Discrete‐Time Neural Network Based Mode Switching Optimal Control Scheme for Mobile Dual‐Manipulator Robots

ABSTRACT In recent years, the mobile dual‐manipulator cooperative robot (MDMCR) has become a significant research platform for modern industrial applications due to their excellent adaptability and flexibility. However, the highly anthropomorphic structure and high motion redundancy pose numerous challenges in developing a high‐precision control scheme that is trackable, predictable, and adaptable in complex scenarios. Most research focuses on single‐task operations, lacking cohesive strategies for multi‐task coordination in dual‐manipulator robots. This paper presents a detailed mathematical model and kinematic analysis of the MDMCR, along with a novel coordination strategy designed to optimize cooperative precision and prevent self‐collision between the arms within a switching mode based on a cooperation index. Additionally, an innovative control scheme for mode switching of the MDMCR is established by incorporating kinematic constraints, joint limit constraints, and obstacle avoidance into a dynamic quadratic programming (DQP) problem, which is efficiently solved using discrete‐time primal‐dual neural network (DT‐PDNN). Through numerical experiments under two modes and comparative experiments with the traditional method, the effectiveness of the cooperation strategy based on the cooperation index has been verified.