Depth-Constrained ASV Navigation with Deep RL and Limited Sensing

2025-04-25

AI Summary

Autonomous surface vehicles (ASVs) operating in shallow water face significant navigation challenges due to sparse depth measurements from single-beam echosounders (SBES), dynamic environmental disturbances, and stringent bathymetric safety constraints. Method: This paper proposes a deep reinforcement learning (DRL) framework integrating Gaussian process regression (GPR) with proximal policy optimization (PPO). GPR is embedded within the DRL closed loop to enable online reconstruction of high-confidence bathymetric maps from sparse sonar data, facilitating efficient sim-to-real transfer. Contribution/Results: Evaluated in both simulation and full-scale field experiments, the method substantially improves generalization and safety: bathymetric constraint violations decrease by 92%, task completion rate reaches 96.7%, and hazardous shallow zones and obstacles are reliably avoided.

Technology Category

• Search and Optimization: Learning to Search
• Planning, Routing, and Scheduling: Learning for Planning and Scheduling
• Intelligent Robots: State Estimation

Application Category

• Search and Retrieval-Augmented AI: Vertical and domain-specific search
• Systems and Infrastructure for Web, Mobile and WoT: Applied ML and AI for Web-based mobile applications
• Responsible Web: Machine-in-the-loop, human agency and autonomy

Abstract

Autonomous Surface Vehicles (ASVs) play a crucial role in maritime operations, yet their navigation in shallow-water environments remains challenging due to dynamic disturbances and depth constraints. Traditional navigation strategies struggle with limited sensor information, making safe and efficient operation difficult. In this paper, we propose a reinforcement learning (RL) framework for ASV navigation under depth constraints, where the vehicle must reach a target while avoiding unsafe areas with only a single depth measurement per timestep from a downward-facing Single Beam Echosounder (SBES). To enhance environmental awareness, we integrate Gaussian Process (GP) regression into the RL framework, enabling the agent to progressively estimate a bathymetric depth map from sparse sonar readings. This approach improves decision-making by providing a richer representation of the environment. Furthermore, we demonstrate effective sim-to-real transfer, ensuring that trained policies generalize well to real-world aquatic conditions. Experimental results validate our method's capability to improve ASV navigation performance while maintaining safety in challenging shallow-water environments.

Problem

• ASV navigation in shallow water with depth constraints
• Limited sensing using single-beam echosounder measurements
• Safe path planning under dynamic disturbances

Innovation

• Deep RL for ASV navigation with depth constraints
• Gaussian Process regression for sparse sonar data
• Effective sim-to-real transfer for real-world conditions