Distributed Optimal Control of Multiple Serial Robot Systems With Kinematics and Dynamics Based on Discrete Neural Dynamics

Multiple serial robot systems (MSRS) are extensively utilized in various scenarios due to their efficiency and flexibility. The current prevailing schemes basically rely mainly on robot kinematics to control the MSRS by joint angles, velocities, or accelerations. However, industrial applications involving tasks such as grinding and cutting require the consideration of contact forces exerted by series robots, which cannot be addressed through robot kinematics alone. Furthermore, these tasks also necessitate maintaining a fixed orientation of the end-effector, which is rarely considered in current prevailing schemes for MSRS. To address these deficiencies, this article proposes a performance index-oriented distributed (PIOD) scheme for cooperative control of MSRS based on robot dynamics while addressing the issue of maintaining the orientation of the end-effector. Accordingly, a discrete neural dynamics (DND) model is proposed to solve the PIOD scheme, and then, the joint torques, utilized for controlling the MSRS, are obtained by the robot dynamics. In addition, the convergence and robustness of the DND model are confirmed through rigorous theoretical analyses. Finally, the proposed PIOD scheme is demonstrated to be feasible and practical through simulations and physical experimentations for the cooperative control of MSRS.