Affiliation:
1. Aliyun School of Big Data, Changzhou University, Changzhou 213164, China
Abstract
In this paper, we consider the cooperative decision-making problem for multi-target tracking in multi-agent systems using multi-agent deep reinforcement learning algorithms. Multi-agent multi-target pursuit has faced new challenges in practical applications, where pursuers need to plan collision-free paths and appropriate multi-target allocation strategies to determine which target to track at the current time for each pursuer. We design three feasible multi-target allocation strategies from different perspectives. We compare our allocation strategies in the multi-agent multi-target pursuit environment that models collision risk and verify the superiority of the allocation strategy marked as POLICY3, considering the overall perspective of agents and targets. We also find that there is a significant gap in the tracking policies learned by agents when using the multi-agent reinforcement learning algorithm MATD3. We propose an improved algorithm, DAO-MATD3, based on dynamic actor network optimization. The simulation results show that the proposed POLICY3-DAO-MATD3 method effectively improves the efficiency of completing multi-agent multi-target pursuit tasks.
Funder
Changzhou Municipal Advanced Technologies Research Center program
Changzhou Sci & Tech Program
Subject
Electrical and Electronic Engineering,Computer Networks and Communications,Hardware and Architecture,Signal Processing,Control and Systems Engineering
Reference31 articles.
1. Decentralized Multi-Agent Pursuit Using Deep Reinforcement Learning;Newbury;IEEE Robot. Autom. Lett.,2021
2. Group chasing tactics: How to catch a faster prey;Janosov;New J. Phys.,2017
3. Optimal Base Station Scheduling for Device-to-Device Communication Underlaying Cellular Networks;Li;IEEE J. Sel. Areas Commun.,2015
4. Collective Predation and Escape Strategies;Angelani;Phys. Rev. Lett.,2012
5. Xie, F., Botea, A., and Kishimoto, A. (2017, January 19–25). A Scalable Approach to Chasing Multiple Moving Targets with Multiple Agents. Proceedings of the Twenty-Sixth International Joint Conference on Artificial Intelligence, Melbourne, VIC, Australia.