Scholarlink

<p>Autonomous vehicles are suited for continuous area patrolling problems. However, finding an optimal patrolling strategy can be challenging for many reasons. Firstly, patrolling environments are often complex and can include unknown environmental factors, such as wind or landscape. Secondly, autonomous vehicles can have failures or hardware constraints, such as limited battery life. Importantly, patrolling large areas often requires multiple agents that need to coordinate their actions collectively. In this work, we consider these limitations and propose a homogeneous multi-agent architecture approach based on model-free, deep multi-agent reinforcement learning. Patrolling agents execute identical policies locally based on their local observations and shared location information. Agents can automatically recharge themselves to support continuous collective patrolling. The patrolling system can tolerate agent failures and allow adding supplementary agents to replace failed agents or increase the overall patrol performance. The overall patrol performance, the efficiency of battery recharging strategies, the overall fault tolerance, and the ability to cooperate with supplementary agents are proved through simulation experiments. The proposed solution outperforms the state-of-the-art Individual Learner architectures-based solution.</p>

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

An energy-aware and fault-tolerant deep reinforcement learning-based approach for multi-agent patrolling problems

  • 期刊论文

  • 发表日期：

    2026年6月30日

查看全文
  • Chenhao Tong,
  • Aaron Harwood,
  • Maria A. Rodriguez,
  • Richard O. Sinnott

摘要

Autonomous vehicles are suited for continuous area patrolling problems. However, finding an optimal patrolling strategy can be challenging for many reasons. Firstly, patrolling environments are often complex and can include unknown environmental factors, such as wind or landscape. Secondly, autonomous vehicles can have failures or hardware constraints, such as limited battery life. Importantly, patrolling large areas often requires multiple agents that need to coordinate their actions collectively. In this work, we consider these limitations and propose a homogeneous multi-agent architecture approach based on model-free, deep multi-agent reinforcement learning. Patrolling agents execute identical policies locally based on their local observations and shared location information. Agents can automatically recharge themselves to support continuous collective patrolling. The patrolling system can tolerate agent failures and allow adding supplementary agents to replace failed agents or increase the overall patrol performance. The overall patrol performance, the efficiency of battery recharging strategies, the overall fault tolerance, and the ability to cooperate with supplementary agents are proved through simulation experiments. The proposed solution outperforms the state-of-the-art Individual Learner architectures-based solution.