Research On Wearable MIMU/sEMG Personal Indoor Relative Positioning Technology

With the development of modern technology,personal relative indoor positioning technology has become the foundation of smart life and plays a vital role in people's future life.At present,the outdoor positioning technology based on satellite is becoming more and more mature,but the satellite signal is blocked by the building and cannot meet the needs of personal relative positioning in indoor scenes.In this paper,relative indoor positioning technology has been researched.In which,Support Vector Machine(SVM)algorithm optimized by Adaptive Particle Swarm Optimization/Cuckoo Search(APSO/CS)for leg movements recognition is proposed,the stride estimation model is constructed and the improved Mahony complementary filtering algorithm is studied.The main content of this paper is as follows:(1)Leg movements recognition algorithm in indoor environment:In order to solve the problem of low accuracy of leg movements recognition by different sensors,SVM algorithm optimized by APSO/CS for leg movements recognition is proposed.The APSO/CS algorithm is used to optimize the kernel function parameter g and the penalty factor c in the SVM model,and the voting mechanism is used to correct the wrong classification labels,and improve the accuracy of recognition.Experimental results show that the average correct rate of recognition is about 98.78%.(2)Stride estimation model:Aiming at the problem of inaccurate stride estimation in the commonly used models at different speed,an improved stride estimation model fusing acceleration signal,step frequency and sEMG signal is proposed.Based on the correct recognition of leg movements,the model can adapt to stride estimation under conditions of different speed.Experimental results show that the average error and error rate of the model are 0.01 m and 1.01%.(3)Improved indoor forward direction estimation algorithm:Aiming at the problem that the Mahony complementary filtering algorithm needs to manually set the controller parameters K_p and K_i when estimating the forward direction of pedestrians,a Mahony complementary filtering algorithm improved by Back Propagation(BP)neural network is proposed.Aiming at the problem of cumulative error in the calculation of the forward direction by Mahony complementary filtering algorithm,a Mahony complementary filtering algorithm optimized by the dominant direction is proposed.Experimental results show that in the process of walking on a rectangular route,both the Mahony complementary filtering algorithm improved by the BP neural network and the Mahony complementary filtering algorithm optimized by the dominant direction can significantly reduce the error of the Mahony complementary filtering algorithm in calculating the forward direction.(4)Wearable personal relative indoor positioning test based on Micro-Inertial Measurement Unit/surface electromyography(MIMU/sEMG):use the recognized leg movements,estimated stride,corrected forward direction and other information to calculate the pedestrian track.Experimental test results show that the above method can better realize the relative indoor positioning of individuals.The wearable MIMU/sEMG personal relative indoor positioning method proposed in this paper can also be applied to personal relative indoor positioning fields such as medical health and smart travel,which can realize the function of personal relative indoor positioning.