Distirbuted Fault Detection of Multiple Unmanned Ships Based on Fuzzy Model
Abstract: This paper proposes a distributed fault detection scheme for Multiple Unmanned Surface Vehicle Systems (MUSVSs) based on a fuzzy model. Initially, the dynamic equations of unmanned surface vehicles (USVs) are presented. Considering the communication topology among the USVs and potential thruster failures, T-S fuzzy models incorporating fault information for each unmanned surface vehicle in the MUSVS are developed. Subsequently, an observer is designed for each USV to generate a residual signal, capable of utilizing measurement information not only from the current USV but also from neighboring USVs and their corresponding observers. Additionally, a fault reference model is employed to enhance the performance of the fault detection system. A fuzzy Lyapunov function-based method and an inverse convex approach are developed, along with the introduction of a free-weighting matrix method to ensure that the obtained sufficient conditions guarantee the asymptotic stability of the fuzzy fault detection system, as well as ${H_\infty }$ performance. However, the constructed matrix inequalities contain coupling terms, which hinder the direct design of the observer, necessitating the construction of corresponding solvability conditions. Finally, a multi-agent system consisting of four USVs is established for simulation verification. In the simulation, the LMI toolbox in MATLAB is used to solve the constructed matrix inequalities, and the gain of the designed observer is determined. By comparing the instances when the residual evaluation function exceeds the threshold, the simulation results validate the effectiveness and applicability of the proposed distributed fault detection scheme.
Submission Number: 235
Loading