Hierarchical Attack-Defense Strategy of Unmanned Surface Vehicles for Domain Protection and Invasion Task
Authors:
- 17 June 2025
- journal article
- Published by World Scientific Pub Co Pte Ltd in Unmanned Systems
- Vol. 14 (04) , 819-838
Abstract
This paper introduces an innovative hierarchical framework tailored to meet the offensive and defensive needs of swarm-based unmanned surface vehicles (USVs) for domain protection and invasion tasks. The framework is structured into three layers: The top layer handles target allocation, the middle layer focuses on action strategy, and the bottom layer ensures safety. For domain protection tasks, the top layer leverages auction theory to effectively match attackers and defenders, enabling the distributed allocation of multiple defenders to respective invaders. The middle layer employs an adaptive guidance method, allowing each defender to follow a guidance law aligned with the optimal interception point. The bottom layer uses an enhanced potential field function to guarantee collision-free interception of invaders. For invasion tasks, the top layer similarly applies auction theory for the allocation of virtual waypoints. The middle layer introduces an Artificial Potential Field (APF)-based angle-adaptive attack strategy, enabling invaders to evade defender interception. The bottom layer remains consistent with the domain protection framework. This paper marks the first hybrid virtual-real experiment in multi-USV confrontation, demonstrating its reliability and effectiveness in real maritime conditions. Experiments conducted within a 5000-m radius, with vessels traveling at speeds up to 12 m/s, confirmed the framework’s robustness and practical applicability in near-realistic combat scenarios.Keywords
Funding Information
- National Natural Science Foundation of China (62325302)
- National Natural Science Foundation of China (62173220)
This publication has 30 references indexed in Scilit:
- Large-scale UAV swarm confrontation based on hierarchical attention actor-critic algorithmApplied Intelligence, 2024
- H3E: Learning air combat with a three-level hierarchical framework embedding expert knowledgeExpert Systems with Applications, 2023
- Pursuit-evasion game strategy of USV based on deep reinforcement learning in complex multi-obstacle environmentOcean Engineering, 2023
- Robust formation tracking of multiple autonomous surface vessels with individual objectives: A noncooperative game-based approachControl Engineering Practice, 2022
- Discovering optimal strategy in tactical combat scenarios through the evolution of behaviour treesAnnals of Operations Research, 2021
- A deep reinforcement learning-based method applied for solving multi-agent defense and attack problemsExpert Systems with Applications, 2021
- Multi-agent hierarchical policy gradient for Air Combat Tactics emergence via self-playEngineering Applications of Artificial Intelligence, 2021
- Path-guided time-varying formation control with collision avoidance and connectivity preservation of under-actuated autonomous surface vehicles subject to unknown input gainsOcean Engineering, 2019
- Discovering Emergent Agent Behaviour with Evolutionary Finite State Machines2018
- Concise deep reinforcement learning obstacle avoidance for underactuated unmanned marine vesselsNeurocomputing, 2018