期刊
NEURAL COMPUTING & APPLICATIONS
卷 34, 期 17, 页码 14599-14607出版社
SPRINGER LONDON LTD
DOI: 10.1007/s00521-022-07244-y
关键词
Deep reinforcement learning; UAV navigation; 3D environment
资金
- national key research and development program of China [2019YFB2102200]
- Natural Science Foundation of China [61902062, 61972086, 62102082]
- Jiangsu Provincial Natural Science Foundation of China [BK20190332, BK20210203]
- Zhejiang Lab [2019NB0AB05]
This paper introduces a goal-conditioned reinforcement learning framework for vision-based UAV navigation. By developing a memory-enhanced model, the proposed approach improves the success rate of navigation and reduces the training steps.
It is a long-term challenging task to develop an intelligent agent that is able to navigate in 3D environment using only visual input in an end-to-end manner. In this paper, we introduce a goal-conditioned reinforcement learning framework for vision-based UAV navigation, and then develop a Memory Enhanced DRL agent with dynamic relative goal, extra action penalty and non-sparse reward to tackle the UAV navigation problem. This enables the agent to escape from the objective-obstacle dilemma. By performing experimental evaluations in high-fidelity visual environments simulated by Airsim, we show that our proposed memory-enhanced model can achieve higher success rate with less training steps compared to the DRL agents without memories.
作者
我是这篇论文的作者
点击您的名字以认领此论文并将其添加到您的个人资料中。
推荐
暂无数据