| Bicycle riding,as a basic lower limb motion,can help amputees avoid atrophy during rehabilitation.Besides,bicycle riding also has prominent effects in balance training,short trip and physical exercise for lower limb prosthesis wearers.Therefore,the study of bicycle riding of intelligent prosthesis wearers is important for lower limb amputees.This thesis mainly forces on three aspects,including bicycle riding data collection,bicycle riding status and phases recognition,and prosthetic knee joint motion control.First of all,the movement characteristics and four bicycle riding phases are analyzed.A motion data sampling system is established.And,the singular value noise reduction method is used for signal de-noising,which provides essential data for bicycle riding recognition and control.Then,decision tree method is used for bicycle riding status recognition of prosthetic wearers.The recognition accuracy rate reaches 98.3%.Next,a Support Vector Machine(SVM)-based bicycle riding phase recognition model is designed.The motion data are combined as a multi-dimensional feature vector and input to the bicycle riding phase recognition model.In order to improve the accuracy of bicycle riding phase recognition,Grey Wolf Optimizer is integrated into the SVM-based recognition model.As a result,the phase recognition accuracy increases to 94%,which verifies the effectiveness and feasibility of the optimization algorithm classification algorithm proposed in this thesis.In addition,compared with BP neural network,GA-SVM and PSO-SVM recognition model,it can be seen that the GWO-SVM classification model has better performance in both recognition accuracy and convergence speed.It means that the bicycle riding status and phase recognition of prosthesis wearers can be effectively solved.In the final part,a prosthesis motion simulation model of bicycle riding control is established.The control method combining finite state machine and iterative learning control is applied in the bicycle riding control model.In the simulation,the key points of bicycle riding phase conversion are taken as the transition conditions.The motion control of the prosthesis joint at different phases is achieved,which lays the foundation for the intelligent prosthetic limb cooperating with the prosthesis wearer to realize the bicycle riding movement. |
