Four-Criterion-Optimization-Based Coordination Motion Control of Dual-Arm Robots

Abstract

In order to address the coordination constraints and physical constraints subjected to the dual-arm robot simultaneously, a novel four-criterion-optimization coordination motion (FCOCM) scheme is proposed, which combines four optimization criteria, namely, the repetitive motion planning (RMP), minimum velocity norm (MVN), maximize manipulability (MM), and infinity-norm velocity minimization (INVM). The scheme can remedy discontinuity in the INVM scheme, eliminate joint angular drift phenomenon, prevent the high joint angular velocity and MM, thereby improving the motion efficiency and ensuring safety and accuracy in the process of tasks. Besides, this scheme also considers real-time trajectory feedback, satisfies physical constraints, and ensures the joint angular velocity is zero at the end of tasks. Furthermore, the improved FCOCM scheme is solved by a novel power-exponent-type variable-parameter recurrent neural network (PET-VPNN) model proposed in this article. A novel Sinh-tunable type activation function which achieves better convergence performance is also proposed. Simulations and experiment are presented to verify the superiority of the proposed coordination motion control method. This research is of great significance for the coordination motion control of dual-arm robots in complex path planning tasks.