# US 12,222,727 B2 · Robot control device, robot control method, and learning model generation device

> Agents: this data is available as JSON from the patentgrep API and MCP server — see https://patentgrep.com/llms.txt

- Number: US 12,222,727 B2
- Document ID: US12222727B2
- Status: Granted
- Kind: B2
- Application number: 17/836,542
- Filed: 2022-06-09
- Published: 2022-09-22
- Granted: 2025-02-11
- Earliest claimed: 2020-02-27
- Assignees: Mitsubishi Electric Corporation
- Applicants: Mitsubishi Electric Corporation
- Inventors: Kei Ota
- Web page: https://patentgrep.com/patent/US12222727B2

## CPC

- G05D 1/0221: {involving a learning process}
- G05D 1/0214: {in accordance with safety or protection criteria, e.g. avoiding hazardous areas (monitoring the location of vehicles within a certain area, e.g. forbidden or allowed areas, in traffic control systems for road vehicles G08G1/13)}
- G05D 1/617: Safety or protection, e.g. defining protection zones around obstacles or avoiding hazards (arrangements for controlling the position or course of two or more vehicles for avoiding collisions therebetween G05D1/693; arrangements for reacting to or preventing system or operator failure G05D1/80)
- G06N 3/08: Learning methods
- G01C 21/00: Navigation; Navigational instruments not provided for in groups G01C1/00 - G01C19/00 (measuring distance traversed on the ground by a vehicle G01C22/00; control of position, course, altitude or attitude of vehicles G05D1/00; traffic control systems for road vehicles involving transmission of navigation instructions to the vehicle G08G1/0968)
- G05D: SYSTEMS FOR CONTROLLING OR REGULATING NON-ELECTRIC VARIABLES
- G06N: COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G01C: MEASURING DISTANCES, LEVELS OR BEARINGS; SURVEYING; NAVIGATION; GYROSCOPIC INSTRUMENTS; PHOTOGRAMMETRY OR VIDEOGRAMMETRY (measuring liquid level G01F; radio navigation, determining distance or velocity by use of propagation effects, e.g. Doppler effects, propagation time, of radio waves, analogous arrangements using other waves G01S)

## Abstract

A robot control device is configured to include: a moving route acquiring unit to acquire a moving route of a robot from a first learning model by giving, to the first learning model, observation data indicating a position of an obstacle being present in a region where the robot moves and state data indicating a moving state of the robot at a movement start point at which the robot starts moving among moving states of the robot in the region where the robot moves; and a control value generating unit to generate a control value for the robot, the control value for allowing the robot to move along the moving route acquired by the moving route acquiring unit.

## Claims

### Claim 1

A robot control device comprising:
a computer processor;
a memory storing instructions executed by the computer processor to
acquire a moving route of a robot from a first learning model, by giving, to the first learning model, observation data indicating a position of an obstacle being present in a region where the robot moves and state data indicating a moving state of the robot at a movement start point where the robot starts moving among moving states of the robot in the region where the robot moves, the first learning model outputting route data including coordinate data of a next position on the moving route to which robot is to move from the movement start point, and
generate at least one control value for the robot, the at least one control value indicating an acceleration by which the robot is to move at a particular time to proceed along the acquired moving route,
wherein the computer processor gives, to a second learning model, state data indicating a moving state of the robot when the robot is moving in the region and the acquired moving route including the route data outputted by the first learning model, and acquires the at least one control value for the robot from the second learning model; and
a controller to control the robot based on the at least one control value.

### Claim 2 (depends on 1)

The robot control device according to claim 1, the executed instructions further causing the computer processor to change a control value related to a speed of the robot to equal to or less than a speed threshold when the control value related to the speed of the robot is larger than the speed threshold, and change a control value related to an acceleration of the robot to equal to or less than an acceleration threshold when the control value related to the acceleration of the robot is larger than the acceleration threshold, among control values for the robot acquired from the second learning model.

### Claim 3

A robot control method comprising:
acquiring, a moving route of a robot from a first learning model, by giving, to the first learning model, observation data indicating a position of an obstacle being present in a region where the robot moves and state data indicating a moving state of the robot at a movement start point where the robot starts moving among moving states of the robot in the region where the robot moves, the first learning model outputting route data including coordinate data of a next position on the moving route to which robot is to move from the movement start point;
generating, at least one control value for the robot, the at least one control value indicating an acceleration by which the robot is to move along the acquired moving route;
giving, to a second learning model, state data indicating a moving state of the robot when the robot is moving in the region and the acquired moving route including the route data outputted by the first learning model, and acquiring the at least control value for the robot from the second learning model; and
controlling the robot based on the at least one control value.

### Claim 4

A learning model generation device comprising:
a computer processor; and
a memory storing instructions executed by the computer processor to
generate teacher data indicating a route along which a robot can move in each of a plurality of regions without colliding with an obstacle being present in each of the regions;
generate a first learning model that learns a moving route of the robot by using observation data indicating a position of the obstacle being present in each of the plurality of regions, state data indicating a moving state of the robot at a movement start point at which the robot starts moving among moving states of the robot in each of the regions, and the generated teacher data, and outputs moving route data for the robot, when observation data indicating a position of an obstacle being present in a region where the robot moves and state data indicating a moving state of the robot at a movement start point at which the robot starts moving among moving states of the robot in the region where the robot moves are given, the moving route data including coordinate data of a next position on the moving route to which the robot is to move from the movement start point, and
generate a second learning model that
learns a control value for the robot by using state data indicating a moving state of the robot when the robot is moving in each of the regions and the generated teacher data, or the moving route data for the robot output from the first learning model, the control value indicating an acceleration by which the robot is to move; and
outputs the control value to a controller, when state data indicating a moving state of the robot when the robot is moving in a certain region and a moving route of the robot output from the first learning model are given, the controller using the control value to control the at least one of the speed and acceleration by which the robot moves at a particular time to proceed along the moving route.

### Claim 5 (depends on 4)

The learning model generation device according to claim 4, wherein
the executed instructions further cause the computer processor to generate a second learning model that
learns a control value for the robot by using the state data indicating the moving state of the robot when the robot is moving in each of the regions and the generated teacher data, and
outputs a control value for the robot, when the state data indicating the moving state of the robot when the robot is moving in the certain region and a moving route of the robot, which is output from a different source than the first learning model, are given.