Research On The Energy Consumption And Fatigue Model Of Erector Spine Muscle Based On Acceleration Signal

In recent years,the rapid development of logistics industry has produced a large number of labor demand.Logistics industry has become a high incidence industry of back pain,because it needs to be carried back and forth,with heavy workload and high work intensity.The fatigue of the erector spine muscle caused by repeated lifting is the main reason for low back pain.It is of great practical significance to evaluate the fatigue effect of the erector spine muscle during the operation of logistics workers to explore the injury mechanism and effective prevention of chronic low back pain.The target of this paper is to evaluate the motion of the erector spine muscles during lifting objects repeatedly.The following four aspects are mainly studied:(1)Kinetics and energy consumption model of erector spine muscle: This article analyzes the motion characteristics of the erector spine muscle when bending and moving objects,establishes an energy consumption model for the erector spine muscle,and verifies the model by monitoring blood oxygen values.(2)Fatigue experiment of erector spine muscle: design the fatigue experiment scheme of erector spine muscle,carry out the fatigue experiment of erector spine muscle for many testers,record the acceleration signal and s EMG signal of the subject,optimize the subjective fatigue state evaluation of the tester,and provide a data base for subsequent research.(3)Acceleration signal preprocessing and characteristic parameters extraction: filter out the high-frequency noise in the original acceleration signal,extract the timedomain and frequency-domain feature parameters of the collected acceleration signal.Principal component analysis was used to reduce the dimensionality of the extracted parameters,and select the features more suitable for determining the degree of fatigue were selected.(4)Recognition of erector spine muscle fatigue: Establish a non-linear classification model based on perceptron and support vector machines,train and identify data samples,proposing a support vector machine optimization model combined with K-means clustering to compare the recognition effect and stability of the model.Height,weight and sample membership are introduced to modify the model parameters to reduce the impact of individual differences.On this basis,a fatigue evaluation method based on the acceleration time domain and frequency domain eigenvalues is proposed.Compared with the traditional muscle fatigue evaluation method,the accuracy and stability of the SVM fatigue recognition model combined with K-means clustering are higher than those based on the EMG characteristics.Through parameter modification,the overall recognition accuracy is over 91%.