Redundant Manipulator Control with Constraints for Subgoals

This work analysed the general form of solution for inverse redundant kinematics (IRK) problem in geometry view, and proposed new methods to formulate the obstacle and singularity avoidance as constraints, which were incorporated into standard non-linear quadratic programming (QP) method to resolve the IRK problem. With this method, the compromise or conflict between multi optimization criteria and the complicated weighted adjustment process in weighted sum method for IRK were removed. Obstacle avoidance and physical limits, such as joint range, velocity and acceleration, were formulated as dynamic inequality constraints. For on-line inescapable singularity avoidance, the vector, according to the minimum singular value of the Jacobian matrix, was used to revise the desired tracking velocity when manipulators move near to singularity. Numerical examples show the validity and effectiveness of the proposed methods