机器人末端加速度的预测研究

Prediction Study of Robot End Acceleration

针对传统的机器人加速度动力学建模求解方法的复杂性和实时性差等问题,提出了基于小波包和改进的灰色预测模型相结合的机器人加速度的测量及预估方法。机器人末端加速度的测量值包括了加速度的真实值、噪声信号和重力加速度g的分量,计算去除重力加速度g分量,利用小波包对信号做了去噪处理,再用改进的灰色预测模型对经过处理的信号进行超前预测,解决了滤波信号的时间滞后问题,提高了加速度测量的实时性和前瞻性。仿真和实验结果验证了该方法在机器人末端加速度测量和预估方面的可行性,为机器人的加速度反馈控制及惯性力预测控制提供了基础。

For the problems of complexity and poor real-time of the traditional robot acceleration dynamics modeling solution method,the measurement and prediction method of robot acceleration based on the combination of wavelet packet and improved gray prediction model is proposed.The measured value of the robot end acceleration includes the real value of acceleration,the noise signal and the component of gravitational acceleration g.The calculation removes the component of gravitational acceleration g.The signal is denoised by using wavelet packets,and then the processed signal is over-predicted by the improved gray prediction model,which solves the time lag problem of the filtered signal and improves the real-time and forward-looking of acceleration measurement.Simulation and experimental results verify the feasibility of the method in robot end acceleration measurement and prediction,which provides a basis for acceleration feedback control and inertial force prediction control of the robot.