Learning automata-based solution to target coverage problem for directional sensor networks with adjustable sensing ranges
The extensive applications of directional sensor networks (DSNs) in a wide range of situations have attracted a great deal of attention. One significant problem linked with DSNs is target coverage, which primarily operate based on simultaneously observing a group of targets occurring in a set area,...
Saved in:
| Main Author: | |
|---|---|
| Format: | Thesis |
| Language: | English |
| Published: |
2017
|
| Subjects: | |
| Online Access: | http://eprints.utm.my/id/eprint/79394/1/MohdNorsyarizadRazaliPFS2017.pdf |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| Summary: | The extensive applications of directional sensor networks (DSNs) in a wide range of situations have attracted a great deal of attention. One significant problem linked with DSNs is target coverage, which primarily operate based on simultaneously observing a group of targets occurring in a set area, hence maximizing the network lifetime. As there are limitations to the directional sensors’ sensing angle and energy resource, designing new techniques for effectively managing the energy consumption of the sensors is crucial. In this study, two problems were addressed. First, a new learning automata-based algorithm is proposed to solve the target coverage problem, in cases where sensors have multiple power levels (i.e., sensors have multiple sensing ranges), by selecting a subset of sensor directions that is able to monitor all the targets. In real applications, targets may have different coverage quality requirements, which leads to the second; the priority-based target coverage problem, which has not yet been investigated in the field of study. In this problem, two newly developed algorithms based on learning automata and greedy are proposed to select a subset of sensor directions in a way that different coverage quality requirements of all the targets could be satisfied. All of the proposed algorithms were assessed for their performances via a number of experiments. In addition, the effect of each algorithm on maximizing network lifetime was also investigated via a comparative study. All algorithms are successful in solving the problems; however, the learning automata-based algorithms are proven to be superior by up to 18% comparing with the greedy-based algorithms, when considering extending the network lifetime. |
|---|