The Research Of Control System Of Lower Limb Exoskeleton Robot Based On EEG Signals

Lower limb exoskeleton robot is a king of human-machine integrative system that can help paraplegia patients and the old to do rehabilitation training.Recent years,BrainComputer Interface(BCI)developed rapidly,and it can decode the user's intents directly.Introducing BCI technology into control system of robot could help subjects with severe motor disabilities do rehabilitation training more effective.In this paper,we have developed a control system of lower limb exoskeleton robots based on Steady-State Visual Evoked Potentials and Motor Imagery,the main work is as follows:The control system of lower limb exoskeleton robots based on EEG signals is built.The exoskeleton consists of mechanical structure and auxiliary equipment,several paraplegia patients have completed rehabilitation training in virtue of it and perform several actions include sit,stand and walk.Interactive systems of SSVEP and MI and control system combined with EEG signals and sensor signals are designed.Classification of two kinds of EEG signals are realized.For the SSVEP signals,the methods of Butterworth filter,fast Fourier transform and canonical correlation analysis are used.As far as MI signals,methods of Chebyshev filter and multi class common spatial patterns are used to extract effective features,and then support vector machine output results by recognizing features.The feasibility of control system is verified.In the offline experiment,EEG signals are collected to train the parameters of the recognition algorithm.In the online experiment,the control system real-time analyze the EEG signals and perform the final action.The results show that control system can implement actions follow subjects' intention with high recognition accuracy,it also verifies that the combination of sensor signals can effectively avoid the wrong actions.