Singularity Avoidance of Humanoid Manipulator Based on Workspace Decomposition

Aiming at the singular problems in the humanoid manipulator. This article is designed with a set of singularity avoidance algorithm that allows the manipulator to through the singular region smooth and maintain a higher cartesian trajectory tracking accuracy. Firstly, on the basis the workspace decomposition of the manipulator, the internal singularity and the boundary singularity in the singular subspace are further distinguished by primary transformation. Then, the algorithms are designed respectively based on the switching strategy: for internal singularities, the null space method is adopted, and for the boundary singularities, the control signal is modified to let the manipulator escape from the singular region. The proposed algorithm increases the effective workspace and ensures the tracking accuracy of the cartesian trajectory on the basis of ensuring the stable operation. Simulation results show the advantages and effectiveness of the method.