Novel brain-controlled intelligent rehabilitation method based on capsule network, and application thereof

Abstract

The invention discloses a novel brain-controlled intelligent rehabilitation method based on a capsule network, and application thereof. According to the method, an upper computer generates corresponding motion stimulation videos according to rehabilitation training items selected by a user, the motion stimulation videos are displayed to the user through a displayer, the user imagines correspondingactions, and synchronous motion imagination electroencephalogram signals of the user are collected through an electroencephalogram collecting device, subjected to amplifying and filtering preprocessing and transmitted to the upper computer in a wireless mode; after the upper computer receives the motion imagery electroencephalogram signals and processes the motion imagery electroencephalogram signals again, the features of the motion imagery electroencephalogram signals are extracted based on a capsule network algorithm, and are classified, the classification results are converted into corresponding rehabilitation training control signals, and the rehabilitation training control signals are transmitted to a rehabilitation exercise device; the rehabilitation exercise device drives the userto perform corresponding training actions and feeds back signal data of the training actions of the user to the upper computer; and the upper computer evaluates the signal data of the training actions of the user by adopting a Fugl-Meyer evaluation method. The safety of the brain-controlled rehabilitation system can be improved.

Description

technical field [0001] The invention relates to a brain-controlled intelligent rehabilitation method. In particular, it relates to a novel brain-controlled intelligent rehabilitation method and its application based on a capsule network. Background technique [0002] Because motor imagery can realize that the limbs do not perform any action when the brain has an action intention, and it does not depend on the patient's residual function and is closely related to the patient's active movement, so motor imagery is usually used in rehabilitation medicine to activate and restore the movement of stroke patients Network to improve motor function in stroke patients. The rehabilitation system based on motor imagery can effectively improve the rehabilitation effect and solve the problems existing in traditional rehabilitation methods. The most critical and important step in this system is to correctly identify and classify motor imagery EEG signals, so as to transmit them to exercis...

AI Technical Summary

Problems solved by technology

Since the sports rehabilitation system not only has high requirements on the accuracy of EEG signal classification, but also requires the system to have real-time online processing capabilities, an overly complex deep neural network will reduce the real-time performance of the system and delay the output of control commands.

Images