Saved in:
Bibliographic Details
Main Authors: Wei, Anning, Liang, Jintao, Lin, Kaiyuan, Li, Ziyue, Zhao, Rui
Format: Preprint
Published: 2024
Subjects:
Online Access:https://arxiv.org/abs/2410.15205
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866913556640825344
author Wei, Anning
Liang, Jintao
Lin, Kaiyuan
Li, Ziyue
Zhao, Rui
author_facet Wei, Anning
Liang, Jintao
Lin, Kaiyuan
Li, Ziyue
Zhao, Rui
contents Existing multi-agent deep reinforcement learning (MADRL) methods for multi-UAV navigation face challenges in generalization, particularly when applied to unseen complex environments. To address these limitations, we propose a Dual-Transformer Encoder-based Proximal Policy Optimization (DTPPO) method. DTPPO enhances multi-UAV collaboration through a Spatial Transformer, which models inter-agent dynamics, and a Temporal Transformer, which captures temporal dependencies to improve generalization across diverse environments. This architecture allows UAVs to navigate new, unseen environments without retraining. Extensive simulations demonstrate that DTPPO outperforms current MADRL methods in terms of transferability, obstacle avoidance, and navigation efficiency across environments with varying obstacle densities. The results confirm DTPPO's effectiveness as a robust solution for multi-UAV navigation in both known and unseen scenarios.
format Preprint
id arxiv_https___arxiv_org_abs_2410_15205
institution arXiv
publishDate 2024
record_format arxiv
spellingShingle DTPPO: Dual-Transformer Encoder-based Proximal Policy Optimization for Multi-UAV Navigation in Unseen Complex Environments
Wei, Anning
Liang, Jintao
Lin, Kaiyuan
Li, Ziyue
Zhao, Rui
Multiagent Systems
Existing multi-agent deep reinforcement learning (MADRL) methods for multi-UAV navigation face challenges in generalization, particularly when applied to unseen complex environments. To address these limitations, we propose a Dual-Transformer Encoder-based Proximal Policy Optimization (DTPPO) method. DTPPO enhances multi-UAV collaboration through a Spatial Transformer, which models inter-agent dynamics, and a Temporal Transformer, which captures temporal dependencies to improve generalization across diverse environments. This architecture allows UAVs to navigate new, unseen environments without retraining. Extensive simulations demonstrate that DTPPO outperforms current MADRL methods in terms of transferability, obstacle avoidance, and navigation efficiency across environments with varying obstacle densities. The results confirm DTPPO's effectiveness as a robust solution for multi-UAV navigation in both known and unseen scenarios.
title DTPPO: Dual-Transformer Encoder-based Proximal Policy Optimization for Multi-UAV Navigation in Unseen Complex Environments
topic Multiagent Systems
url https://arxiv.org/abs/2410.15205