Deep-Reinforcement-Learning-Based Autonomous UAV Navigation With Sparse Rewards

强化学习 计算机科学 马尔可夫决策过程 人工智能 构造(python库) 比例(比率) 过程(计算) 机器学习 领域(数学分析) 状态空间 自主代理人 国家(计算机科学) 方案(数学) 马尔可夫过程 算法 统计 操作系统 量子力学 物理 数学分析 程序设计语言 数学
作者
Chao Wang,Jian Wang,Jingjing Wang,Xudong Zhang
出处
期刊:IEEE Internet of Things Journal [Institute of Electrical and Electronics Engineers]
卷期号:7 (7): 6180-6190 被引量:126
标识
DOI:10.1109/jiot.2020.2973193
摘要

Unmanned aerial vehicles (UAVs) have the potential in delivering Internet-of-Things (IoT) services from a great height, creating an airborne domain of the IoT. In this article, we address the problem of autonomous UAV navigation in large-scale complex environments by formulating it as a Markov decision process with sparse rewards and propose an algorithm named deep reinforcement learning (RL) with nonexpert helpers (LwH). In contrast to prior RL-based methods that put huge efforts into reward shaping, we adopt the sparse reward scheme, i.e., a UAV will be rewarded if and only if it completes navigation tasks. Using the sparse reward scheme ensures that the solution is not biased toward potentially suboptimal directions. However, having no intermediate rewards hinders the agent from efficient learning since informative states are rarely encountered. To handle the challenge, we assume that a prior policy (nonexpert helper) that might be of poor performance is available to the learning agent. The prior policy plays the role of guiding the agent in exploring the state space by reshaping the behavior policy used for environmental interaction. It also assists the agent in achieving goals by setting dynamic learning objectives with increasing difficulty. To evaluate our proposed method, we construct a simulator for UAV navigation in large-scale complex environments and compare our algorithm with several baselines. Experimental results demonstrate that LwH significantly outperforms the state-of-the-art algorithms handling sparse rewards and yields impressive navigation policies comparable to those learned in the environment with dense rewards.
最长约 10秒,即可获得该文献文件

科研通智能强力驱动
Strongly Powered by AbleSci AI
更新
PDF的下载单位、IP信息已删除 (2025-6-4)

科研通是完全免费的文献互助平台,具备全网最快的应助速度,最高的求助完成率。 对每一个文献求助,科研通都将尽心尽力,给求助人一个满意的交代。
实时播报
今后应助连冷安采纳,获得10
刚刚
1秒前
领导范儿应助Haha采纳,获得10
1秒前
1秒前
天天快乐应助调皮傲旋采纳,获得10
3秒前
S8发布了新的文献求助10
3秒前
黄金城发布了新的文献求助10
4秒前
浮游应助Jimmy Ko采纳,获得10
5秒前
5秒前
量子星尘发布了新的文献求助10
5秒前
5秒前
6秒前
白熊发布了新的文献求助30
6秒前
7秒前
7秒前
9秒前
zhu发布了新的文献求助10
10秒前
10秒前
独特听芹完成签到,获得积分10
12秒前
zz完成签到,获得积分10
12秒前
13秒前
晚风完成签到,获得积分10
14秒前
jwj发布了新的文献求助10
15秒前
15秒前
白熊完成签到 ,获得积分10
15秒前
16秒前
李健应助北齐冲浪的鱼采纳,获得10
17秒前
17秒前
王一鸣发布了新的文献求助10
18秒前
ikutovaya完成签到,获得积分10
18秒前
18秒前
奋斗的妙松完成签到,获得积分10
19秒前
老实莫言完成签到,获得积分10
19秒前
20秒前
量子星尘发布了新的文献求助150
20秒前
wop111应助morph采纳,获得20
20秒前
追寻的冬寒完成签到 ,获得积分10
21秒前
22秒前
吼吼吼吼发布了新的文献求助10
22秒前
善学以致用应助生动念烟采纳,获得10
22秒前
高分求助中
(应助此贴封号)【重要!!请各用户(尤其是新用户)详细阅读】【科研通的精品贴汇总】 10000
NMR in Plants and Soils: New Developments in Time-domain NMR and Imaging 600
Electrochemistry: Volume 17 600
Physical Chemistry: How Chemistry Works 500
SOLUTIONS Adhesive restoration techniques restorative and integrated surgical procedures 500
Energy-Size Reduction Relationships In Comminution 500
Principles Of Comminution, I-Size Distribution And Surface Calculations 500
热门求助领域 (近24小时)
化学 医学 生物 材料科学 工程类 有机化学 内科学 生物化学 物理 计算机科学 纳米技术 遗传学 基因 复合材料 化学工程 物理化学 病理 催化作用 免疫学 量子力学
热门帖子
关注 科研通微信公众号,转发送积分 4950711
求助须知:如何正确求助?哪些是违规求助? 4213460
关于积分的说明 13104286
捐赠科研通 3995337
什么是DOI,文献DOI怎么找? 2186837
邀请新用户注册赠送积分活动 1202090
关于科研通互助平台的介绍 1115359