Research On Gait Recognition And Gait Planning Of Bionic Intelligent Lower Limb Prosthesis

The research of the bionic intelligent lower limb prosthesis aims at using the rapidly developing artificial intelligence technology and the matured humanoid robot technology to enable the prosthesis to recognize the different gait of the amputee and to make the planned prosthetic gait have better biomimetic performance.This will enable people and prostheses to coordinate their movements and help disabled people regain vitality.Therefore,the research in this paper has social and economic benefits.This article starts from gait recognition and extends before and after.It mainly studies the following contents:(1)Based on the analysis of the lower limbs of the human body,the inertial motion capture system was selected as the motion data acquisition device,and the motion data was collected for the five gaits of the human body in different collection environments.Based on the experiment,500 exercise data of 10 people under 5 gaits were obtained.For the discrete time domain signal obtained,wavelet transform denoising is used.Knee joint angle signals were analyzed,and the curve fitting of the single cycle signal was performed using the MATLAB curve fitting toolbox.The analysis of the signal changes was consistent with the human body movement law and verified the correctness of the collected data.(2)In order to establish a gait recognition model,the eigenvalues of the signal are extracted.The absolute mean,variance and peak are selected as feature vectors.Because the initial parameters of the traditional BP neural network model are setted randomly,the artificial bee colony algorithm is used to optimize the weights and thresholds,and a gait recognition model for bee colony optimized BP neural network is established.Using 300 eigenvectors as training samples and 200 eigenvectors as test samples,the network was trained and tested.The final gait recognition accuracy rate reached 90%,verifying the correctness and superiority of the model.(3)A linear inverted pendulum model was established to obtain the human body level walking center-of-mass trajectory,and the five-bar model was used to find the angles of the supporting phase joints.Taking into account the plantar reaction force,the linear inverted pendulum model was modified to obtain a joint angle more in line with the human body movement law.The angle of the swing related section is obtained,the gait planning result of the complete gait cycle is obtained,and the optimized centroid trajectory is obtained from the energy curve of the inverted pendulum and the geodesic differential equation.(4)The BRHL virtual simulation platform was used to simulate the bionic intelligent lower limb prosthesis level walking.The gait planning results were verified based on ADAMS,Pro/Engineer and Simulink.The results show that the trajectory of the center of mass and the centroid of the foot is in agreement with the ideal value,and the gait has a good anthropomorphism,which verifies the correctness and validity of the gait planning method.