Evaluating sensor performance and configuration for autonomous operation of small ferries in maritime environments

Abstract Autonomous ferries offer a promising solution to challenges in public waterway transport, such as crew shortage and environmental impact. Safe navigation requires reliable sensor performance, particularly in adverse maritime conditions like fog, rain, and low visibility. This paper evaluates LiDAR, mmWave radar, and infrared/RGB cameras under real-world conditions aboard a medium-sized autonomous ferry prototype. A virtual test field in Unreal Engine 5 was used to assess sensor configurations and field trials validated performance in varying weather conditions. Results show LiDAR’s range and density loss in fog and rain, the robustness of mmWave radar for long-range detection, and the benefit of pairing it with PTZ-mounted infrared cameras for improved tracking in low-visibility scenarios. We propose an optimized multi-sensor configuration combining these technologies to maximize perception accuracy. By linking simulation and experimental findings, this study provides actionable recommendations for weather-resilient perception systems and informs sensor fusion strategies for small to medium-sized autonomous vessels.