Dynamic Obstacle Avoidance of Unmanned Surface Vehicles in Maritime Environments Using Gaussian Processes Based Motion Planning

2024-12-10

arXiv.org

AI Summary

Unmanned surface vehicles (USVs) operating in open waters must simultaneously avoid static obstacles (e.g., islands, reefs) and dynamic obstacles (e.g., other vessels), posing significant challenges for real-time, safe motion planning. Method: This paper proposes a Gaussian process (GP)-based motion planning framework that fuses multi-source obstacle information. It extends GPMP2 by embedding a real-time state-estimated multivariate Gaussian distribution within a factor graph optimization framework, enabling dynamic obstacle modeling and online replanning. The approach integrates GP-based trajectory representation, factor graph optimization, and multivariate uncertainty propagation. Results: Evaluated in a high-fidelity ROS maritime simulator and MATLAB benchmarks, the method achieves a 23.6% improvement in dynamic obstacle avoidance success rate and reduces trajectory jerk (integrated jerk) by 31.4%, while satisfying both real-time performance and safety requirements.

Technology Category

• Planning, Routing, and Scheduling: Replanning and Plan Repair
• Intelligent Robots: Motion and Path Planning
• Multiagent Systems: Multiagent Planning

Application Category

• Graph Algorithms and Modeling for the Web: Efficient manipulation of static and dynamic Web-related graphs
• Systems and Infrastructure for Web, Mobile and WoT: Experiences and lessons learnt from Web-based algorithms and system deployments
• User Modeling, Personalization and Recommendation: Attacks and countermeasures in recommendation systems

Abstract

During recent years, unmanned surface vehicles are extensively utilised in a variety of maritime applications such as the exploration of unknown areas, autonomous transportation, offshore patrol and others. In such maritime applications, unmanned surface vehicles executing relevant missions that might collide with potential static obstacles such as islands and reefs and dynamic obstacles such as other moving unmanned surface vehicles. To successfully accomplish these missions, motion planning algorithms that can generate smooth and collision-free trajectories to avoid both these static and dynamic obstacles in an efficient manner are essential. In this article, we propose a novel motion planning algorithm named the Dynamic Gaussian process motion planner 2, which successfully extends the application scope of the Gaussian process motion planner 2 into complex and dynamic environments with both static and dynamic obstacles. First, we introduce an approach to generate safe areas for dynamic obstacles using modified multivariate Gaussian distributions. Second, we introduce an approach to integrate real-time status information of dynamic obstacles into the modified multivariate Gaussian distributions. Therefore, the multivariate Gaussian distributions with real-time statuses of dynamic obstacles can be innovatively added into the optimisation process of factor graph to generate an optimised trajectory. The proposed Dynamic Gaussian process motion planner 2 algorithm has been validated in a series of benchmark simulations in the Matrix laboratory and a dynamic obstacle avoidance mission in a high-fidelity maritime environment in the Robotic operating system to demonstrate its functionality and practicability.

Problem

• Avoiding static and dynamic obstacles for unmanned surface vehicles
• Integrating real-time obstacle data into motion planning
• Ensuring compliance with maritime collision prevention regulations

Innovation

• Uses Gaussian processes for dynamic motion planning
• Integrates real-time obstacle status into Gaussian distributions
• Incorporates maritime collision regulations into algorithm