Decentralized Non-communicating Multiagent Collision Avoidance with Deep\n Reinforcement Learning

2016/09/26 by Yu Fan Chen, Chen, Yu Fan, Miao Liu +5 · 16 citations
Computer Science · #FOS: Computer and information sciences #Multi-Agent Systems and Negotiation #Multiagent Systems (cs.MA) #Reinforcement Learning in Robotics #Robotic Path Planning Algorithms

paper · pdf · doi:10.48550/arxiv.1609.07845

openalex publication_date 2016/09/26 · openalex created_date 2022/10/02 · openalex updated_date 2026/07/28

Abstract

Finding feasible, collision-free paths for multiagent systems can be\nchallenging, particularly in non-communicating scenarios where each agent's\nintent (e.g. goal) is unobservable to the others. In particular, finding time\nefficient paths often requires anticipating interaction with neighboring\nagents, the process of which can be computationally prohibitive. This work\npresents a decentralized multiagent collision avoidance algorithm based on a\nnovel application of deep reinforcement learning, which effectively offloads\nthe online computation (for predicting interaction patterns) to an offline\nlearning procedure. Specifically, the proposed approach develops a value\nnetwork that encodes the estimated time to the goal given an agent's joint\nconfiguration (positions and velocities) with its neighbors. Use of the value\nnetwork not only admits efficient (i.e., real-time implementable) queries for\nfinding a collision-free velocity vector, but also considers the uncertainty in\nthe other agents' motion. Simulation results show more than 26 percent\nimprovement in paths quality (i.e., time to reach the goal) when compared with\noptimal reciprocal collision avoidance (ORCA), a state-of-the-art collision\navoidance strategy.\n

Cited by

• Intrinsic-Motivation Multi-Robot Social Formation Navigation with Coordinated Exploration
• SINRL: Socially Integrated Navigation with Reinforcement Learning using Spiking Neural Networks
• A New Trajectory-Oriented Approach to Enhancing Comprehensive Crowd Navigation Performance
• Deadlock-Free Hybrid RL-MAPF Framework for Zero-Shot Multi-Robot Navigation
• Symmetry-Breaking in Multi-Agent Navigation: Winding Number-Aware MPC with a Learned Topological Strategy
• CoRL-MPPI: Enhancing MPPI With Learnable Behaviours For Efficient And Provably-Safe Multi-Robot Collision Avoidance
• Hybrid DQN-TD3 Reinforcement Learning for Autonomous Navigation in Dynamic Environments
• A short methodological review on social robot navigation benchmarking
• Integrating Offline Pre-Training with Online Fine-Tuning: A Reinforcement Learning Approach for Robot Social Navigation
• Learning Social Robot Navigation By Sensing Human Legs
• Learning Social Heuristics for Human-Aware Path Planning
• Multi-Agent Reinforcement Learning in Intelligent Transportation Systems: A Comprehensive Survey
• Sequential Convex Programming with Filtering-Based Warm-Starting for Continuous-Time Multiagent Quadrotor Trajectory Optimization
• Multi-Robot Navigation in Social Mini-Games: Definitions, Taxonomy, and Algorithms
• Deep Reinforcement Learning with anticipatory reward in LSTM for Collision Avoidance of Mobile Robots
• Towards Generalizable Safety in Crowd Navigation via Conformal Uncertainty Handling

Related

• Multi-agent navigation based on deep reinforcement learning and traditional pathfinding algorithm
• Reciprocal Collision Avoidance for General Nonlinear Agents using Reinforcement Learning
• Towards Optimally Decentralized Multi-Robot Collision Avoidance via Deep Reinforcement Learning
• Motion Planning Among Dynamic, Decision-Making Agents with Deep Reinforcement Learning
• Fully Distributed Multi-Robot Collision Avoidance via Deep Reinforcement Learning for Safe and Efficient Navigation in Complex Scenarios