Research On EEG Pattern Recognition Technology Based On Upper Limb Rehabilitation Training Movements

There are many stroke patients in China,of which 70%-85% are associated with hemiplegia.Hemiplegic patients would suffer from disability of the upper extremity.After that,if there is no rehabilitation treatment in time,it may easily cause problems such as muscle atrophy and joint stiffness.Therefore,patients need rehabilitation training with the help of rehabilitation equipment.Brain-computer interface(BCI)refers to the establishment of a human-computer interaction system between the human brain and a computer or other electronic device that does not rely on the peripheral nervous system and muscle tissue for information transmission.The system can directly translate the human electroencephalogram(EEG)into external device control commands and control the external device to execute.Based on the principle of neurophysiological plasticity,the patient's EEG can be used to directly controls the rehabilitation training robot for rehabilitation training,which can promote the repair of damaged motor nerves.Therefore,this article aims to study EEG pattern recognition technology based on upper limb rehabilitation training actions.Applying this technology to the upper limb rehabilitation training system can enable patients to achieve active upper limb rehabilitation training and speed up the rehabilitation process.In this paper,pattern recognition technology is researched on the EEG data set of upper limb movement intention.Finally,the research results technology is applied to the upper limb rehabilitation training system and functional test of the system is performed.The main contents of this article include:(1)Traditional method of EEG energy feature extraction is improved.Studies have shown that the energy characteristics of EEG based on the movement intention of upper limbs may be quite different in different time periods.In response to this phenomenon,the energy features of EEG in different time periods of the same action is separately extracted,and the stability of the energy feature values of different time periods is tested by setting weights.Thus the validity of the energy feature value extracted by the EEG is ensured.It has been verified by experiments that the improved energy feature extraction method can effectively improve the accuracy of feature classification results.(2)The gray wolf algorithm(GWO)is applied to EEG for feature selection.Genetic algorithm(GA)is one of the classic algorithms currently used in EEG for feature selection.In this paper,the GWO algorithm and the GA algorithm are applied to the feature selection of energy eigenvalues respectively.Through comparative experiments,the superiority of the GWO algorithm applied to EEG for feature selection is fully verified.(3)An independently designed and implemented BCI system software that can be applied to upper limb rehabilitation robots is introduced.This software realized the pattern recognition technology studied above.The software combines BCI with upper limb rehabilitation training robot,and finally realizes a BCI based upper limb rehabilitation training system.This system mainly verifies the feasibility of applying the pattern recognition technology studied in this paper to the upper limb rehabilitation training system.The system test results show that the system's online action recognition rate can reach 76.50%.