# US 6,724,364 B2 · Learning-type movement control apparatus, method therefor, and distribution medium therefor > Agents: this data is available as JSON from the patentgrep API and MCP server — see https://patentgrep.com/llms.txt - Number: US 6,724,364 B2 - Document ID: US6724364B2 - Status: Granted - Kind: B2 - Application number: 09/733,682 - Filed: 2000-12-08 - Published: 2001-10-11 - Granted: 2004-04-20 - Earliest claimed: 1999-12-10 - Assignees: Sony Corporation - Inventors: Jun Tani - Web page: https://patentgrep.com/patent/US6724364B2 ## CPC - G06F 3/038: Control and interface arrangements therefor, e.g. drivers or device-embedded control circuitry - G05B 2219/35438: Joystick - G05B 2219/40116: Learn by operator observation, symbiosis, show, watch - G06F: ELECTRIC DIGITAL DATA PROCESSING (computer systems based on specific computational models G06N) - G05B: CONTROL OR REGULATING SYSTEMS IN GENERAL; FUNCTIONAL ELEMENTS OF SUCH SYSTEMS; MONITORING OR TESTING ARRANGEMENTS FOR SUCH SYSTEMS OR ELEMENTS (systems for controlling or regulating non-electric variables G05D; systems for regulating electric or magnetic variables G05F; control devices or systems insofar as characterised by mechanical features only G05G) ## Abstract A learning-type movement control apparatus that learns the movement of an operation control device, predicts the movement thereof, and drives it so as to automatically move. The apparatus comprises an operation control device having a predetermined portion that is displaced according to a force exerted in an arbitrary direction, outputs the amount of the displacement at least as one-dimensional position-representing information, receives a feedback signal carrying information generated by adding displacement information to the position-representing information, and drives the predetermined portion according to a direction and a displacement that are based on the feedback signal. The apparatus also includes a learning section that receives the position-representing information and performs learning of the movement of the operation control device. The apparatus further includes a predicting section that performs prediction of a displacement of the operation control device according to the position-representing information the learning result of the learning section, and generates the feedback signal by adding the displacement information to the position-representing information. ## Claims ### Claim 1 A learning-type movement control apparatus comprising: an operation control device that has a predetermined portion which is displaced according to a force exerted in an arbitrary direction, that outputs the amount of the displacement at least as one-dimensional position-representing information, that receives a feedback signal carrying information generated by adding displacement information to the position-representing information, and that drives said predetermined portion according to a direction and a displacement that are based on the feedback signal; a learning section that receives the position-representing information which is input in time series from said operation control device and that performs learning of the movement of said operation control device; and a predicting section that performs prediction of the displacement of said operation control device according to the position-representing information that is input in time series from said operation control device and a result of the learning by said learning section, that generates the feedback signal by adding the displacement information to the position-representing information, and that outputs the feedback signal to said operation control device. ### Claim 2 (depends on 1) A learning-type movement control apparatus as claimed in claim 1, further comprising a recurrent neural network that comprises: said learning section; said predicting section; an input layer; a hidden layer; and an output layer. ### Claim 3 (depends on 2) A learning-type movement control apparatus as claimed in claim 2, wherein said recurrent neural network performs feedback from said output layer to said input layer. ### Claim 4 (depends on 1) A learning-type movement control apparatus as claimed in claim 1, wherein said learning section performs the learning of the movement of said predetermined portion of said operation control device according to an error propagation method. ### Claim 5 (depends on 2) A learning-type movement control apparatus as claimed in claim 2, wherein said learning section performs the learning of the movement of said predetermined portion of said operation control device according to an error propagation method. ### Claim 6 (depends on 3) A learning-type movement control apparatus as claimed in claim 3, wherein said learning section performs the learning of the movement of said predetermined portion of said operation control device according to an error propagation method. ### Claim 7 A learning-type movement control method comprising the steps of: performing learning of a time-series pattern of movement information output by an operation control device by using a recurrent neural network, where said movement information output by said operation control device is based on movement of a predetermined portion of said operation control device according to a pattern of force applied to said operation control device; performing prediction of information regarding movement of said predetermined portion of said operation control device by using said recurrent neural network according to a result of the learning; providing a feedback signal to said operation control device, where said feedback signal is the sum of the movement information output by the operation control device and the predicted movement information; and driving said operation control device to move according to said feedback signal. ### Claim 8 (depends on 7) A learning-type movement control method as claimed in claim 7, wherein said recurrent neural network performs feedback from an output layer to an input layer. ### Claim 9 (depends on 7) A learning-type movement control method as claimed in claim 7, wherein said learning section performs the learning of the movement of said predetermined portion of said operation control device according to an error propagation method. ### Claim 10 (depends on 8) A learning-type movement control method as claimed in claim 8, wherein said learning section performs the learning of the movement of said predetermined portion of said operation control device according to an error propagation method. ### Claim 11 A distribution medium for distributing a computer-readable program that allows an information-processing unit to execute processing comprising the steps of: performing learning of a time-series pattern of movement information output by an operation control device by using a recurrent neural network, where said movement information output by said operation control device is based on movement of a predetermined portion of said operation control device according to a pattern of force applied to said operation control device; performing prediction of information regarding movement of said predetermined portion of said operation control device by using said recurrent neural network according to a result of the learning; providing a feedback signal to said operation control device, where said feedback signal is the sum of the movement information output by the operation control device and the predicted movement information; and driving said operation control device to move according to said feedback signal. ### Claim 12 (depends on 1) A learning-type movement control apparatus as claimed in claim 1, where the operation control device is for controlling the movement of a position indicating item on a display in a computer system. ### Claim 13 (depends on 3) A learning-type movement control apparatus as claimed in claim 3, wherein said input layer includes one or more input context neurons and said output layer includes one or more output context neurons, and wherein each output context neuron provides data to a corresponding input context neuron. ### Claim 14 (depends on 7) A learning-type movement control method as claimed in claim 7, where the movement of said operation control device controls the movement of a position indicating item on a display in a computer system. ### Claim 15 (depends on 8) A learning-type movement control method as claimed in claim 8, wherein said input layer includes one or more input context neurons and said output layer includes one or more output context neurons, and wherein each output context neuron provides data to a corresponding input context neuron. ### Claim 16 (depends on 11) A distribution medium as claimed in claim 11, where the movement of said operation control device controls the movement of a position indicating item on a display in a computer system. ### Claim 17 (depends on 11) A distribution medium as claimed in claim 11, wherein: said recurrent neural network performs feedback from an output layer to an input layer, said input layer includes one or more input context neurons, said output layer includes one or more output context neurons, and each output context neuron provides data to a corresponding input context neuron.