Multiple sensor fault-tolerant predictive control for autonomous surface vehicle formation.

Area of Science:

• Robotics and Control Systems
• Marine Engineering
• Fault-Tolerant Systems

Background:

• Cooperative control of Autonomous Surface Vehicles (ASVs) is challenged by multiple sensor faults, impacting navigation and formation stability.
• Existing methods struggle to maintain robust formation control under dynamic and multiple sensor degradation scenarios.

Purpose of the Study:

• To propose a predictive fault-tolerant control strategy for ASV formation path following under multiple sensor faults.
• To enhance the reliability and accuracy of ASV cooperative control systems in the presence of sensor anomalies.

Main Methods:

• Development of a Distributed Extended Kalman Filter (DEKF) state estimator integrating fault detection and auxiliary data from other ASVs.
• Implementation of a Model Predictive Control (MPC) controller utilizing DEKF-estimated states for robust trajectory tracking.
• Establishment of an ASV formation path following model using a virtual leader-follower structure.

Main Results:

• The proposed DEKF-MPC approach demonstrated superior performance compared to APF-MPC and RANSAC-EKF-MPC in simulations.
• Accurate trajectory tracking and consistent heading maintenance were achieved by ASVs, even with multiple sensor faults.
• The DEKF state estimator effectively estimated ASV position, Speed Through Water (STW), and current speed under fault conditions.

Conclusions:

• The DEKF-MPC strategy provides an effective solution for fault-tolerant cooperative control of ASV formations.
• This approach significantly improves the robustness and reliability of ASV systems operating in complex marine environments with sensor uncertainties.