Adaptive Multiobjective Control With Collective Neural Dynamics for Motion Planning of Redundant Manipulators

Kinematic control of redundant manipulators plays a fundamental and crucial role in modern robotics. However, many existing motion planning schemes induce excessive joint movements by neglecting motion sparsity, leading to reduced operational efficiency. To address this problem, an adaptive multiobjective control (AMOC) scheme for redundant manipulators is proposed. At the core of the AMOC scheme lies the formulation and solution of a new multiobjective optimization problem that simultaneously minimizes the sparsity penalty and the task tracking accuracy. Critically, the inclusion of a motion sparsity objective renders this problem nonconvex, making it difficult for traditional optimization methods to solve effectively. Therefore, an adaptive collective neural dynamics (ACND) algorithm is constructed. Theoretical analyses verify the convergence of the ACND algorithm. Through simulations and physical experiments on a Franka Emika Panda manipulator, the effectiveness and superiority of the proposed AMOC scheme are demonstrated.