Research On Control Method Of Rehabilitation Robot Based On Motion Imagination

In the 21st century,the mortality rate of cardiovascular and cerebrovascular diseases ranks first among all causes of death.About 15 million people worldwide suffer from cardiovascular and cerebrovascular diseases such as stroke,resulting in permanent paralysis.Cardiovascular and cerebrovascular diseases not only bring great inconvenience to patients’ food,clothing and housing,but also bring huge economic pressure to their families.However,a single passive exercise can only ensure that the patient’s muscles do not shrink and cannot restore the relevant functions of the nervous system from the root cause.As a result,more and more patients are expecting rehabilitation treatment to regain normal motor function in the limbs.With the growth of the patient base,the rehabilitation medical resources are in short supply,and more and more rehabilitation robots have been developed to alleviate the tension.However,most products on the market can only perform passive,mechanical predetermined actions,which can not help patients reshape the neural pathways,and can not achieve the healing effect.Only by effectively combining the patient’s active exercise intention with the machine feedback can the patient be motivated to perform rehabilitation exercises.In the method of collecting the patient’s motion intention,the electroencephalogram(EEG)based on the motion imagination is not only easy to operate,but also does not cause damage to the human body,and is a good choice for robot control.In this thesis,the control method of rehabilitation robot based on motion imaging is studied.By collecting EEG signals based on motion imaging,EEG-Images are generated through denoising and signal reconstruction,and EEG signals are classified by deep learning to control ankle.Joint medical rehabilitation exercise equipment to establish an effective and safe system.This thesis first introduces the practical application of brain-computer interface technology in the field of rehabilitation robots,and discusses the research status of brain-computer interface and EEG signal processing at home and abroad,which leads to the main research content of this thesis.Secondly,this thesis carries out the work related to the acquisition and preprocessing of EEG signals.The high frequency noise contained in the original EEG signal is filtered by the method of wavelet decomposition,and the power spectrum information of EEG is extracted.Finally,Then,using the wavelet reconstruction algorithm,the optimal reconstruction scheme is explored,and the effect of filtering out high-frequency interference signals without losing useful information is achieved.Thirdly,this thesis studies the EEG signal classification method based on deep learning,combines the power spectrum information in the EEG signal with the electrode position to generate EEG-Images of 32×32 size,and uses four network models for data testing,which is higher.Classification accuracy.Finally,this thesis carried out the construction of the back-end system of the ankle rehabilitation robot,which not only built a simple and easy-to-use graphical interface for the upper computer,but also transmitted the control signal to the lower computer safely,and provided an interface for the feedback of the lower computer information.The patient’s exercise status is displayed in real time in the graphical interface to form a safe,stable and easy-to-use rehabilitation exercise robot platform.