Session: 15-01-02 Seakeeping and Operability

Submission Number: 176096

Power Damage State Identification for Unmanned Surface Vehicles Using Motion Data 

Unmanned surface vehicles (USVs) are valued for their cost-effectiveness, maneuverability, and operational efficiency in maritime missions. However, their power systems are critical yet vulnerable components; once damage or failure occurs, the vehicle’s navigation and task execution are seriously affected. Traditional fault diagnosis methods for manned vessels are unsuitable for USVs due to their autonomous operation and distributed control. This study proposes a motion data based method for identifying power damage states in small and medium-sized USVs. By analyzing the differences between unmanned and manned vehicles, the study establishes a motion propulsion mapping model derived from measured data to correlate motion behavior with power conditions. An outlier detection algorithm ensures data validity, while the concept of damage degree quantifies propulsion degradation. A deep neural network based dynamic identification algorithm is developed to improve recognition accuracy and generalization under varying operating conditions. Experiments under controlled damage scenarios verify the proposed approach. The results show that the algorithm accurately identifies different power damage levels and adjusts thruster commands in real time to maintain direction and limited maneuverability. The proposed method demonstrates strong applicability for motion based fault diagnosis and provides an effective foundation for intelligent fault management and resilient control of unmanned surface vehicles.

Presenting Author: Mengyuan Xu Shanghai Jiao Tong University

Presenting Author Biography: Xu Mengyuan is a Ph.D. candidate at Shanghai Jiao Tong University. His research focuses on intelligent ship maneuvering modeling, advised by Professor Wang Hongdong. Current research progress involves integrating hydrodynamic principle models with data-driven methods to address the key challenge of predicting ship behavior under uncertain environmental disturbances.

Authors:

Mengyuan Xu Shanghai Jiao Tong University
Jiankun Lou Shanghai Jiao Tong University
Mingyang Zhang Shanghai Jiao Tong University
Yongjin Guo Shanghai Jiao Tong University
Hongdong Wang Shanghai Jiao Tong University