Citation: Li, L.; Chen, H. UAV
Enhanced Target-Barrier Coverage
Algorithm for Wireless Sensor
Networks Based on Reinforcement
Learning. Sensors 2022, 22, 6381.
https://doi.org/10.3390/s22176381
Academic Editor: Matteo Anedda
Received: 8 July 2022
Accepted: 23 August 2022
Published: 24 August 2022
Publisher’s Note: MDPI stays neutral
with regard to jurisdictional claims in
published maps and institutional affil-
iations.
Copyright: © 2022 by the authors.
Licensee MDPI, Basel, Switzerland.
This article is an open access article
distributed under the terms and
conditions of the Creative Commons
Attribution (CC BY) license (https://
creativecommons.org/licenses/by/
4.0/).
Article
UAV Enhanced Target-Barrier Coverage Algorithm for Wireless
Sensor Networks Based on Reinforcement Learning
Li Li and Hongbin Chen *
School of Information and Communication, Guilin University of Electronic Technology, Guilin 541004, China
* Correspondence: chbscut@guet.edu.cn
Abstract:
Target-barrier coverage is a newly proposed coverage problem in wireless sensor networks
(WSNs). The target-barrier is a closed barrier with a distance constraint from the target, which
can detect intrusions from outside. In some applications, detecting intrusions from outside and
monitoring the targets inside the barrier is necessary. However, due to the distance constraint, the
target-barrier fails to monitor and detect the target breaching from inside in a timely manner. In
this paper, we propose a convex hull attraction (CHA) algorithm to construct the target-barrier
and a UAV-enhanced coverage (QUEC) algorithm based on reinforcement learning to cover targets.
The CHA algorithm first divides the targets into clusters, then constructs the target-barrier for the
outermost targets of the clusters, and the redundant sensors replace the failed sensors. Finally, the
UAV’s path is planned based on QUEC. The UAV always covers the target, which is most likely to
breach. The simulation results show that, compared with the target-barrier construction algorithm
(TBC) and the virtual force algorithm (VFA), CHA can reduce the number of sensors required to
construct the target-barrier and extend the target-barrier lifetime. Compared with the traveling
salesman problem (TSP), QUEC can reduce the UAV’s coverage completion time, improve the energy
efficiency of UAV and the efficiency of detecting targets breaching from inside.
Keywords:
wireless sensor networks (WSNs); target-barrier coverage; Unmanned Aerial Vehicle
(UAV); trajectory planning; reinforcement learning
1. Introduction
In recent years, UAVs have played a crucial role in sensor networks, and UAV-aided
wireless sensor networks can significantly improve coverage [
1
]. The rise of UAV-aided
wireless sensor networks has brought new opportunities for many applications, such as
agriculture [
2
], environmental monitoring [
3
], data collection [
4
,
5
], animal detection [
6
],
etc. Generally, coverage in WSNs can be classified into target coverage, area coverage,
and barrier coverage [
7
]. Selecting different coverage types for different applications can
significantly reduce the cost of WSNs. This paper mainly studies barrier coverage which
can detect intrusions. There have been many studies on barrier coverage, which can be
classified into the open and closed barrier. The open barrier is defined as constructing
a continuous barrier that extends from one side to the opposite side. It fails in forming
in an end-to-end fashion and can only detect intrusions from one side. Conversely, the
closed barrier connects the head to the end of the barrier and forms a ring that can detect
intrusions from any direction.
It is extremely critical to timely detect abnormal situations in some applications, such
as wildlife monitoring, epidemic area monitoring, oil leak monitoring, etc. For example,
in a wildlife monitoring scenario, it is necessary to detect intrusions from the outside to
prevent humans from entering by accident or intruding illegally. At the same time, it is
essential to monitor wild animals to detect the animals leaving their habitat or being in
an unusual situation in time. The applications mentioned above must deploy a closed
barrier with a distance constraint between the barrier and targets, and targets inside the
Sensors 2022, 22, 6381. https://doi.org/10.3390/s22176381 https://www.mdpi.com/journal/sensors
