Project

Communication Aware Multi-AUV Dynamic Coverage Path Planning

Lead researcher

Nadav Schwalb

Project based at

Lead Partner Organisation

This project explores how to support efficient, reliable multi-AUV survey missions that align with ARIAM’s broader goals of infrastructure inspection, monitoring, and digital twin development. While single-AUV surveys often follow simple lawnmower paths, scaling to multi-AUV operations introduces the challenge of Coverage Path Planning (CPP) under real-world underwater conditions. Traditional Multi-robot CPP (MCPP) methods assume perfect communication or static environments—assumptions that break down in the harsh acoustic conditions underwater, where bandwidth is limited, errors are frequent, and communication quality depends on range, orientation, and vehicle dynamics. 

This project aims to research and develop novel dynamic path planning algorithms for multi-AUV coverage under communication constraints. The outcome will be demonstrated in real-world survey missions with AUVs and surface support. A connected AUV fleet could efficiently survey a region of interest using homogenous sensors to create a multi-modal high-resolution map as well as dynamically allocating new tasks to AUVs during operation. This research closes a critical gap by coupling communication-aware planning with cooperative autonomy, directly supporting ARIAM’s mission to improve infrastructure inspection and monitoring in complex marine environments. 

Research Activities: 

  • Building simulation infrastructure – creating a virtual environment to test and evaluate multi-AUV strategies under realistic conditions such as noise, currents, and communication limits. 
  • Extending AUV capabilities for mesh communication – enabling vehicles to exchange information directly with each other and the surface vessel to support cooperative missions. 
  • Exploring Multi-robot Coverage Path Planning (MCPP) methods – investigating approaches like Artificial Potential Fields, Monte Carlo Search Trees, Multi-Agent Reinforcement Learning, and Genetic Algorithms to improve how multiple AUVs coordinate their survey paths dynamically. 

Expected Impact: 

This research will enable more efficient and reliable underwater infrastructure inspection by allowing multiple AUVs to coordinate surveys dynamically, reducing mission time, energy use, and operational costs. By developing communication-aware planning methods, fleets will remain effective even in complex underwater environments, improving both data quality and robustness. AUV Fleets with robust communication will be better equipped to dynamically allocate new tasks and adapt to changing or unknow environments. These advancements will make large-scale inspection tasks such as ports, pipelines, and offshore energy facilities more feasible, while also supporting broader applications in marine science, environmental monitoring, and digital twin creation. 

 

 

Associated Researchers