Cooperative Navigation via Relational Graphs and State Abstraction

IEEE Networking Letters Pub Date : 2023-06-12 DOI:10.1109/LNET.2023.3285295

Salwa Mostafa;Mohamed K. Abdel-Aziz;Mehdi Bennis

引用次数: 0

Abstract

We consider a cooperative-navigation problem in a partially observable MADRL framework. We investigate how agents cooperate to learn a communication protocol given a very large state space while generalizing to a new environment. The proposed solution leverages the notion of structured observation and abstraction, in which the raw-pixel observations are converted into a relational graph that is then used for learning abstraction. Abstraction is performed based on compression using a relational graph autoencoder (RGAE) and a multilayer perceptron (MLP) to remove irrelevant information. The results show the effectiveness of the proposed MLP and RGAE in learning better policies with better generalization capabilities. It is also shown that communication among agents is instrumental in improving the navigation task performance.

查看原文

微信好友朋友圈 QQ好友复制链接

本刊更多论文

通过关系图和状态抽象实现协同导航

我们考虑了部分可观测 MADRL 框架中的合作导航问题。我们研究了在一个非常大的状态空间中，代理如何合作学习通信协议，同时适应新的环境。我们提出的解决方案利用了结构化观测和抽象的概念，将原始像素观测转换成关系图，然后用于学习抽象。抽象是在使用关系图自动编码器（RGAE）和多层感知器（MLP）压缩的基础上进行的，以去除无关信息。结果表明，提议的 MLP 和 RGAE 在学习具有更好泛化能力的更佳策略方面非常有效。结果还表明，代理之间的交流有助于提高导航任务的性能。

本文章由计算机程序翻译，如有差异，请以英文原文为准。

求助全文

约1分钟内获得全文去求助

来源期刊

IEEE Networking Letters

自引率

0.00%

发文量

期刊最新文献

Table of Contents IEEE Networking Letters Author Guidelines IEEE COMMUNICATIONS SOCIETY IEEE Communications Society Optimal Classifier for an ML-Assisted Resource Allocation in Wireless Communications