Robot control system, robot control method, and program

A robot control system comprising:

• a control device configured to input a current output of an environment sensor, which acquires an environmental condition of a robot device, to a trained model, the trained model being obtained by machine learning, to generate an operation command of the robot device based on the current output of the environment sensor output from the trained model, and control an operation of the robot device,
• wherein the control device comprises:
• a memory storing a program;
• and
• a processor which when executing the program configures the processor to:
• calculate a matching degree between an environmental condition of the robot device determined when the trained model has machine-learned the operation command of the robot device and a current environmental condition of the robot device based on the current output of the environment sensor;
• and
• adjust at least one of a content of the operation command of the robot device and a setting of the environment sensor based on a calculation result of the calculation performed by the processor,
• wherein the trained model is a trained model obtained by performing machine learning to generate a reproduction signal of the output of the environment sensor in addition to the operation command of the robot device based on the output of the environment sensor, and
• wherein the processor is further configured to:
• generate the operation command of the robot device corresponding to the current output of the environment sensor and the reproduction signal of the output of the environment sensor at a time of the learning when the current output of the environment sensor is input;
• and
• calculate a difference between the reproduction signal of the output of the environment sensor at the time of the learning output by the trained model and the current output of the environment sensor as the matching degree between the environmental conditions.

A robot control method of a robot control system including a robot device and an environmental sensor, the method comprising the steps of:

• acquiring, by the environment sensor, an environmental condition of the robot device;
• inputting, by a control device, a current output of the environment sensor to a trained model, the trained model being obtained by machine learning, to generate an operation command of the robot device corresponding to the current output of the environment sensor output from the trained model, and controlling, by the control device, an operation of the robot device;
• calculating, by the control device, a matching degree between an environmental condition of the robot device determined based on the trained model having machine-learned the operation command of the robot device and a current environmental condition of the robot device based on the current output of the environment sensor;
• adjusting, by the control device, at least one of a content of the operation command of the robot device and a setting of the environment sensor based on a calculation result of the matching degree between the environmental conditions;
• and
• causing the robot device to operate, based on at least one of the operation command of the robot device and the setting of the environment sensor adjusted in the adjustment step,
• wherein the trained model is a trained model obtained by performing machine learning to generate a reproduction signal of the output of the environment sensor in addition to the operation command of the robot device based on the output of the environment sensor, and
• wherein the method further comprises the steps of:
• generating the operation command of the robot device corresponding to the current output of the environment sensor and the reproduction signal of the output of the environment sensor at a time of the learning based on input of the current output of the environment sensor;
• and
• calculating a difference between the reproduction signal of the output of the environment sensor at the time of the learning output by the trained model and the current output of the environment sensor as the matching degree between the environmental conditions.

A robot control method of a robot control system including a robot device and an environmental sensor, the method comprising the steps of:

• acquiring, by the environment sensor, an environmental condition of a robot device;
• inputting, by the control device, a current output of the environmental sensor to a trained model, the trained model being obtained by machine learning, to generate an operation command of the robot device corresponding to the current output of the environment sensor output from the trained model, and controlling an operation of the robot device;
• calculating, by the control device, a matching degree between an environmental condition of the robot device determined based on the trained model having machine- learned the operation command of the robot device and a current environmental condition of the robot device based on the current output of the environment sensor;
• and
• adjusting, by the control device, at least one of a content of the operation command of the robot device and a setting of the environment sensor based on a calculation result of the matching degree between the environmental conditions,
• wherein the trained model is a trained model obtained by performing machine learning to generate a reproduction signal of the output of the environment sensor in addition to the operation command of the robot device based on the output of the environment sensor, and
• wherein the method further comprises the steps of:
• generating the operation command of the robot device corresponding to the current output of the environment sensor and the reproduction signal of the output of the environment sensor at a time of the learning based on input of the current output of the environment sensor;
• and
• calculating a difference between the reproduction signal of the output of the environment sensor at the time of the learning output by the trained model and the current output of the environment sensor as the matching degree between the environmental conditions.