LAUSR

In this paper, we propose a recurrent neural network (RNN) for the tracking control of surgical robots while satisfying remote center-of-motion (RCM) constraints. RCM constraints enforce rules suggesting that the surgical tip should not go beyond the region of incision while tracking the commands of the surgeon. Violations of RCM constraints can result in serious injury to the patient. We unify the RCM constraints with the tracing control by formulating a single constrained optimization problem using a penalty-term approach. The penalty-term actively rewards the optimizer for satisfying the RCM constraints. We then propose an RNN-based metaheuristic optimization algorithm called “Beetle Antennae Olfactory Recurrent Neural Network (BAORNN)” for solving the formulated optimization problem in real time. The proposed control framework can track the surgeon’s commands and satisfy the RCM constraints simultaneously. Theoretical analysis is performed to demonstrate the stability and convergence of the BAORNN algorithm. Simulations using LBR IIWA14, a 7-degree-of-freedom robotic arm, are performed to analyze the performance of the proposed framework.