal bit rate. DSSS uses a signal structure in which the sequence of chips produced by the transmitter is already known by the receiver. The receiver may apply the known PR sequence to counteract the effect of the PR sequence on the received signal in order to reconstruct the information signal.
Frequency-hopping spread spectrum (FHSS) is a method of transmitting radio signals by rapidly switching a carrier among many frequency channels, using a pseudorandom sequence known to both transmitter and receiver.
The controller apparatus 210 may comprise a sensor component 212 configured to provide sensory input to the controller. In some implementations of a bStem™ integrated platform, the component may comprise a discrete camera module and/or microphone component. In one or more implementations wherein the controller 210 may comprise a portable computing and communications device (e.g., a tablet, a smartphone device) the component 214 may comprise a built-in camera and/or microphone of the computing device. The component 212 may provide information related to gestures, voice commands, and/or other indications configured to convey intentions of a user.
By way of an illustration, the user may whirl an arm around and/or point a finger at the fan 220 in order to indicate a start operation. The user gesture intention may be accompanied by a transmission of an “ON” command using the remote controller 202. In some implementations, the controller 210 may be configured to establish a data link with (pair to) the remote controller 202. Pairing of the remote controller 202 and the learning controller 210 may enable transmission of the user commands from the handset 202 to the learning controller 210. In some implementations establishment of the link 206 may be facilitated by, e.g., obtaining link protocol parameters specifications (e.g., the spreading code, list of device IDs) for given remote controller device 202.
In order to learn associations between user commands and context associated with the task, the learning controller 210 may comprise a receiver component (e.g., the receiver component 234 of the controller apparatus 230 shown in
The controller 210 may comprise a processing component configured to operate a learning process configured to learn associations between the user intentions (e.g., the gesture) and the control command provided via transmissions 206.
Referring now to
The processing component 240 may comprise a specialized computerized apparatus (e.g., bStem™) and/or a general use computerized apparatus (e.g., a tablet computer, a smartphone, a computer, and/or other computerized device) configured to operate a learning process (e.g., by BrainOS library). In one or more implementations, the learning process operation may be by a remote server interfaced to the processing component 240 via a wired or wireless link.
In some implementations, the controller apparatus 230 may comprise a user interface component 238 configured to communicate to the learning process indications configured to start, stop, pause, and/or resume learning. In some implementations, learning may commence automatically based on detection of one or more remote transmissions 206 and/or context in the sensory input. In one or more implementations, learning may terminate based on receipt of a stop command via the interface 238, expiration of a time interval, number of context-command occurrences, and/or another indication (e.g., absence of transmissions 206 within a given time period) automatically based on detection of one or more remote transmissions 206 and/or context in the sensory input.
Subsequent to activation of learning, the learning process may detect one or more control instructions within the remote transmissions (e.g., 206 in
In one or more implementations, data from two or more sensor types may be combined (e.g., via an adaptive data fusion process). A false positive of features detected using video, may be filtered out using audio signal provided in order to disambiguate context detection. By way of an illustration of one implementation of audio/visual sensor input, a representation of a person leaning towards a cell phone may not trigger an association with a TV mute action. However, a representation of the person leaning towards a cell phone when accompanied by the cell phone ring may cause muting of the TV. In another such example, an audio/visual sensor input may detect when there are multiple raised hands among a group of people, and then an automatic camera may select the appropriate person with a raised hand gesture (from the group of people) based on a voice indication (e.g., “I have a question”).
In some implementations, wherein the learning controller operation may be effectuated by a portable communications device (e.g., a smartphone) determination of the associations between the context and user control instructions may be effectuated by the portable device using sensory data obtained by a sensor component of the portable device (e.g., a microphone, a camera, and/or other component).
In some implementations, determination of the associations between the context and user control instructions may be effectuated by a computing entity (e.g., a local computer and/or a remote computing Cloud) in data communication with the apparatus 230 of
Returning now to
Some computerized appliances (as originally provided by a manufacturer) may not be provided with remote control functionality (e.g., dishwasher, refrigerator, and/or other). Some appliance operations, may benefit from an operational trajectory that may prove too difficult for a non-expert user to navigate, too time consuming, too/boring (e.g., follow a regular pattern when vacuuming/mowing a lawn) and/or otherwise not desirable to be performed by the user using manual remote control operations. Some operations may prove too complex for remote control operations (e.g., navigating an unmanned aerial vehicle around subjects of interest while avoiding obstacles, e.g., as described in co-pending and co-owned owned U.S. patent application Ser. No. 14/332,322 entitled “APPARATUS AND METHODS FOR TRACKING USING AERIAL VIDEO”, filed Jul. 15, 2014, the foregoing being incorporated by reference in its entirety). In some implementations, a user may elect to remotely operate computerized appliances using interfaces that may originally not intended for use with the appliance. By way of an illustration, the user may elect to employ gestures instead of using a remote controller and/or an app running on a tablet computer and/or smartphone in order to operate a home appliance ordinarily operated via e.g., an infrared remote (e.g., TV, fan, HVAC, and or other devices).
In some implementations, a user may train a learning computerized device to develop an association between a context derived from sensory information and an action. In some implementations, the context may comprise user gesture, audible signal (a whistle, a spoken word, a clap, and/or other indication); the action may comprise activation of a user interface element (e.g., a soft/hard button press, a swipe, touch screen activation) configured cause generation of an event indicative of the context. By way of an illustration of using a home automation system comprising a user interface device, the user may teach the learning computerized device to trigger the event by associating the context (e.g., hand gesture) with a pressing of a button on the user interface device. In one or more implementations, the context change may be determined based on a history (within time window) of a given characteristic. By way of an illustration of audio based context, if a background sound may be present initially (e.g., due to operation of a motor appliance, e.g. a washer), the sudden absence of the sound may identify context and trigger an n inspection action. Initial presence of an object on the table (e.g., a flower vase) followed by sudden absence of that flower vase from the table may a cleaning action.
The apparatus 310 may comprise a user interface device 310, e.g., a smartphone, comprising a camera, a display, and a touch interface, and the learning controller operation may be effectuated by executing a plurality of instructions (e.g., a software application and/or a library) by the device 310.
During training, the device 310 may be operable to determine a context (e.g., user gesture and/or sound) using a camera and/or a microphone. Contemporaneous with the provision of the gesture and/or the sound, the user may activate one or more elements of the user interface (e.g., by touching soft buttons 312, 314). The software application executed by the device 310 may be configured to produce an event comprising an event identification (EventID) associated with activation of a given button, e.g., Button1ID. Training of the associations may be effectuated using any applicable methodologies including these described herein.
Subsequent to training, the user may link given EventID with an action to be performed by a home appliance. In some implementations, the link may be effectuated by providing executable instructions to the device 310, configured to turn on a given appliance. By way of an illustration, the user may provide code configured to produce a plurality of RF or IR pulses to operate a fan and/or TV remotely. In one or more implementations, if this than that (IFTTT) framework may be employed wherein a trigger (e.g., the event associated with a hand gesture) may be linked to an action (e.g., turn on/off TV) via a recipe (instructions configured to cause transmission of IR pulses).
Subsequent to training the device 320 to detect an indication (e.g., a gesture) and providing executable instruction, a user may utilize the apparatus 310 in order to operate a given appliance. The apparatus 320 may detect a gesture indication by the user in video input denoted by arrow 328 in
The system 330 may comprise a learning controller apparatus 340 and a user interface device 332, e.g., a smartphone, a tablet computer and/or other. The device 332 may comprise a display, a touch interface, and a communications component. In one or more implementations, the controller apparatus 340 may comprise the controller 100, 110, 210, 230 described above with respect to
During training, the apparatus 340 may be operable to determine a user indication using sensory input 338 obtained by the sensor component 334. In one or more implementations, sensory input 338 may comprise one or more video image characterizing a gesture (e.g., the user pointing a hand at a ceiling fan). Contemporaneous with the provision of the gesture and/or the sound indication, the user may activate one or more elements of the user interface device 332 (e.g., by touching soft buttons 313, 315). Software application executed by the device 332 may be configured to produce an event comprising an event identification (EventID) associated with activation of a given button, e.g., Button1ID corresponding to activation of the button 313. The EventID may be communicated from the device 332 to the apparatus 340 via a communications link 336. In one or more implementations, the link 336 may comprise one or more of a wired or wireless links (e.g., USB, Ethernet, WiFi, Bluetooth, IR, pulse modulated RF as on the 27 MHz and 49 MHz bands, and/or other communication connections).
Training of determining the associations may comprise multiple repetitions of the gesture/button press combinations and may be effectuated using any applicable methodologies including these described herein (e.g., using a perceptron and/or a look up table).
Subsequent to training the apparatus 340 to determine the associations between a given user indication and the corresponding event, the event may be linked to an action by an appliance device, e.g., as shown and described in the implementation of
In one or more implementations, the linking may be implemented using a configurator application executed by the controller 340, a user interface device 332 of
In one or more implementations, an event from the event list (e.g., the event 502) may first be associated with an action (e.g., the ON action 522). Subsequently, the action may be applied to one or more devices (e.g., the devices 512, 514). For example, a user transitioning from standing to sitting position may be associated with turning on the television, turning on the associated speaker system and dimming lighting. The event-action linking methodology may be configured to enable a given event (e.g., user gesture) that corresponds to (different) contexts that may be associated with (different) actions based on a location of the user providing the gesture. By way of an illustration, a user gesture comprising pointing a hand at a wall proximate a wall switch may be associated with the light turn on/off action; the same pointing gesture away from the wall switch may be associated with no action.
As illustrated in
In some implementations, wherein the component 360 may comprise a general use computing and/or communications apparatus (e.g., a tablet computer, a smartphone, a laptop a desktop computer) and/or a remote computing cloud entity) context determination and/or event-action link operations may be effectuated by an application executed by the general used computing apparatus. By way of an illustration of one exemplary implementation, sensory input 358 may be pre-processed (e.g., down-sampled, and/or compressed) and communicated to a cloud computing entity for analysis. The analysis may provide an event ID based on a match with a previously determined configuration (e.g., LUT).
In some implementations, the component 362 may comprise a dedicated computerized device (e.g., a home automation controller) provided, e.g., as a part of home automation system installation. In one or more implementations, the component 362 may comprise a general use computing and/or communications device (e.g., a tablet computer, a smartphone, a laptop a desktop computer and/or a remote computing cloud computer) operating a home automation software application (e.g., an operating system, a library e.g., the HomeKit™ by Apple, an app, and/or another software configuration).
In one or more implementations wherein the component 362 may comprise a general purpose computing apparatus (e.g., iPad™), functionality of the components 360 and 362 may be implemented by the apparatus 362. By way of an illustration, the iPad apparatus may comprise one or more sensors and be configured to operate a BrainOS-based application configured to enable learning of the association between user indications and events (e.g., as described above with respect to
Based on receiving an event ID, the component 362 may determine an action associated with the given event ID. For example, the component 362 may utilize a table produced e.g., as output of the configuration process described with respect to
By way of an illustration, the user gesture may correspond to the event ID 00001 corresponding to fan ON action ID. Based on the determination of the action ID, the component 362 may communicate a control command to the appropriate destination (e.g., the appliance 364) using communication link 356. The link 356 may comprise wired (e.g., serial bus, Ethernet, 2 wire, controller area network (CAN) bus, and/or other wired link) and/or wireless interface (e.g., Bluetooth, WiFi, 3G-40 cellular, IR, and/or other wireless data transmission link, in one or more implementations.
In some implementations, wherein components 360 and 362 may comprise software applications executed by a computerized device, the link 354 may be effectuated using e.g., a message, a semaphore, a queue, a memory location and/or any applicable data exchange methodology between two or more software applications.
In some implementations, wherein the component 362 may comprise a dedicated computerized device, e.g., a home automation controller provided as a part of home automation (HA) system installation, the configurator application described above with respect to
In one or more implementations, the configuration application may be executed by a remote computing resource (e.g., by cloud computing service) and interfaced to a user via a user interface computing device (e.g., a smartphone).
In one or more implementations the apparatus 410 may comprise a user interface device 310, 332, e.g., a smartphone/tablet computer, comprising a camera, a display, and a touch interface. In one or more implementations (e.g., such as described above with respect to
In one or more implementations, the apparatus 410 may be configured to obtain sensory information related to the task (e.g. the input 108 in
In some implementations wherein operation of the learning process may require computational power in excess of that that may be provided by a processing module of the apparatus 410, one or more portions of the learning process may be operable by the remote computerized entity 406. In one or more implementations, the remote computerized entity 406 may be configured to receive the sensory information (e.g. the input 108 described above with respect to
In
In one or more applications that may require computational power in excess of that that may be provided by a processing module of the apparatus 410, the local computerized interface device 404 may be configured to perform computations associated with context determination, obtaining user instruction, determining of association between the context and the user instructions and/or other operations (e.g., provision of commands to the appliance device). The local computerized interface device 404 may comprise a variety of computing devices including, for example, a desktop PC, a laptop, a notebook, a tablet, a phablet, a smartphone (e.g., an iPhone®), a printed circuit board and/or a system on a chip (SOC) comprising one or more of general processor unit (GPU), field programmable gate array (FPGA), multi-core central processing unit (CPU), an application specific integrated circuit (ASIC), and/or other computational hardware.
In
In
In
A user may utilize the interface 500 in order to perform an action-event association as follows. Upon activating the configuration application, the user may be presented with one or more available events and/or appliances that may be controlled. The user may select an event (e.g., by activating element 502 in panel 501). For the selected event, the user may select an appliance that is to be controlled (e.g., by activating element 514 in panel 510). The configuration application may update the interface 500 to present one or more actions (e.g., ON, OFF, MUTE) that may be performed by the selected appliance (e.g., a TV). The user may select a target action (e.g., the turn TV ON action) by activating the respective element 522. It is noteworthy, that the panel 520 may be updated in accordance with the selected appliance in order to present actions supported by the selected appliance.
In some implementations, a given event may be linked with an action to be performed by a plurality of devices (e.g., power off all components of a home entertainment system).
In one or more implementations, the user interface 530 of
In
In
In
The configuration application may update the device panel (e.g., 550, 551) of the interface 530 in order to match appliances to the respective action. Representations for unavailable (e.g., powered off and/or unsupported) appliance devices may be not displayed and/or rendered unselectable (e.g., grayed out). By way of an illustration, based on the user selecting the POWER action (e.g., by activating the element 542) the appliance panel 550 may be configured to display the TV appliance icon 552 and the DVD appliance icon 554. Based on the user selecting the PLAY action (e.g., by activating the element 544) the corresponding appliance panel 551 may be configured to display the DVD appliance icon 555. The user may select one or more target appliance (e.g., the TV, the DVD) by activating the respective element(s) 552, 554.
A user may utilize the interface 530 of
The controller apparatus may be incorporated into existing user premises infrastructure. In some implementations, the learning apparatus 610 may comprise the apparatus 210 of
The computerized furniture appliance 620 may comprise a motor actuator configured to operate one or more components of the appliance (e.g., slide a drawer 622, open/close a door/lid, adjust shelve height and/or orientation, and/or perform other actions). It will be appreciated by those skilled in the arts that although a drawer operation is illustrated in
The apparatus 620 may comprise a receiver component configured to detect remote transmissions 606 in
The sensor component 602 may be configured to provide sensory input to the learning controller. In some implementations, the component 602 may comprise a camera, a radio wave sensor, a light sensor, an ultrasonic sensor, and/or other sensor capable of providing information related to operation of the appliance apparatus 620.
An exemplary training operation of the system 600 comprising a camera sensor component 602 may be described as follows, in some implementations.
The appliance 620 and the camera component 602 may be arranged such that to enable the camera to obtain digital images of the appliance and the user with sufficient spatial and/or temporal resolution. In one or more implementations, spatial resolution may be characterized by image size selected between 60×80 pixels to 2000×4000 pixels. Greater image pixel count may be employed provided computational and/or energy use capabilities of the processing electronic may accommodate increased pixel count. In some implementations of home and/or vehicle automation, temporal resolution may be characterized by frame rate selected between 2 frames per second (fps) and 100 fps. It will be appreciated by those skilled in the arts that lower or greater frame rates may be employed in accordance with parameters of a specific application (e.g., rate of change of context) may utilize a remote controller
The user may employ a remote control device in order to communicate instructions to the apparatus 620 via, e.g., transmissions 606. In one or more implementations, the transmissions 606 may comprise infrared waves, pressure waves, radio waves, visible light, and/or combination thereof. The appliance apparatus 620 may comprise a detector component configured consistent with the transmission carrier used to communicate transmissions 606.
In one or more implementations, the transmissions 606 may comprise infrared waves, pressure waves, radio waves, visible light, and/or combination thereof. The apparatus 620 and 610 may comprise a detector component configured consistent with the transmission carrier used to communicate transmissions 606.
The user may utilize posture (denoted by icons 614, 644 in
During operation, the apparatus 640 may detect sensory context change. By way of an illustration, the apparatus 640 may determine user position change from a lean-in posture (depicted by icon 614 in
The controller apparatus 710 may comprise a sensor component. In one or more implementations, the sensor component may comprise a camera sensor 712 (described above with respect to
The controller apparatus 710 may comprise a data communications interface configured to communicate instruction (depicted by arrow 736) related to operation of the apparatus 720. In one or more implementations, the communications interface may comprise a wireless interface utilizing one or more of e.g., radio frequency (RF), infrared (IR), pressure (sound), light, and/or other carrier waves. In some implementations, (e.g., such as described above with respect to
In some implementations, the transmissions 736 may be detected by the apparatus 720 (e.g., a receiver embodied within the appliance). In one or more implementations (e.g., such as described above with respect to
The system 700 may be employed by a user (depicted by icon 704 in
The user may alter sensory context and provide an indication of an action. In one or more implementations, the context change may comprise a user gesture, posture change (e.g., lean towards the appliance 720), a sound, an audible signal (click, clap), and/or other indication. The controller apparatus 710 may detect the context change using the sensor component (e.g., a camera of a smartphone). In some implementations, the context change may be detected using pixel wise discrepancy estimation methodology, e.g., such as described below with respect to FIG. CC. In one or more implementations, the action indication may comprise activation of a remote controller device (e.g., pressing left/right button on an RF/IR remote controller), activation of a GUI element (e.g., pressing a soft button on the phone touch screen), physically causing the desired movement of the device if it has back-driven motor feedback (e.g., opening a drawer and/or a curtain) and/or performing other action. In some implementations, wherein the action indication may be communicated by the user via the remote controller, the apparatus 710 may comprise a receiver component configured of detect signal transmissions of appropriate type (e.g., RF/IR).
The user may repeat several times the combination of the context change/action indication. The apparatus 710 may operate a learning process configured to associate the action indication with the context change using any applicable methodology (e.g., a perceptron artificial neuron network, a LUT, a classifier, a K-nearest neighbors (K-NN) algorithm and/or other approach).
The learning process may be configured to associate the action indication with an action by the apparatus 720. By way of an illustration, the button press (associated with the transmission 706) may be configured to cause drawer opening of the furniture appliance. In some implementations, wherein the transmission 706 comprises an RF/IR transmission effectuated by a remote control device associated with the appliance, the appliance may detect the transmission (e.g., using receiver component embodied therewith) and perform respective action (e.g., open the drawer).
In some implementations (e.g., such as shown and described with respect to
In one or more implementations, the controller may be configured (prior to training or subsequent to training) to training to store action-indication associations (mapping). By way of an illustration, a user may utilize methodology described above with respect to
Based on the detection action indication (e.g., using transmission 706) and the action indication-action mapping, the controller apparatus 710 may communicate a respective command for the apparatus 720 (either directly or via a home automation controller).
In some implementations wherein operation of the controlled device may comprise physical manifestations (e.g., drawer opening/closing) that may interfere with the context determination (e.g., user posture detection), the controller apparatus 710 may be configured not to relay the command to the apparatus 720 so as to reduce potential for interference with the context determination process.
In one or more implementations, wherein the controller apparatus 740 may be configured to communicate with an intermediary device (e.g., a home automation controller, an appliance controller, a protocol converter, and/or other apparatus and/or software process) the instruction 736 may comprise, e.g., a message comprising the appliance identification (e.g., NIGHT_STAND) and the action (e.g., CLOSE) communicated to, e.g., the home automation controller. Upon receipt of the instruction, the intermediate entity (e.g., home automation controller (e.g., the component 362 in
System 800 shown in
In one or more implementations, the apparatus 810 may comprise user premises equipment (e.g., a security camera) adapted to execute the learning process instructions. The controller apparatus 810 may comprise a sensor component, e.g., a camera described above with respect to
The system 800 may be employed by a user in order to train the controller apparatus 810 to operate the computerized apparatus 804 (e.g., turn on/off light) based on occurrence of a context change. In one exemplary implementation, the user may employ the system 800 to activate the apparatus 804 (turn on the light) responsive to opening of a book. An exemplary training operation may be performed as follows:
The user may alter sensory context and provide an indication of an action. In one or more implementations, the context change may comprise opening of a book 802. In one or more implementations, context (e.g., book opening) may be configured based on a distance (e.g., 0.5 m to 3 m) between the book and the controlled apparatus (e.g., the lamp 804) and/or location of the book being opened. By way of an illustration of operating a trained controller in a library environment, a user opening a book at or near a work desk may trigger the desk lamp, while another user browsing (opening) books near a bookshelf may not trigger the desk lamp. The controller apparatus 810 may detect the context change using the sensor component. In some implementations, the context change may be detected using pixel-wise discrepancy estimation methodology, e.g., such as described below with respect to
The user may repeat several times the combination of the context change/action indication. The apparatus 810 may operate a learning process configured to associate the action indication with the context change using any applicable methodology (e.g., a perceptron artificial neuron network, a LUT, a classifier, a K-NN process, and/or other approach).
The learning process may be configured to associate the action indication with an action by the apparatus 804 using any applicable methodology described herein including, e.g.,
In one or more implementations, wherein the controller apparatus 840 may be configured to communicate with an intermediary device (e.g., a home automation controller, an appliance controller, a protocol converter, and/or other apparatus and/or software process) the instruction 836 may comprise, e.g., a message comprising the appliance identification (e.g., NIGHT_LAMP) and the action (e.g., ON) communicated to, e.g., the home automation controller. Upon receipt of the instruction, the intermediate entity (e.g., home automation controller (e.g., the component 362 in
Methodology described herein may be utilized in home automation applications.
The appliance apparatus may comprise a camera 920 configured to provide sensory input related to environment within the room containing the appliance 916 to the learning process of the apparatus 910. The camera 920 may be characterized by field of view 914. Images provided by the camera 920 may be utilized by the learning process to determine changes in the context associated with the environment within the view field 914. In some implementations context change determination may be effectuated using other data e.g., time of day/year, day of the week, weather, daylight, sound, user(s) content preferences, state of other equipment within the premises (e.g., receiver ON state), and/or other information associated with user premises.
In some implementations, the context information may comprise number, position and/or posture of users. By way of an illustration, a single user watching a movie may elect to suspend (pause) the playback in order to get a drink and/or attend to an issue outside the room. Pause command issued by the user via the handset 904 may correspond to the following context data: a single user getting up. In one or more implementations, the context may be based on information related to presence of a person obstructing view of TV screen.
In some implementations, the context may comprise information related to weather, time of day, day of the week and/or year, number of people in the room, identity of a person (e.g., a male adult versus a child), content being displayed, and/or other information. A given context may be associated with a respective control command(s) produced by the apparatus 910. For example, context comprising representation of a male adult user may cause commands to switch the TV to a sports channel while a representation of a child may correspond to a command to switch the TV to a cartoon channel. In some implementations of multi-screen video projection devices (e.g., virtual and/or physical multi-screen TV, tablets, and/or computer monitors), users may configure content for individual screens depending on time of day, day of week, weather, and or other parameters (e.g., occurrence of a sports and/or cultural event). In some implementations, the content may be configured based on presence and/or absence of one or more objects in a room: e.g., presence of a toy (e.g., from a Toy Story cartoon) character in the room may cause selection of a Disney channel and/or related TV channel.
In some implementations, the context may comprise user uses gestures that may be provided via Microsoft Kinect and/or other visual motion and position detection system. In one or more implementations, a user may utilize language commands that may be converted into some representation (e.g., a hash, a voiceprint), and used as a sensory context. Individual words of language commands (spoken language tags) may have a meaning associated therewith, and/or may be meaningless (in a given language) provided the spoken language tags consistently accompany a given action by the robotic device.
The context information (e.g., context change) may be associated with one or more actions by the apparatus 910. The association may be based on operation of a neuron network (e.g., effectuating a perceptron), a look up table and/or other classification approach, e.g., look up table.
In some implementations, a perceptron may comprise ANN network comprising one or more output artificial neuron connected to a plurality of input artificial neurons via an array (layer) of efficacies. Efficacies of the connecting layer may be adapted using e.g., correlative and/or error-based gradient descent learning process in order attain target output. Individual neurons may be operable in accordance with an arctangent, sigmoid, and/or other output generating process. The output neuron(s) may compete with one another in a softmax winner-take-all process to produce a given output.
In some implementations of controlling a digital video recorder/playback apparatus (e.g., TiVo®) coupled to a TV (e.g., via output 908 in
In some implementations wherein the learning apparatus 910 may be configured to control other devices, additional output units may be utilized and/or additional perceptron components may be employed.
Various other implementations of predictor/classifier learning process may be employed, such as, e.g., a multilayer perceptron (with a hidden layer), a K-NN predictor, or other classifier/predictors.
In one or more implementations, the action may comprise an operation executed by the appliance 916 (e.g., volume, channel, playback adjustment) and/or operations by one or more devices in communication with the appliance via interface 908. By way of an illustration, the action may comprise a pause command issued to an external playback/record device.
In one or more implementations, e.g., such as described above with respect to
In some implementations, e.g., such as described in U.S. patent application Ser. No. 14/244,888, entitled “LEARNING APPARATUS AND METHODS FOR CONTROL OF ROBOTIC DEVICES VIA SPOOFING”, filed Apr. 3, 2014; Ser. No. 14/244,892 entitled “SPOOFING REMOTE CONTROL APPARATUS AND METHODS”, filed Apr. 3, 2014; and Ser. No. 14/244,890, entitled “APPARATUS AND METHODS FOR REMOTELY CONTROLLING ROBOTIC DEVICES”, filed Apr. 3, 2014, each of the foregoing previously incorporated herein by reference in their entirety supra, the context change may be associated with an action by the appliance 916 using remote control device 904.
In some implementations, user voice commands may be combined with user actions via a remote control (e.g., in
The apparatus 910 may be configured to provide control instructions via data link 908 to one or more home appliances disposed remote from the apparatus 916. In one or more implementations, the data link 908 may comprise a wireless (e.g., Bluetooth, Wi-Fi, 3G-4G cellular, and/or wired link (e.g., serial, Ethernet and/or other link implementation).
Subsequent to learning the associations, upon detecting occurrence of a given context, the apparatus 910 may issue control instructions to the appliance apparatus 916 that may be associated with the detected context. For example, upon detecting that the user stood up (using sensory input from the camera 920) the apparatus 910 may issue commands to the apparatus 916 to pause content playback; upon detecting an adult user in the room at 6 pm during a week day the apparatus 910 may issue commands to display one or more news stations on one or more screens of the apparatus 916; upon detecting a change in weather (using e.g., a news feed, wired and/or wireless sensor component) the apparatus 910 may issue instructions to a remote window actuator via link 908 to close window, and/or instructions to an irrigation controller to suspend watering. It will be recognized by those skilled in the arts that the learning controller 910 may be employed to learn to operate other home appliances, such as, e.g., HVAC system, fan, heater, humidifier, sound system, security system, furniture, assisted living devices (e.g., motorized bed), and/or other controllable devices.
The processing component may interface with the mechanical 1018, sensory 1020, electrical 1022, power 1024, and user interface 1018 components via driver interfaces and/or software abstraction layers. Additional processing and memory capacity may be used to support these processes. It will be appreciated that these components may be fully controlled by the robotic brain. The memory and processing capacity may aid in management of learning configuration (e.g., loading, replacement), operations during a startup, and/or other operations. Consistent with the present disclosure, the various components of the device may be remotely disposed from one another, and/or aggregated. For example, the learning configuration may be executed on a server apparatus, and control the mechanical components via network or radio connection while memory or storage capacity may be integrated into the brain. Multiple mechanical, sensory, or electrical units may be controlled be a learning apparatus via network/radio connectivity.
The mechanical components 1018 may include virtually any type of device capable of motion or performance of a desired function or task. These may include, without limitation, motors, servos, pumps, hydraulics, pneumatics, stepper motors, rotational plates, micro-electro-mechanical devices (MEMS), electroactive polymers, and/or other mechanical components.
The user interface components 1018 may comprise virtually any type of device capable of registering inputs from and/or communicating outputs to a user. These may include, without limitation, display, touch, proximity sensitive interface, light, sound receiving/emitting devices, wired/wireless input devices and/or other devices.
The sensory devices 1020 may enable the learning apparatus to accept sensory input from external entities. These may include, without limitation, video, audio, capacitive, radio, vibrational, ultrasonic, infrared, and temperature sensors radar, lidar and/or sonar, and/or other sensory devices.
The electrical components 1022 may include virtually any electrical device for interaction and manipulation of the outside world. This may include, without limitation, light/radiation generating devices (e.g. LEDs, IR sources, light bulbs, and/or other devices), audio devices, monitors/displays, switches, heaters, coolers, ultrasound transducers, lasers, and/or other electrical components. These devices may enable a wide array of applications for the robotic apparatus in industrial, hobbyist, building management, medical device, military/intelligence, and other fields (as discussed below).
The communications interface 1026 may include one or more connections to external computerized devices to allow for, inter alia, management of the robotic device, e.g., as described above with respect to
The power system 1024 may be tailored to the needs of the application of the device. For example, for a small-sized lower power appliance, a wireless power solution (e.g. battery, solar cell, inductive (contactless) power source, rectification, and/or other) may be appropriate. For building management applications, battery backup/direct wall power may be superior. In addition, in some implementations, the power system may be adaptable with respect to the training of the appliance apparatus (e.g., 720 in
In some implementations, methods 1100, 1120, 1200, 1220, 1240, 1300, 1400, 1500 may be implemented in one or more processing devices (e.g., a digital processor, an analog processor, a digital circuit designed to process information, an analog circuit designed to process information, a state machine, and/or other mechanisms for electronically processing information and/or execute computer program components). The one or more processing devices may include one or more devices executing some or all of the operations of methods 1100, 1120, 1200, 1220, 1240, 1300, 1400, 1500 in response to instructions stored electronically on an electronic storage medium. The one or more processing devices may include one or more devices configured through hardware, firmware, and/or software to be specifically designed for execution of one or more of the operations of methods 600, 700, 800, 820, 840, 900. The operations of methods 1100, 1120, 1200, 1220, 1240, 1300, 1400, 1500 may be implemented by a learning controller apparatus (e.g., 110 in
At operation 1102 of method 1100, illustrated in
At operation 1104 sensory input conveying context associated with environment of the apparatus and occurring contemporaneous with the transmissions may be analyzed. In one or more implementations, such as gesture recognition, the sensory input may be provided by a sensor component of the learning controller (e.g., camera 352 of the apparatus 360 in
In some implementations, the context determination may be configured based on detection of a sudden change (e.g., within a time window selected between 1 seconds and 10 seconds) of a given sensory characteristic (e.g., image pixel change in color, brightness, local contrast, local binary pattern, sound level, and/or other parameter). In one or more implementations, e.g., such as described with respect to
Advantageously, such configuration may enable continuing functioning of the learning controller when relocated (e.g., rotated or translated), as the previously acquired (reference) image of the environment may be utilized in order to be used to compensate for the relocation and prior learned associations to continue to function as desired/expected by the user. By way of an illustration of a learning controller embodied in a light fixture (e.g., 610 of
In some implementations wherein user privacy may be of concern, the learning controller may be embodied in a mobile platform (rover). Upon a request from a user, the rover may enter a given room, acquire image input, activate/deactivate one or more appliances, devices, and be dismissed (with image acquisition disabled) if privacy is desired. The learning controller may operate a Wi-Fi on/off switch in order to connect/disconnect the device from the Internet and ensure privacy of video capture.
In one or more implementations, the reference state (the image I0) may be determined based on an output of a high pass filter of the image sequence. The filter may comprise an exponential filter operation, e.g., described with respect to
At operation 1106 associations may be developed between the control instructions determined at operation 1102 and the corresponding context. In one or more implementations, the associations may be configured based on one or more LUT characterizing the relationship between sensory input (e.g., user gesture) and the control instruction (e.g., POWER ON) obtained by the learning controller during training.
At operation 1108 the association information may be stored. In some implementations, the information storing of operation 1108 may comprise storing one or more entries of a LUT in internal memory of the learning controller apparatus (e.g., the memory 1012, 1014 in
In some implementations, it may be of use (e.g., due to a development of an erroneous association and/or given functionality no longer needed) to enable the learning controller to selectively not to trigger a given action (e.g., to selectively not turn on TV) based on occurrence of corresponding context (a given person entering a room). In some implementations, the controller learning process may be trained to select no-action output with the given context by, e.g., utilizing a GUI element configured to convey the DO NOTHING or NO ACTION indication. By way of an illustration, a user may utilize a remote control device (e.g., 202 in
The learning controller may comprise functionality configured to enable removal of no longer needed learned associations by “unlearning/‘forgetting’ respective association between the context (e.g., person entering the room) and the action (turn TV ON) while preserving other trained associations (e.g., turn lights ON).
In some implementations, the association removal may be effectuated using a “remove” GUI element, implemented e.g., by the user interface device 310, 320, and/or 710 of
When a given context occurs (e.g., a person entering a room), the user may activate the respective action (turn TV ON) indication element (e.g., the button 322 of the apparatus 320 in
In one or more implementations, a user may be provided with a graphical interface configured to enable removal of associations from the learned configuration implemented e.g., by the user interface device 310, 320, and/or 710 of
At operation 1124 of method 1120 a context may be determined using any applicable methodology described herein. In some implementations, the context may comprise user posture (e.g., 702 in
At operation 1126 a determination may be made as to whether an unlearn indication present. In some implementations, the unlearn (also referred to as undo) indication may comprise a transmission associated with a button press of a GUI device. In one or more implementations, the unlearn indication may be configured based on activation of unlearn component of a GUI.
Responsive to determination at operation 1126 that the unlearn indication may be present, the method may proceed to operation 1128 wherein context-action association corresponding to the context may be determined. In some implementation wherein a single action may correspond to a given context (e.g., turn TV ON when a given person enters the room), the action determination (e.g., turn TV ON) may be based on identification of the context determined at operation 1124. In one or more implementation wherein multiple actions may correspond to a given context (e.g., turn TV ON and turn light ON when the person enters the room), action identification may be configured based on one or more of (i) action indication provided to the learning process (e.g., activating the button 322 of the apparatus 320 described with respect to
In some implementations of a LUT-based learning process configured to provide a plurality of outputs (e.g., in absence of a winner takes all competition process) an entry in the LUT associated with the context determined at operation 1124 may be removed, a score corresponding to the [context+control instruction] combination may be decremented. In one or more implementations of a neuron network-based learning process configured to provide a plurality of outputs (e.g., in absence of a winner takes all competition process), efficacy of connections associated with outputting the instruction determined at operation 1122 may be reduced.
In one or more implementations of a LUT-based learning process configured to provide a specific output (e.g., using a winner takes all competition process) an entry in the LUT associated with a competing action for the context determined at operation 1124 may be added, a score corresponding to the [context+competing action] combination may be incremented. In one or more implementations of a neuron network-based learning process configured to provide a given outputs (e.g., based on a winner takes all competition process), efficacy of connections associated with outputting the a NULL command may be increased
In one or more implementations, “SEND NOTHING” indication may be utilized. By way of an illustration, a user interface element (e.g., unlearn button) may be provided to indicate to the learning process to ignore a particular input. In some implementations, the unlearn button may be configured to cause the learning process to unlearn sending any action, unlearn sending the most recent action and/or another action composition. The unlearn functionality may enable operation of the learning process to remove false positives cannot be untaught. This helps obviate the disadvantage of false positives, erroneous and/or no longer needed associations and/or other actions.
At operation 1130 learning process may be adapted so as to inhibit execution of the action corresponding with the context-action association. In some implementations, the inhibition may be configured based on potentiation of the NO ACTION output configured to cause the learning process to activate NO ACTION output based on detection of the context. In some implementations, the learning process adaptation may comprise depression of the output associated with the action by e.g., reducing efficacy of the ANN nodes associated with enabling the TV ON action output or decrementing the respective entry in the LUT.
In some implementations, wherein a controlled appliance may be operable in a mode wherein it may expect a command, the NULL association may cause transmission of a NULL command (e.g., a packet with empty payload). Upon detecting such command/packet, the appliance may take no action (e.g., the TV controller may continue listening to commands without altering its state). In one or more implementations, wherein a controlled appliance may be operable via a message queue (e.g., respond when there is a command, otherwise perform no action), the NO ACTION association may cause no command transmission to the transmission to the appliance. At operation 1132 the adapted configuration of the learning process may be stored. In some implementations, e.g., such as illustrated with respect to
Operations of method 1120 may be used to implement “undo” functionality. In some implementations, upon receiving the undo indication the learning process may be operable to unlearn (e.g., suppress) the most recently learned context-action association (remove learned “robotic behavior”). In one or more implementations, the undo indication may cause removal of learned behaviors/associations during a given period of time. Specific associations may be indicated to the learning process by providing the relevant context (e.g., gesture). By way of an illustration, upon realizing that the controller learned an erroneous association causing an undesired behavior of an appliance (e.g., pausing video playback whenever one user may raise an arm while stretching) the user may issue the undo indication. Subsequently, during a given interval, the user may raise an arm to provide the context. The learning process may determine one or more actions (commands) that may be associated with that context (e.g., user gesture of raising an arm). The most recent action from the one or more of the actions associated with that context may be suppressed (removed) using methodology described herein. In one or more implementations, all actions associated with that context may be suppressed (removed).
The interval during which context for unlearning associations may be provided may be selected between 1 s and 300 s. In some implementations, the interval may be selected arbitrarily long (e.g., until another undo indication may be provided to indicate end of the undo interval). In one or more implementations, the interval may be configured based on an amount of ambient sensory input that may have been analyzed during context detection during training. For example, in some implementations, wherein during training, a substantial amount of background movement and/or sounds may have been present (e.g., during a party with 7 or more people), undo interval may be increased commensurate with the duration of the training.
At operation 1202, context may be determined. In some implementations, the context determination may be configured based on analysis of sensory input. In some implementations, the sensory input may comprise video image sequence and/or audio signal. Analysis may comprise, e.g., operations described with respect to
At operation 1204 user indication associated with the context of operation 1202 may be detected. In one or more implementations, the indication may comprise activation of a user interface element (e.g., GUI button 312, 314 in
In some implementations of training, the user indication may be provided contemporaneously (e.g., within one-2 seconds) with the context (e.g., the user uses one arm to produce a gesture and another hand to press a button on a user interface device). In some implementations, different physical users (referred colloquially as the user) may cooperate to provide the context (e.g., a gesture) and the indication (e.g., the button press). In some implementations, different physical users (referred colloquially as the user) may cooperate to provide the context (e.g., a gesture) and the indication (e.g., the button press). The user indication may closely follow the context and/or context change. In one or more implementations (e.g., described with respect to Eqn. 1) wherein the reference sensory input may comprise output of a filter characterized by a time constant Tf, the user indication may follow the context within the window duration. During operation of the trained controller, the action (e.g., command transmission 216 in
At operation 1206 a determination may be made as to whether an association between the context and the indication may exist. In some implementations of the learning process configured based on use of a LUT, a row or a column element of the LUT corresponding to the context may be accessed in order to determine as to whether the given context has previously occurred. An entry (e.g., a cell) of the selected row/column LUT element may be accessed in order to determine as to whether the given indication has previously occurred for the given context.
Responsive to the determination at operation 1206 that the association may exist, the method may proceed to operation 1208 wherein a tag may be produced. The tag may be used to convey the existence and/or identification of the association between the context and the indication.
At operation 1222 one or more tags may be retrieved. In some implementations, the tag retrieval may comprise presentation of the one or more tags (e.g., EventIDs) by a graphical user interface, e.g., the event list 501 comprising tags 502, 504, 506, that may have been associated previously (e.g., the events produced by pressings of buttons 313, 315 in
At operation 1224 a set of instructions may be provided for a given tag. In some implementations, the instructions may be provided by specifying an appliance and a respective command for the appliance, (e.g., DVD player 555 and play command 544 in
In some implementations, wherein a learning controller process (e.g., 214 in FIG. 2A) may be configured to communicate to an intermediate device, (e.g., home automation controller 362 and/or home automation application) the instructions when executed by a processor may be configured to cause transmission of a command to the automation controller 362 and/or home automation application.
At operation 1224 one or more sets of instructions provided at operation 1224 and associated with the one or more tags may be stored in a non-volatile medium. In some implementations, the instructions may be stored in memory of a local learning controller apparatus (e.g., 210 in
At operation 1242 an event tag may be associated with an action. In some implementations, the tag-action association may be effectuated via a GUI, e.g., as described with respect to
At operation 1243 context in sensory input may be determined. In some implementations, the context determination may be effectuated based on analysis of a plurality of images, e.g., using functionality described with respect to
At operation 1244 a determination may be made as to whether an association between the context and the tag may exist. In some implementations of the learning process configured based on use of a LUT, a row or a column element of the LUT corresponding to the context may be accessed in order to determine as to whether the given context has previously occurred. An entry (e.g., a cell) of the selected row/column LUT element may be accessed in order to determine as to whether the given tag has previously occurred for the selected context.
Responsive to the determination at operation 1244 that the association may exist, the method may proceed to operation 1246 wherein a set computer readable instructions corresponding to the tag may be retrieved. In some implementations, execution of instructions retrieved at operation 1246 may be configured to communicate a command to the computerized appliance apparatus.
At operation 1248 the command may be communicated to the apparatus thereby causing the apparatus to execute an action in accordance with the context detected at operation 1243. In some implementations, the action may comprise one or more actions described with respect to
At operation 1302 of method 1300 illustrated in
At operation 1304 context corresponding to the transmissions described with respect to operation 1302 may be determined. In one or more implementations, the context may be determined based on analysis of sensory input, e.g., the input described above with respect to
At operation 1306 an association between the context and the respective remote control instruction determined at operation 1302 may be developed. In some implementations, the association may be produced based on one or more LUT characterizing the relationship between the context (e.g., user rising the right hand) and control signal (e.g., RF code corresponding to POWER ON) obtained by the learning controller during training. In one or more implementations, the associations may comprise a trained configuration of a network of artificial neurons configured to implement an adaptive predictor and/or combiner of the control system described, e.g., in U.S. patent application Ser. No. 13/842,530 entitled “ADAPTIVE PREDICTOR APPARATUS AND METHODS”, filed Mar. 15, 2013; and Ser. No. 13/842,562 entitled “ADAPTIVE PREDICTOR APPARATUS AND METHODS FOR ROBOTIC CONTROL”, filed Mar. 15, 2013, each of the foregoing previously incorporated herein by reference in their entirety supra.
At operation 1308, based on the association developed at operation 1306 and on detecting the context occurring at operation 1304, a second transmission of a second wireless carrier may be automatically produced. Second carrier transmissions may comprise IR, RF (e.g., configured using Direct Sequence Spread Spectrum (DSSS), and/or frequency hopping spread spectrum” (FHSS)), pressure, and/or other technology may be utilized. The second transmission may be configured communicate one or more instructions to the appliance apparatus, (e.g., the fan 220 in
At operation 1402 of method 1400 sensory input may be analyzed. In one or more implementations, the sensory input may comprise output of a video camera, audio, electromagnetic, pressure wave sensor array, a combination thereof, and/or other data (e.g., time of day/week/year, motion parameters provided by, e.g., an inertial measurement component. In one or more implementations the sensory input may be provided by a sensor module of the learning controller (e.g., 212 in
At operation 1404 a user action may be determined. In some implementations, the user action may comprise a gesture, a posture change, a voice command, an eye movement, an audible indication, activation of a remote control device (e.g., and/or other 202, 320 in
At operation 1406 a determination may be made as to whether the action determined at operation 1404 has previously occurred and an association exists for the user action. The association may comprise a relationship between the user action and one or more control instructions configured to cause the computerized appliance to execute an operation (e.g., turn on/off, open close). In one or more implementations, determination as to whether the association exists may be based on an analysis of a LUT configured to store associations between the action and the corresponding control instruction. In some implementations of the learning process configured based on use of a LUT, a row or a column element of the LUT corresponding to the user action may be accessed in order to determine as to whether the given action has previously occurred. An entry (e.g., a cell) of the selected row/column LUT element may be accessed in order to determine as to whether one or more control instructions have been previously associated with the given action.
Responsive to determination at operation 1406 that the association may exist, the method 1400 may proceed to operation 1408 wherein one or more control instructions associated with the user action may be retrieved. In some implementations, wherein protocol specification of the control communication between the user handset (e.g., 202 in
At operation 1410 the one or more control instructions determined at operation 1408 may be automatically provided to the appliance thereby enabling execution of the operations (e.g., closing the drawer by the night stand 620 in
At operation 1502 reference instance I0 of sensory input may be obtained. In some implementations, the reference instance may comprise an initial image (e.g., shown in
I0=ΣaiI(n−i−1) (Eqn. 1)
where ai denote filter coefficients. In some implementations, the filter of Eqn. 1 may comprise an exponential filter with half-life decay window selected between 1 and 10 second for image sequence refreshed at 10 fps.
At operation 1504 Obtain current instance of sensory input I(n) may be obtained. In some implementations wherein camera sensor may comprise a general purpose camera (e.g., embodies in a smartphone 310 of
At operation 1506 a discrepancy map between the currents instance of sensory input and the reference instance may be determined. In some implementations of image analysis, the discrepancy map Ir may be configured based on a pixel-wise difference of the current image and the reference image I0 as follows:
Ir(n)=I(n)−I0 (Eqn. 2)
At operation 1508 most relevant (also referred to as salient) feature may be determined based on an analysis of the discrepancy map. In some implementations, the discrepancy map analysis may comprise cluster analysis (e.g., comprising thresholding and determination of largest contiguous (in pixel space) portion and/or portion with the greatest cumulative pixel value, and/or other techniques). The discrepancy map analysis may comprise a winner takes all (WTA) evaluation wherein a single portion may be selected (e.g., the largest area). In some implementations, the discrepancy map analysis may comprise a signal range squashing operation (e.g., based on a sigmoid, arctangent, and/or other operation configured to transform wide range input into a bounded range output). In some implementations, the discrepancy analysis may comprise a soft-max operation expressed as:
F=max(S(Ir(n))), (Eqn. 3)
wherein F denotes feature output, and S denotes a sigmoid operation.
At operation 1510 a hash representation of the map representing the relevant feature may be determined. In one or more implementations, the hash representation may be determined based on size, location, and/or energy content of the detected relevant feature. Hashing of feature representations may enable more economical storage and/or faster retrieval of features during training and/or operation of the learning process.
At operation 1512, hash representation may be stored, in, e.g., memory of the learning controller (e.g., 1012, 1014 in
Various image sources may be utilized with the context detection methodology described herein such as digital images comprising plurality of pixels provided by a video camera (e.g., 212 in
A reference image may be selected for use by the context detection process. In some implementations, the learning process may be configured to periodically acquire a reference image of the environment.
In one or more implementations, the reference image I0 may be determined based on an output of a high pass filter of the image sequence. The filter may comprise an exponential filter operation characterized by decay half life time window selected between 1 and 10 seconds.
During operation of context detection process, a given image may be compared to the reference image in order to determine the context based on changes in sensory input.
As shown in
Context detection methodology shown described with respect to
In one or more implementations, data from two or more sensor types may be combined (e.g., via an adaptive data fusion process). By way of an illustration, false positive of features detected using video, may be filtered out using audio, proximity, and/or other signal provided in order to disambiguate context detection.
The methodology described herein may advantageously enable automatic operation of computerized appliances a trained controller apparatus. The learning controller may automatically trigger an action execution by the appliance based on detecting a context that may have been associated with the action during training. In one or more implementations, the action execution may be effectuated based on provision of one or more control instructions to the appliance and/or to an intermediary device in data communication with the appliance.
Use of computerized controllers to automatically operate appliances may obviate need to users to perform mundane tasks (e.g., closing of doors/windows in bad weather), configuring home entertainment system in accordance with detected audience (e.g., based on age, sex, and/or other characteristics), allow users to perform other tasks without interruptions (e.g., automatic TV muting may enable the user to verify caller ID and answer a phone call before it goes to voice mail). Use of training, may enable users to operate home appliances by using gestures, posture, location, sounds, instead of and/or in addition to remote control devices provided by manufacturers. Use of computerized controllers for robot operation may afford users with added functionality that may not have been available otherwise.
In some implementations, a given learning controller (e.g., 210 in
Multiple learning controllers may be configured to operate a given computerized appliance and/or robotic device. In one or more implementations, multiple learning controllers may be disposed in different rooms of a house. The learning controller disposed in a given room may provide control instructions to a robotic appliance (e.g., vacuum cleaner) operating in that room. When a robotic appliance (e.g., vacuum cleaner) moves to another room, another learning controller disposed in that other room may take over the control operations of the appliance. In some implementations, two or more learning controllers may cooperate when operating a given appliance (e.g., the vacuum cleaner). By way of an illustration, one learning controller may collect and process sensory data, while another learning controller may communicate commands to the appliance, one controller may provide processing resources to another controller while processing sensory input (e.g., video). Various other configurations may be employed (e.g., one controller processing video while another controller processing audio sensory data).
In some implementations of a learning controller embodied in a portable user interface device (e.g., 710/740 of
The methodology described herein may be applicable to perform operations with a variety of household, office, industrial and/or other appliances and/or robotic devices. Various tasks may be executed using trained controllers described herein, including, e.g., cleaning, object removal/sorting/rearrangement, content recording/playback, premises security, premises maintenance, and/or other tasks.
It will be recognized that while certain aspects of the technology are described in terms of a specific sequence of steps of a method, these descriptions are only illustrative of the broader methods of the technology, and may be modified as required by the particular application. Certain steps may be rendered unnecessary or optional under certain circumstances. Additionally, certain steps or functionality may be added to the disclosed implementations, or the order of performance of two or more steps permuted. All such variations are considered to be encompassed within the technology disclosed and claimed herein.
While the above detailed description has shown, described, and pointed out novel features of the technology as applied to various implementations, it will be understood that various omissions, substitutions, and changes in the form and details of the device or process illustrated may be made by those skilled in the art without departing from the disclosure. The foregoing description is of the best mode presently contemplated of carrying out the technology. This description is in no way meant to be limiting, but rather should be taken as illustrative of the general principles of the technology. The scope of the technology should be determined with reference to the claims.
| Number | Name | Date | Kind |
|---|---|---|---|
| 4484120 | Olex et al. | Nov 1984 | A |
| 5046022 | Conway et al. | Sep 1991 | A |
| 5063603 | Burt | Nov 1991 | A |
| 5079491 | Nose et al. | Jan 1992 | A |
| 5092343 | Spitzer et al. | Mar 1992 | A |
| 5245672 | Wilson et al. | Sep 1993 | A |
| 5355435 | DeYong et al. | Oct 1994 | A |
| 5388186 | Bose | Feb 1995 | A |
| 5408588 | Ulug | Apr 1995 | A |
| 5467428 | Ulug | Nov 1995 | A |
| 5638359 | Peltola et al. | Jun 1997 | A |
| 5673367 | Buckley | Sep 1997 | A |
| 5875108 | Hoffberg et al. | Feb 1999 | A |
| 5937143 | Watanabe et al. | Aug 1999 | A |
| 6009418 | Cooper | Dec 1999 | A |
| 6014653 | Thaler | Jan 2000 | A |
| 6363369 | Liaw et al. | Mar 2002 | B1 |
| 6458157 | Suaning | Oct 2002 | B1 |
| 6545705 | Sigel et al. | Apr 2003 | B1 |
| 6545708 | Tamayama et al. | Apr 2003 | B1 |
| 6546291 | Merfeld et al. | Apr 2003 | B2 |
| 6581046 | Ahissar | Jun 2003 | B1 |
| 6643627 | Liaw et al. | Nov 2003 | B2 |
| 7164971 | Ferla et al. | Jan 2007 | B2 |
| 7395251 | Linsker | Jul 2008 | B2 |
| 7426920 | Petersen | Sep 2008 | B1 |
| 7672920 | Ito et al. | Mar 2010 | B2 |
| 7849030 | Ellingsworth | Dec 2010 | B2 |
| 7937164 | Samardzija et al. | May 2011 | B2 |
| 7945349 | Svensson et al. | May 2011 | B2 |
| 7970492 | Matsushima | Jun 2011 | B2 |
| 8015130 | Matsugu et al. | Sep 2011 | B2 |
| 8315305 | Petre et al. | Nov 2012 | B2 |
| 8467623 | Izhikevich et al. | Jun 2013 | B2 |
| 8494677 | Mizutani | Jul 2013 | B2 |
| 8600166 | Adhikari | Dec 2013 | B2 |
| 8990133 | Ponulak et al. | Mar 2015 | B1 |
| 9111441 | Yano | Aug 2015 | B2 |
| 9164596 | Lee et al. | Oct 2015 | B1 |
| 20020038294 | Matsugu | Mar 2002 | A1 |
| 20030050903 | Liaw et al. | Mar 2003 | A1 |
| 20040056625 | Sano et al. | Mar 2004 | A1 |
| 20040102862 | Kato et al. | May 2004 | A1 |
| 20040136439 | Dewberry et al. | Jul 2004 | A1 |
| 20040153211 | Kamoto | Aug 2004 | A1 |
| 20040193670 | Langan et al. | Sep 2004 | A1 |
| 20040225416 | Kubota et al. | Nov 2004 | A1 |
| 20050015351 | Nugent | Jan 2005 | A1 |
| 20050036649 | Yokono et al. | Feb 2005 | A1 |
| 20050071047 | Okabayashi et al. | Mar 2005 | A1 |
| 20050149227 | Peters | Jul 2005 | A1 |
| 20050283450 | Matsugu et al. | Dec 2005 | A1 |
| 20060094001 | Torre et al. | May 2006 | A1 |
| 20060145647 | Kitatsuji et al. | Jul 2006 | A1 |
| 20060161218 | Danilov | Jul 2006 | A1 |
| 20070022068 | Linsker | Jan 2007 | A1 |
| 20070053513 | Hoffberg | Mar 2007 | A1 |
| 20070176643 | Nugent | Aug 2007 | A1 |
| 20070208678 | Matsugu | Sep 2007 | A1 |
| 20070276690 | Ohtani et al. | Nov 2007 | A1 |
| 20070290885 | Tanabe | Dec 2007 | A1 |
| 20080085048 | Venetsky et al. | Apr 2008 | A1 |
| 20080086241 | Phillips et al. | Apr 2008 | A1 |
| 20080100482 | Lazar | May 2008 | A1 |
| 20090043722 | Nugent | Feb 2009 | A1 |
| 20090287624 | Rouat et al. | Nov 2009 | A1 |
| 20100036457 | Sarpeshkar et al. | Feb 2010 | A1 |
| 20100081958 | She | Apr 2010 | A1 |
| 20100086171 | Lapstun | Apr 2010 | A1 |
| 20100100240 | Wang et al. | Apr 2010 | A1 |
| 20100119214 | Shimazaki et al. | May 2010 | A1 |
| 20100166320 | Paquier | Jul 2010 | A1 |
| 20100225824 | Lazar et al. | Sep 2010 | A1 |
| 20110016071 | Guillen et al. | Jan 2011 | A1 |
| 20110119214 | Breitwisch et al. | May 2011 | A1 |
| 20110119215 | Elmegreen et al. | May 2011 | A1 |
| 20110208355 | Tsusaka | Aug 2011 | A1 |
| 20120011090 | Tang et al. | Jan 2012 | A1 |
| 20120079866 | Kuwayama et al. | Apr 2012 | A1 |
| 20120109866 | Modha | May 2012 | A1 |
| 20120171927 | Yu | Jul 2012 | A1 |
| 20120209428 | Mizutani | Aug 2012 | A1 |
| 20120265370 | Kim | Oct 2012 | A1 |
| 20120295662 | Haubrich | Nov 2012 | A1 |
| 20120303091 | Izhikevich | Nov 2012 | A1 |
| 20120308076 | Piekniewski et al. | Dec 2012 | A1 |
| 20120308136 | Izhikevich | Dec 2012 | A1 |
| 20130073484 | Izhikevich et al. | Mar 2013 | A1 |
| 20130073491 | Izhikevich et al. | Mar 2013 | A1 |
| 20130073492 | Izhikevich et al. | Mar 2013 | A1 |
| 20130073495 | Izhikevich et al. | Mar 2013 | A1 |
| 20130073496 | Szatmary et al. | Mar 2013 | A1 |
| 20130073498 | Izhikevich et al. | Mar 2013 | A1 |
| 20130073499 | Izhikevich et al. | Mar 2013 | A1 |
| 20130073500 | Szatmary et al. | Mar 2013 | A1 |
| 20130077831 | Momozono et al. | Mar 2013 | A1 |
| 20130151448 | Ponulak | Jun 2013 | A1 |
| 20130151449 | Ponulak | Jun 2013 | A1 |
| 20130151450 | Ponulak | Jun 2013 | A1 |
| 20130184860 | Ota et al. | Jul 2013 | A1 |
| 20130218821 | Szatmary et al. | Aug 2013 | A1 |
| 20130251278 | Izhikevich et al. | Sep 2013 | A1 |
| 20130272570 | Sheng et al. | Oct 2013 | A1 |
| 20130297542 | Piekniewski et al. | Nov 2013 | A1 |
| 20130310977 | TSUSAKA; Yuko et al. | Nov 2013 | A1 |
| 20130325768 | Sinyavskiy et al. | Dec 2013 | A1 |
| 20130325773 | Sinyavskiy et al. | Dec 2013 | A1 |
| 20130325775 | Sinyavskiy et al. | Dec 2013 | A1 |
| 20130325776 | Ponulak et al. | Dec 2013 | A1 |
| 20130325777 | Petre et al. | Dec 2013 | A1 |
| 20130326839 | Cho | Dec 2013 | A1 |
| 20140012788 | Piekniewski | Jan 2014 | A1 |
| 20140041042 | Wong et al. | Feb 2014 | A1 |
| 20140079297 | Tadayon et al. | Mar 2014 | A1 |
| 20140081895 | Coenen et al. | Mar 2014 | A1 |
| 20140114477 | Sato et al. | Apr 2014 | A1 |
| 20140116469 | Kim | May 2014 | A1 |
| 20150012943 | Mampaey | Jan 2015 | A1 |
| 20160075015 | Izhikevich et al. | Mar 2016 | A1 |
| 20160243701 | Gildert | Aug 2016 | A1 |
| Number | Date | Country |
|---|---|---|
| 102226740 | Oct 2011 | CN |
| H0487423 | Mar 1992 | JP |
| 2108612 | Apr 1998 | RU |
| WO-2008083335 | Jul 2008 | WO |
| Entry |
|---|
| Abbott L. F. and Nelson S.B. (2000), “Synaptic plasticity: taming the beast”, Nature Neuroscience, 3, 1178-1183. |
| Bohte, ‘Spiking Nueral Networks’ Doctorate at the University of Leiden, Holland, Mar. 5, 2003, pp. 1-133 [retrieved on Nov. 14, 2012]. Retrieved from the interne &It;ahref=″http://homepages.cwi.nl/˜sbohte/publication/phdthesis.pdf″>http://homepages.cwi.nl/˜sbohte/publication/phdthesis.pdf&It/a>&It;url: /&g;. |
| Brette et al., Brian: a simple and flexible simulator for spiking neural networks, The Neuromorphic Engineer, Jul. 1, 2009, pp. 1-4, doi: 10.2417/1200906.1659. |
| Cessac et al. ‘Overview of facts and issues about neural coding by spikes.’ Journal of Physiology, Paris 104.1 (2010): 5. |
| Cuntz et al., ‘One Rule to Grow Them All: A General Theory of Neuronal Branching and Its Paractical Application’ PLOS Computational Biology, 6 (8), Published Aug. 5, 2010. |
| Davison et al., PyNN: a common interface for neuronal network simulators, Frontiers in Neuroinformatics, Jan. 2009, pp. 1-10, vol. 2, Article 11. |
| Djurfeldt, Mikael, The Connection—set Algebra: a formalism for the representation of connectivity structure in neuronal network models, implementations in Python and C++, and their use in simulators BMC Neuroscience Jul. 18, 2011 p. 1 12(Suppl 1):P80. |
| Dorval et al. ‘Probability distributions of the logarithm of inter-spike intervals yield accurate entropy estimates from small datasets.’ Journal of neuroscience methods 173.1 (2008): 129. |
| Fidjeland, et al., “Accelerated Simulation of Spiking Neural Networks Using GPUs,” WCCI 2010 IEEE World Congress on Computational Intelligience, Jul. 18-23, 2010—CCIB, Barcelona, Spain, pp. 536-543, [retrieved on Nov. 14, 2012]. Retrieved from the Internet: URL:http://www.doc.ic.ac.ukl-mpsha/IJCNN10b.pdf. |
| Floreano et al., ‘Neuroevolution: from architectures to learning’ Evol. Intel. Jan. 2008 1:47-62, [retrieved Dec. 30, 2013] [retrieved online from URL:http://inforscienee.eptl.cb/record/112676/files/FloreanoDuerrMattiussi2008.pdf<, http: />. |
| Gewaltig et al.. ‘NEST (Neural Simulation Tool)’, Scholarpedia, 2007. pp. I-15, 2(4): 1430, doi: 1 0.4249/scholarpedia.1430. |
| Gleeson et al., NeuroML: A Language for Describing Data Driven Models of Neurons and Networks with a High Degree of Biological Detail, PLoS Computational Biology, Jun. 2010, pp. 1-19 vol. 6 Issue 6. |
| Gollisch et al., ‘Rapid neural coding in the retina with relative spike latencies.’ Science 319.5866 (2008): 1108-1111. |
| Goodman et al., Brian: a simulator for spiking neural networks in Python, Frontiers in Neuroinformatics, Nov. 2008, pp. 1-10, vol. 2, Article 5. |
| Gorchetchnikov et al., NineML: declarative, mathematically-explicit descriptions of spiking neuronal networks, Frontiers in Neuroinformatics, Conference Abstract: 4th INCF Congress of Neuroinformatics, doi: 1 0.3389/conffninf,2011.08.00098. |
| Graham, Lyle J., The Surf-Hippo Reference Manual, http://www.neurophys.biomedicale.univparis5. fr/graham/surf-hippo-files/Surf-Hippo%20Reference%20Manual.pdf, Mar. 2002. pp. 1-128. |
| Harmony Ultimate User Guide, Version 4, pp. 1-54 Oct. 31, 2013. |
| Izhikevich E.M. (2006) Polychronization: Computation With Spikes. Neural Computation, 18:245-282. |
| Izhikevich et al., ‘Relating STDP to BCM’, Neural Computation (2003) 15, 1511-1523. |
| Izhikevich, ‘Simple Model of Spiking Neurons’, IEEE Transactions on Neural Networks, vol. 14, No. 6, Nov. 2003, pp. 1569-1572. |
| Jin, X., Rast, A., F. Galluppi, F., S. Davies., S., and Furber, S. (2010) “Implementing Spike-Timing-Dependent Plasticity on SpiNNaker Neuromorphic Hardware”, WCCI 2010, IEEE World Congress on Computational Intelligence. |
| Karbowski et al., ‘Multispikes and Synchronization in a Large Neural Network with Temporal Delays’, Neural Computation 12. 1573-1606 (2000). |
| Khotanzad. ‘Classification of invariant image representations using a neural network’ IEEE. Transactions on Acoustics, Speech, and Signal Processing, vol. 38, No. 6, Jun. 1990, pp. 1028-1038 [online], [retrieved on Dec. 10, 2013]. Retrieved from the Internet <URL: http://www-ee.uta.edu/eeweb/IP/Courses/SPR/Reference/ Khotanzad.pdf. |
| Laurent, ‘Issue 1—nnql Refactor Nucleus into its own file—Neural Network Query Language’ [retrieved on Nov. 12, 2013]. Retrieved from the Internet: URL:https://code.google.com/p/nnql/issues/detail?id=1. |
| Laurent, ‘The Neural Network Query Language (NNQL) Reference’ [retrieved on Nov. 12, 2013], Retrieved from the Internet: <URL′https://code.google.com/p/ nnql/issues/detail?id=1>. |
| Lazar et a]. ‘Multichannel time encoding with integrate-and-fire neurons.’ Neurocomputing 65 (2005): 401-407. |
| Lazar et al. ‘A video time encoding machine’, in Proceedings of the 15th IEEE International Conference on Image Processing (ICIP '08 2008, pp. 717-720. |
| Lazar et al. ‘Consistent recovery of sensory stimuli encoded with MIMO neural circuits.’ Computational intelligence and neuroscience (2010): 2. |
| Masquelier, Timothee, ‘Relative spike time coding and STOP-based orientation selectivity in the early visual system in natural continuous and saccadic vision: a computational model.’ Journal of computational neuroscience 32,3 (2012): 425-441. |
| Nichols, A Reconfigurable Computing Architecture for Implementing Artificial Neural Networks on FPGA, Master's Thesis, The University of Guelph, 2003, pp. 1-235. |
| Paugam-Moisy et al., “Computing with spiking neuron networks” G. Rozenberg T. Back, J. Kok (Eds.), Handbook of Natural Computing, Springer-Verlag (2010) [retrieved Dec. 30, 2013], [retrieved online from link.springer.com ]. |
| Pavlidis et al. Spiking neural network training using evolutionary algorithms. In: Proceedings 2005 IEEE International Joint Conference on Neural Networkds, 2005, IJCNN'05, vol. 4, pp. 2190-2194 Publication Date Jul. 31, 2005 [online] [Retrieved on Dec. 10, 2013] Retrieved from the Internet &It;URL: http://citeseerx.ist.psu.edu! viewdoc/download?doi=0.1.1.5.4346&rep-repl&type-pdf. |
| Sato et al., ‘Pulse interval and width modulation for video transmission,’ Cable Television, IEEE Transactions on 4 (1978): 165-173. |
| Schemmel, J., Grub!, A., Meier, K., Mueller, E.: Implementing synaptic plasticity in a VLSI spiking neural network model. In: Proceedings of the 20061ntemational Joint Conference on Neural Networks (IJCNN'06), IEEE Press (2006) Jul. 16-21, 2006, pp. 1-6 [online], [retrieved on Aug. 24, 2012]. Retrieved from the Internet &It;URL: http://www.kip.uniheidelberg.deNeroeffentlichungen/download.cgi/4620/ps/1774.pdf> Introduction. |
| Simulink.RTM. model [online], [Retrieved on Dec, 10, 2013] Retrieved from <:URL: http://www.mathworks.com/ products/simulink/index.html>. |
| Sinyavskiy et al. ‘Reinforcement learning of a spiking neural network in the task of control of an agent in a virtual discrete environment’ Rus. J. Nonlin. Dyn., 2011, vol. 7, No. 4 (Mobile Robots), pp. 859-875, chapters 1-8 (Russian Article with English Abstract). |
| Sjostrom et al., ‘Spike-Timing Dependent Plasticity’ Scholarpedia, 5(2)1 362 (2010), pp. 1-18. |
| Szatmary et al,, “Spike-timing Theory of Working Memory” PLoS Computational Biology, vol. 6, Issue 8, Aug. 19, 2010 [retrieved on Dec. 30, 2013]. Retrieved from the Internet: URL: http://www.ploscompbioLorg/article/info%3Adoi% 2F10.1371%2Fjournal,pcbi.1000879<url:><url:>. |
| Wang ‘The time dimension for scene analysis.’ Neural Networks, IEEE Transactions on 16.6 (2005): 1401-1426. |
| Shepard et al., “Evaluation of Smart Grid and Civilian UAV Vulnerability to GPS Spoofing Attacks”, Northrop Grumman Information Systems, 2012 ION GNSS Conference Nashville, TN, Sep. 19-21, 2012. |
| Number | Date | Country | |
|---|---|---|---|
| 20160075015 A1 | Mar 2016 | US |
