Research on dynamic parameter identification of space hyper-redundant manipulators using iterative WLS-TCS algorithm

ObjectiveIn the dynamic modeling of space hyper-redundant manipulators, the complicated structure, many degrees of freedom and weak rigidity cause the dynamic coupling problems, and it is difficult to obtain an accurate dynamic model. An iterative WLS-TCS algorithm was proposed to identify the ground dynamics parameters of the space hyper-redundant manipulator, which laid a foundation for obtaining the high-precision dynamic model of the manipulator and the research of space in-orbit dynamic control.MethodsFirstly, the terminal crossover and steering-based particle swarm optimization (TCS-PSO) algorithm was used to design periodic Fourier series satisfying multiple constraints as the optimal excitation trajectory. Secondly, the iterative weighted least squares （IWLS） method was used to obtain the minimum parameter set, and the outliers in the data were gradually removed by the iterative weighting to make the identification results more robust and accurate.ResultsThe test results show that the trajectory regression matrix obtained by the TCS optimization method is smaller and better satisfies the constraints given in the excitation trajectory. In parameter identification, the iterative weighted least squares identification method is compared with the recursive least squares method, and the torque residual root mean square (RMS) value is reduced by about 2.22% on average. Compared with the weighted least squares method, the torque residual RMS value is reduced by about 4.85% on average. Finally, the obtained parametric model was brought into the zero-force control test, and the actual effect was in line with expectations.