康复训练患肢动力学模型参数辨识算法的研究

摘要

Because rehabilitants have various characteristics, their limb' s mass and strength are different , and the muscle spasm may happen during the rehabilitation training. All these affect the rehabilitation training robot system' s stability and operability as well as the smoothness of the slave' s movement, thus the rehabilitation training efficiency is lowered. To solve these problems, a new kinetics parameter I-dentification algorithm was presented by solving over-determined equations, according to the training limb' s force and position information. The simulation results have proven the effectiveness of the methods and its robustness under certain noise circumstances. By using this parameter identification algorithm, reliable information can be provided for rehabilitation training robot system.%康复训练患者的病情千差万别,不同的患者有不同质量的患肢和不同的肌张力甚至可能发生肌肉痉挛,这些都将严重影响康复训练机器人系统的稳定性、可操作性以及从手运动的平滑性,降低了康复训练效率.为了解决这些问题,根据患肢训练时力和位置等信息反馈,提出一种用解超定方程组的方法在线辨识患肢的动力学参数,实现了患肢动力学模型的在线辨识.仿真结果证明该方法辨识患肢动力学参数效果良好,并且在一定的噪声范围内具有较强的鲁棒性,可以用于康复训练机器人的患肢动力学参数辨识,为实际康复训练机器人系统的实时控制提供了较为准确的依据.