repository
A novel deep reinforcement learning based clustering scheme for WSN
| DC Field | Value | Language |
|---|---|---|
| dc.contributor.advisor | 최영준 | - |
| dc.contributor.author | YAN CHENGLONG | - |
| dc.date.accessioned | 2025-01-25T01:35:52Z | - |
| dc.date.available | 2025-01-25T01:35:52Z | - |
| dc.date.issued | 2023-08 | - |
| dc.identifier.other | 32920 | - |
| dc.identifier.uri | https://dspace.ajou.ac.kr/handle/2018.oak/24310 | - |
| dc.description | 학위논문(석사)--아주대학교 일반대학원 :인공지능학과,2023. 8 | - |
| dc.description.tableofcontents | Chapter 1 Introduction 1 <br> 1.1 Introduction to WSN 1 <br> 1.2 Research Motivation and Method Overview 2 <br>Chapter 2 Related Works 4 <br> 2.1 Traditional Methods 4 <br> 2.2 RL-based Methods 5 <br>Chapter 3 System Model 7 <br> 3.1 Network Model 7 <br> 3.2 Energy Model 8 <br> 3.2.1 Data sensing 8 <br> 3.2.2 Data logging 9 <br> 3.2.3 Data aggregation and processing 10 <br> 3.2.4 Transmission and Receiving 10 <br> 3.2.5 Network energy consumption 11 <br>Chapter 4 Proposed Method 13 <br> 4.1 Reinforcement Learning 13 <br> 4.2 RL based novel clustering method 17 <br>Chapter 5 Results 21 <br> 5.1 Experimental Environment 21 <br> 5.2 Simulation Results 21 <br>Chapter 6 Conclusion 28 | - |
| dc.language.iso | eng | - |
| dc.publisher | The Graduate School, Ajou University | - |
| dc.rights | 아주대학교 논문은 저작권에 의해 보호받습니다. | - |
| dc.title | A novel deep reinforcement learning based clustering scheme for WSN | - |
| dc.type | Thesis | - |
| dc.contributor.affiliation | 아주대학교 대학원 | - |
| dc.contributor.department | 일반대학원 인공지능학과 | - |
| dc.date.awarded | 2023-08 | - |
| dc.description.degree | Master | - |
| dc.identifier.localId | T000000032920 | - |
| dc.identifier.url | https://dcoll.ajou.ac.kr/dcollection/common/orgView/000000032920 | - |
| dc.subject.keyword | WSN | - |
| dc.subject.keyword | clustering | - |
| dc.subject.keyword | lifetime | - |
| dc.subject.keyword | reinforcement learning | - |
| dc.subject.keyword | transmission quality | - |
| dc.description.alternativeAbstract | For a long time in the past, the development of Wireless Sensor Networks (WSNs) has received a lot of attention. As a promising technology, WSN can be used in various applications including environmental monitoring, surveillance, and healthcare. The efficient utilization of limited resources in WSNs is crucial to prolong network lifetime and ensure optimal performance. Clustering is an effective approach to organize network nodes into groups, where one node acts as a cluster head to coordinate intra-cluster communication and data aggregation. <br>Traditional clustering algorithms often rely on pre-defined parameters or heuristics and proceed clustering by cluster head selection and cluster formation individually, which may not adapt well to dynamic network conditions. A Reinforcement Learning (RL) based clustering protocol is proposed in this paper that integrates cluster head selection and cluster formation as one step. It considers both energy efficiency and inter cluster interference in the model-free design, which achieves longer network lifetime and higher quality of packet transmission. <br>We conduct extensive experiments using the WSN model to verify the proposed scheme, and results is compared with Low Energy Adaptive Clustering Hierarchy (LEACH) and Greedy Energy Efficient Clustering Scheme (GEECS). The experimental results show that the proposed scheme improves the overall lifetime by 65% and 29% respectively. <br>In conclusion, this paper presents a novel deep RL based clustering scheme for WSNs, offering significant advantages over traditional approaches. By harnessing the capabilities of DRL, our proposed scheme optimizes resource utilization and extends the lifetime of WSNs. | - |
- Appears in Collections:
- Graduate School of Ajou University > Department of Artificial Intelligence > 3. Theses(Master)
- Files in This Item:
- There are no files associated with this item.
Items in DSpace are protected by copyright, with all rights reserved, unless otherwise indicated.
-
- License
- STATISTICS
- Total Visit :5,303,341
- Total Download :2,140
- Today View :3,980
