A brain-computer collaborative digital twin reinforcement learning control method and system

Abstract

The invention discloses a brain-machine cooperative digital twin reinforcement learning control method and system. By constructing a brain-computer cooperative control model, the operator gives a virtual robot direction instruction, and at the same time collects the brain information when the operator gives the virtual robot direction instruction. According to the collected EEG signals, the corresponding speed command of the virtual robot is given to complete the specified action, and the brain-computer collaborative control model is rewarded according to the quality of completion, and the training of the brain-computer collaborative control model is completed. The double-loop information interaction mechanism between the brain and the machine is realized by means of intensive learning, and the interaction between the information layer and the instruction layer between the brain and the machine is realized. The brain state of the robot compensates and regulates the instructions of the robot to achieve precise control. Compared with other brain-computer cooperation methods, it improves the robustness and generalization ability, and realizes the mutual adaptation and mutual growth between the brain-computer.

Description

technical field [0001] The invention belongs to the field of brain-computer interface and artificial intelligence comprehensive technology, and relates to a brain-computer cooperative digital twin reinforcement learning control method and system. Background technique [0002] With the development of robotics technology, there is an increasingly urgent demand for intelligent robots with human-like advanced perception and cognition capabilities and the ability to perform non-set tasks in highly complex environments. However, to achieve intelligent robots with human thinking and reasoning methods, autonomous discovery and feature extraction, online incremental learning, and comprehensive processing of various information capabilities only with artificial intelligence technology, the current technology is still not up to it. Utilizing the fusion of human-machine intelligence and giving full play to the advantages of different intelligences of humans and computers is an important...

AI Technical Summary

Problems solved by technology

[0003] After research, it is found that there are still two problems in the field of precision control: (1) the lack of two-way interaction of information between the operator and the robot, and the inability to realize the precise perception of the operator's intention; (2) due to the distraction of the human brain , mental fatigue and excessive mental load, etc., resulting in poor performance of the brain-computer hybrid intelligent system, and even danger

In view of the deterioration of the performance of the brain-computer system caused by the mental state of the human brain, currently only the one-way compensation of the control command is considered, and there is a lack of a dual-loop interaction mechanism between the brain and the computer.

To sum up, for the "human-in-the-loop" brain-computer collaborative control method, there is still no integrated brain-computer collaboration model established at this stage, and it is impossible to effectively realize the deep integration of the information layer and the command layer, and the brain-computer collaborative control. Accuracy, stability and safety need to be improved

Images