Disturbance-observer-based adaptive neuro-fuzzy sliding mode control for collaborative ASV and URV systems

Engineering notes

The controller parameters are adaptively tuned online using an improved grey wolf optimizer with chaotic theory (IGWO-CT), whose optimization performance is first evaluated using standard benchmark functions. Quantitative results show that the proposed controller achieves yaw-rate tracking RMSE values of 0.00956, 0.01915, and 0.00185 rad/s for the ASV, URV1, and URV2, respectively.

Chinese explanation / 中文解读

中文解读待补充：本站会优先为端到端自动驾驶、BEV感知、3D目标检测、轨迹预测、路径规划、LiDAR感知等高价值论文补充中文说明。

Original abstract

This paper proposes a disturbance-observer-based adaptive neuro-fuzzy sliding mode control (DO-ANFSMC) framework for collaborative autonomous surface vehicle (ASV) and underwater robotic vehicle (URV) systems operating in search-and-rescue missions under environmental disturbances and obstacle-avoidance constraints. The proposed framework integrates an adaptive neuro-fuzzy system for approximating unknown nonlinear dynamics, a nonlinear disturbance observer for estimating lumped environmental disturbances and residual uncertainties, and a robust sliding-mode control law for trajectory tracking. The controller parameters are adaptively tuned online using an improved grey wolf optimizer with chaotic theory (IGWO-CT), whose optimization performance is first evaluated using standard benchmark functions. MATLAB simulations are conducted for one ASV and two URVs under identical mission conditions with eight static obstacles. Quantitative results show that the proposed controller achieves yaw-rate tracking RMSE values of 0.00956, 0.01915, and 0.00185 rad/s for the ASV, URV1, and URV2, respectively. The proposed method also reduces the travelled path length to 68.781 m, 68.171 m, and 79.399 m for the ASV, URV1, and URV2, respectively, while maintaining safe obstacle clearance and negligible final target error. Comparative and ablation results show that the combined use of disturbance observation, adaptive neuro-fuzzy approximation, sliding-mode robustness, and IGWO-CT-based parameter tuning improves trajectory-tracking accuracy, disturbance rejection, and navigation efficiency for collaborative marine robotic systems.

Links and sources

• PDF from original source

Comments