基于运动意图识别的上肢助力外骨骼复合控制方法研究

Compound Control Method for Human-powered Augmentation Upper Exoskeleton Based on Motion Intent Recognition

针对上肢助力外骨骼助力时响应延时问题,提出一种基于人机交互力信号、位姿信号数据融合的运动意图识别算法,使用支持向量机(Support vector machine,SVM)分类器对人体运动信号进行分类,确定穿戴者运动意图,对控制参数做出适应性调节,改变人机协同控制系统响应速度。采用果蝇优化算法(Fruit fly optimization algorithm,FOA)优化支持向量机,提高分类准确率;引入有限状态机(Finite state machine,FSM),处理非合理运动意图。为降低外骨骼控制系统跟踪误差,设计一种模糊控制与阻抗控制结合的复合控制方法,提高控制参数更新速度,实时调整轨迹。开展运动意图识别实验,结果表明,识别准确率可达97.93%,可快速检测出非合理运动意图,与使用肌电信号作为运动数据相比,降低信号处理难度的同时,保持了较高的准确率;通过控制系统性能仿真与外骨骼助力性能测试实验证明了控制方法的可行性。

Aiming at the problem of response delay when the human-powered augmentation upper exoskeleton,a motion intent recognition algorithm based on data fusion of human-robot interaction force signal and pose signal is proposed.Classify,determine the wearer's movement intent,make adaptive adjustments to the control parameters,and change the response speed of the Human-robot collaboration system.The fruit fly optimization algorithm(FOA) is used to optimize the support vector machine to improve the classification accuracy;Finite state machine(FSM) is introduced to deal with unreasonable motion intents.In order to reduce the tracking error of the exoskeleton control system,a compound control method of exoskeleton combining fuzzy control and impedance control was designed to improve the update speed of control parameters and adjust the trajectory in real time.The motion intent recognition experiment was set up.The results show that the recognition accuracy can reach 97.93%,which can quickly detect unreasonable motion intents,compared to using EMG signals as motion data,reduce the difficulty of signal processing,and maintain a high accuracy.The exoskeleton boosting performance test experiment proved the feasibility of the control method.