机器人负载的动力学参数辨识

为解决机器人末端负载的时变性给高速运动的机器人带来控制精度降低的问题,研究了参数差值法、力矩求解法、全局参数辨识法的机器人末端负载动力学参数辨识的方法,以提高末端负载的辨识精度.得到的负载动力学参数用于动力学控制以提高机器人动态精度.通过建立拉格朗日动力学线性辨识模型,以最优激励轨迹进行实时数据采集,采样数据经过低通滤波及中心差分的处理后,代入相应的负载辨识方程式,并用加权最小二乘法解决线性方程组,可辨识到不同负载的动力学参数.实验验证了负载辨识方法的可行性.

Identification methods for robot payload dynamical parameters

To solve the problem of low precision control in high speed motion robots due to a time-varying payload and to improve the identification accuracy of the payload, the dynamical parameter identification methods for robot end payload were studied. These methods included the parameter difference method, the moment method, and the global parameter identification method. The parame-ters were used to improve the dynamic precision in robot dynamic control. The dynamical parameters for different payloads can be iden-tified with the Lagrange dynamic linear identification model, which acquires real-time data using the optimal incentive path and solves the equations by using the least square weight algorithm after the sample data was processed with a low-pass filter and central difference processes. The feasibility of the method was verified by experiments.