Multi-UAV Conflict Resolution with Graph Convolutional Reinforcement\n Learning

2021/11/29 by Ralvi Isufaj, Isufaj, Ralvi, Marsel Omeri +3
Engineering · #Air Traffic Management and Optimization #Autonomous Vehicle Technology and Safety #FOS: Computer and information sciences #Machine Learning (cs.LG) #Multiagent Systems (cs.MA) #Robotics (cs.RO) #Traffic control and management

paper · pdf · doi:10.48550/arxiv.2111.14598

openalex publication_date 2021/11/29 · openalex created_date 2022/11/07 · openalex updated_date 2026/07/28

Abstract

Safety is the primary concern when it comes to air traffic. In-flight safety\nbetween Unmanned Aircraft Vehicles (UAVs) is ensured through pairwise\nseparation minima, utilizing conflict detection and resolution methods.\nExisting methods mainly deal with pairwise conflicts, however due to an\nexpected increase in traffic density, encounters with more than two UAVs are\nlikely to happen. In this paper, we model multi-UAV conflict resolution as a\nmulti-agent reinforcement learning problem. We implement an algorithm based on\ngraph neural networks where cooperative agents can communicate to jointly\ngenerate resolution maneuvers. The model is evaluated in scenarios with 3 and 4\npresent agents. Results show that agents are able to successfully solve the\nmulti-UAV conflicts through a cooperative strategy.\n

Related

• Conflict Resolution under Degraded Surveillance in Air Corridors Using Multi-Agent Reinforcement Learning
• Graph Attention-based Reinforcement Learning for Trajectory Design and Resource Assignment in Multi-UAV Assisted Communication
• Automating the resolution of flight conflicts: Deep reinforcement learning in service of air traffic controllers
• Multi-Agent Reinforcement Learning for Unmanned Aerial Vehicle Coordination by Multi-Critic Policy Gradient Optimization
• A Hierarchical Reinforcement Learning Framework for Multi-UAV Combat Using Leader-Follower Strategy