A wavelet approximation method for dynamic analysis and parameters identification of series robot with high degree of freedom

Multidegree of freedom series robots are used in various fine work scenarios. The redundant DOF (short for degree of freedom) robots have excess degrees of freedom and can easily avoid obstacles, their own singularities, and joint limits. But the dynamics model of redundant DOF robot is complicated, difficult to solve, and slow to calculate. These are not conducive to the monitoring and analysis of the robot's state and parameters. A method of robot dynamics calculation based on wavelet approximation is proposed in this article. First, the robot dynamics model is established based on Lagrange method. Then, a multidimensional analysis model of robot dynamics wavelet approximation is established. Next, a 6-DOF UR5 robot is taken as an object to compare these methods. The six-dimensional wavelet approximation method has high accuracy with a long calculation time. One-dimensional wavelet approximation method has the best performance both in computational efficiency. The higher the degree of freedom of the robot, the greater the advantage of the wavelet analysis method because it does not need to solve the exponential increasing matrix in the dynamics model. In the identification and analysis of robot load and dynamic parameters, one-dimensional wavelet approximation method has the accuracy close to the dynamic model, and the minimum identification error is within 9.04%. The proposed method provides a new technical approach for dynamic modeling and control of high degree of freedom robots.