Dual-Criteria Robot Control: Optimization with End-Effector Orientation Constraints

A dual-criterion control scheme for redundant robotic arms is introduced in this paper, incorporating end-effector orientation control to address the challenges of high computational complexity and the absence of orientation tracking in existing neural network-based solutions. The proposed scheme employs a training-free neural dynamic algorithm (NDA) to manage both orientation-tracking and physical constraints simultaneously. Notably, the NDA solver exhibits lower computational complexity compared to current methods addressing similar tasks. In view of this, we leverage the NDA solver and achieve global and exponential convergence to the theoretical solution for robotic motion generation. Furthermore, to validate and illustrate the effectiveness and feasibility of our proposed control scheme, numerical and virtual simulations using a Franka Emika Panda manipulator are conducted.