A Hierarchical Dynamic Obstacle Avoidance Strategy Based on Decision-Making and Control Architecture for Cable-Driven Continuum Robots

Abstract

This article presents a hierarchical decision-making and control architecture for cable-driven continuum robots (CRCRs) operating in complex dynamic obstacle scenarios. At the decision-making level, a finite state machine (FSM) is designed to generate appropriate driving signals (such as avoidance, overtaking, and deceleration) in response to position changes of dynamic obstacles. These signals are then transmitted to a two-layer model predictive control (MPC) system at the control level. The two-layer MPC, integrated with a super twist observer (STO), ensures the generation of safe and collision-free trajectories as well as accurate and fast trajectory tracking. This hierarchical approach leverages the adaptability of the FSM-based decision-making module and the robustness of the two-layer MPC control to achieve high performance of CDCRs in complex scenarios. The effectiveness and superior performance of this framework have been validated through experiments across a range of dynamic obstacle scenarios.