基于可穿戴机械手臂参数估计与动力学模型分析

Parameter Estimation and Dynamical Model Analysis Based on Wearable Mechanical Arm

针对参数估计方法中存在忽略个人差异和测量设备昂贵的问题,提出了一种人体手臂惯性矩、摩擦和肌肉强度等参数的动力学建模和估计方法。建立人体手臂的动力学模型,在同时进行多关节估计的情况下,使用可穿戴机器人在人体关节附近布置驱动器,将输入应用到多个关节上。根据相关的输入和响应,从而实现对物理参数的估计。结果表明:相比单关节,所提方法考虑个人差异,采用多输入特征,既能减少误差,只有10%的位置误差出现在重力项。又可以缩短估计时间,同时估计三关节节省约6.8 s的时间,并且不需要昂贵的精密测量设备。因此,采用所提方法估计参数准确性较高,性能较优越。

In view of the problems of ignoring individual differences and expensive mesuring equipment in parameter estimation methods,dynamic modeling and estimation method of human arm parameters such as the moment of inertia,friction,and muscle strength are proposed,and a dynamic model of the human arm is established.In the case of simultaneous multi-joint estimation,a wearable robot is used to place actuators near the human joints and apply the input to multiple joints.Based on the relevant inputs and responses,the physical parameters can be estimated.Simulation and experimental results show that compared with a single joint,the proposed method can reduce the error,only 10%of the position error occurs in the gravity term,and it can shorten the estimation time,the three-joint estimation achieves about 6.8s faster,and it does not need expensive precision measurement equipment.The proposed method has better accuracy and performance in parameter estimation by considering individual differences and adopting multi-input characteristics.