Message-Aware Graph Attention Networks for Large-Scale Multi-Robot Path Planning
Message-Aware Graph Attention Networks for Large-Scale Multi-Robot Path Planning
复制标题
用于大规模多机器人路径规划的消息感知图注意网络
DOI：
10.1109/lra.2021.3077863
复制
发表时间：
2020
影响因子：
5.2
通讯作者：
Amanda Prorok
中科院分区：
文献类型：
--
作者：
Qingbiao Li;Weizhe Lin;Zhe Liu;Amanda Prorok
运输和物流领域越来越依赖自主移动机器人来处理和分配乘客或资源。在大系统规模下，寻找分散的路径规划和协调解决方案是高效系统性能的关键。最近，图神经网络（gnn）由于能够在分散的多智能体系统中学习通信策略而变得流行。然而，香草gnn依赖于简单的消息聚合机制，防止代理对重要信息进行优先级排序。为了应对这一挑战，在这封信中，我们扩展了我们以前的工作，通过结合一种新机制来允许消息依赖的注意，在多智能体路径规划中利用gnn。我们的消息感知图注意网络（MAGAT）基于类似键查询的机制，该机制确定从各种相邻机器人接收的消息中特征的相对重要性。我们表明，MAGAT能够达到接近于耦合集中式专家算法的性能。此外，消融研究和与几个基准模型的比较表明，我们的注意机制在不同机器人密度下都非常有效，并且在不同的通信带宽约束下表现稳定。实验表明，我们的模型能够在以前看不见的问题实例中很好地进行泛化，并且即使在比训练实例×100大的非常大规模的实例中，它也比基准成功率提高了47%。
The domains of transport and logistics are increasingly relying on autonomous mobile robots for the handling and distribution of passengers or resources. At large system scales, finding decentralized path planning and coordination solutions is key to efficient system performance. Recently, Graph Neural Networks (GNNs) have become popular due to their ability to learn communication policies in decentralized multi-agent systems. Yet, vanilla GNNs rely on simplistic message aggregation mechanisms that prevent agents from prioritizing important information. To tackle this challenge, in this letter, we extend our previous work that utilizes GNNs in multi-agent path planning by incorporating a novel mechanism to allow for message-dependent attention. Our Message-Aware Graph Attention neTwork (MAGAT) is based on a key-query-like mechanism that determines the relative importance of features in the messages received from various neighboring robots. We show that MAGAT is able to achieve a performance close to that of a coupled centralized expert algorithm. Further, ablation studies and comparisons to several benchmark models show that our attention mechanism is very effective across different robot densities and performs stably in different constraints in communication bandwidth. Experiments demonstrate that our model is able to generalize well in previously unseen problem instances, and that it achieves a 47% improvement over the benchmark success rate, even in very large-scale instances that are ×100 larger than the training instances.
影响因子：
5.2
作者：
Sartoretti, Guillaume;Kerr, Justin;Choset, Howie
通讯作者：
Choset, Howie