Research On Intelligent Optimization Of Upper And Lower Limb Rehabilitation Trainer Based On EMG Signal

Functional impairment of the upper and lower limbs caused by external injuries or neurological diseases will have a serious impact on People’s Daily life.However,due to the remolding of the central nervous system,the upper and lower limb function can be restored by the auxiliary therapy of the upper and lower limb rehabilitation equipment.At present,the single rehabilitation plan of the upper and lower limb rehabilitation equipment and the periodic excessive rehabilitation training are easy to cause muscle fatigue in the rehabilitation treatment of patients,thus causing secondary injuries to the injured limbs of patients.Therefore,in the process of rehabilitation training for patients with nerve injury such as stroke,the patient’s muscle state can be obtained non-invasively with the help of the surface em G data of the patient’s muscles,which can effectively prevent excessive training in the rehabilitation process of patients.Because human EMG data easily contain noises and interference from many aspects,it can not accurately reflect the state of human muscles,and thus can not accurately detect muscle fatigue.In this paper,EMG denoising based on convolutional neural network and muscle fatigue detection method and system based on improved Elman neural network are designed,which can effectively eliminate the noise in EMG data and improve the accuracy of muscle fatigue detection.The research contents and achievements of this paper are as follows:(1)state of human upper limb muscle fatigue detection system is introduced in this paper the whole process of thought,to carry out the human upper limb muscle electrical signal acquisition experiments,will be collected the original upper limbs for multi-channel semg feature extraction,and entered into the convolutional neural network to deal with the noise,the system interface are optimized by the onset of rehabilitation information display real-time muscles.(2)The improved Elman neural network in this paper is used for muscle fatigue recognition of electromyography data after denoising,which has high recognition accuracy and real-time performance for muscle fatigue state recognition of patients trained by upper and lower limb rehabilitation trainers.(3)When patients use the upper and lower limb rehabilitation trainers for rehabilitation training,they can not only collect the em G data of patients in the training process,so as to obtain the user’s muscle movement and detect the user’s muscle fatigue.In addition,adaptive electric stimulation intensity adjustment and adaptive training speed adjustment strategies were first added to assist rehabilitation therapy,which effectively prevented lazy training and overtraining of patients.This paper makes full use of the patient’s muscle power data in time domain and frequency domain characteristics,using neural network for patients in rehabilitation training in the process of detecting state of muscle fatigue,and according to the state of fatigue design the electrical stimulation intensity and training speed adaptive adjustment strategy,protect patients in the case of excessive fatigue,excessive training.At the same time,ensure that patients will not be lazy training,reduce the rehabilitation effect,slow down the rehabilitation process.Physiotherapists can intervene,control and adjust rehabilitation treatment programs in real time according to muscle information to ensure the effectiveness and safety of rehabilitation training.