Finite-Time Model Predictive Tracking Control of Position and Orientation for Redundant Manipulators
Long Jin, Fan Zhang, Mei Liu, Sendren Sheng‐Dong Xu
Abstract
Most schemes of trajectory tracking control of redundant manipulators consider only the position of the end-effector with little regard for its orientation, and reduce the feasible region of decision variables when introducing multiple levels of joint limits. To remedy these weaknesses, this article constructs a model predictive tracking control of position and orientation (MPTCPO) scheme for redundant manipulators, which is able to achieve the simultaneous control of position and orientation of the end-effector. The proposed MPTCPO scheme simultaneously minimizes the tracking error, joint-velocity norm, and joint-acceleration norm. In the meantime, it directly handles three levels of joint limits of joint angle, joint velocity, and joint acceleration without reducing the feasible region of decision variables without conversion required. In addition, a finite-time convergent neural dynamics (FTCND) model is designed to find the optimal solution of the proposed scheme. Computer simulations and physical experiments confirm the feasibility and accuracy of the proposed MPTCPO scheme solved by the FTCND model.