USV-Tracker: A Novel USV Tracking System for Surface Investigation with Limited Resources USV-Tracker：面向有限资源水面探测任务的新型无人艇跟踪系统

This paper introduces USV-Tracker, a tracking system for Unmanned Surface Vehicles (USVs) designed for surface investigation and target tracking. The system addresses perception robustness, tracking concealment, and planning efficiency. The work includes: (1) a multi-sensor fusion framework using an Extended Kalman Filter (EKF) for target detection and positioning with cameras, LiDAR, GPS, and IMU sensors; (2) a two-stage path-planning algorithm that generates occlusion-avoidance trajectories and uses a virtual elastic-force constraint to maintain relative positioning, with additional FOV orientation constraints in dense obstacle environments; and (3) a visibility-aware control strategy that maintains target observability through EKF-based trajectory prediction. Gazebo simulations and physical experiments evaluate the system under representative scenarios. The computational workload runs on a constrained on-board computer.