Human Upper Body Movement Recognition And Muscle Fatigue Detection Based On SEMG Signal

"Cerebral stroke" is an acute cerebrovascular disease caused by a sudden burst of blood vessels in the brain or a blood vessel that cannot cause blood to flow into the brain and cause damage to the brain tissue.Stroke has a high incidence of morbidity and mortality.It has become one of the main causes of human death.Some of the patients who survived after the illness failed to recover their autokinetic function,the normal work and life of the patient are severely affected.Therefore,it is of great importance to rehabilitate the patients in order to reduce the bad impact of disease,improve their daily life ability,and relieve the family and social burden.Various disciplines have also conducted studies on rehabilitation training systems for stroke patients.The surface EMG signal,which is easy to collect and can directly reflect the muscle condition,has a great significance for the study of rehabilitation training.The surface EMG signal is a weak electrical signal that is synthesized in time and space by action potentials of the exercise unit that are distributed by each movement unit participating in muscle activity detected by the surface electrode on the surface of the human skin during the activity of the muscle.Mainly research areas of this paper is use surface EMG signals to carry out pattern recognition and muscle fatigue detection for normal human upper limb movements.The main work is as follows:1.sEMG signal acquisition experiment scheme.Three muscles of the upper limbs were collect: biceps brachii,deltoid muscle and triceps brachii muscle.Eleven upper limb motion were design for surface EMG signal acquisit experiments to build a sEMG signal database for later study.2.Application of improved TKE Algorithm in starting and endpoint detection of sEMG signals.Adding information gain as the best criterion for multi-threshold evaluation,the accuracy and flexibility of the the new algorithm were improved.3.A muscle fatigue detection method based on sEMG signal was proposed.By extracting MPF,MF,RMS and IEMG values from the sEMG signal,a method for estimating the moment of muscle fatigue was proposed.The experimental results showed that this method has certain feasibility.4.A method of recognizing various motion patterns based on sEMG signals was studied.Thirty-one features were extract from sEMG signals in time domain,frequency domain and time-frequency domain.Single-feature and multi-feature combinational classification tests that the probabilistic neural network?PNN?used as classifier were perform.The feature combination WL + AR.+ db4_5RMS + MAV has high classification accuracy and meets real-time requirements for time consumption,and it has a certain level of noise immunity.5.Completed real-time identification of human upper limb motion software based on sEMG signal.Combining the above research,a real-time upper limb motion recognition by sEMG signals based on MATLAB GUI software was developed,which enabled the addition,deletion,and modification of the input personnel information and achieved real-time recognition of upper limb movements.