An Energy Efficient Load Balancing Tree-Based Data Aggregation Scheme for Grid-Based Wireless Sensor Networks
Abstract
:1. Introduction
2. Related Work
3. The Proposed Scheme
3.1. Grid Construction
3.2. Tree Structure Construction
- (a)
- The node (cell head) does not exist in the tree and has the minimum energy consumption.
- (b)
- The current number of child nodes (cell heads) connected to the node (cell head) i is Cmi, where Cmi ≤ Cm.
- (c)
- The current depth of the node (cell head) i is Lmi, where Lmi ≤ Lm.
Algorithm 1: The tree-like path establishment algorithm of LB-TBDAS. |
Step 1: System initialization |
(1) Sensor nodes are randomly deployed in the specific network area. |
(2) The network area is partitioned into M × N cells of a grid. |
(3) The sensor node with the highest residual energy is elected to be the cell head in each cell. |
Step 2: Tree initialization |
(1) The BS is responsible for serving as the root node. |
(2) The network depth is Lm and the maximum number of child nodes (cell heads) is Cm. |
Step 3: Tree construction |
3.3. Data Transmission
4. Simulation Results
4.1. Number of Rounds Versus Node Death Percentages
4.2. Number of Rounds when 50% of Nodes Die versus Number of Nodes
4.3. Number of Rounds Versus Depth of Network
4.4. Total Consumed Energy versus Number of Rounds
4.5. Energy Distribution for Sensor Nodes
5. Conclusions
Author Contributions
Funding
Institutional Review Board Statement
Informed Consent Statement
Data Availability Statement
Conflicts of Interest
References
- Ramson, S.R.J.; Moni, D.J. Applications of Wireless Sensor Networks—A Survey. In Proceedings of the International Conference on Innovations in Electrical, Electronics, Instrumentation and Media Technology, Coimbatore, India, 3–4 February 2017; pp. 325–329. [Google Scholar]
- Sharma, S.; Kaur, A. Survey on Wireless Sensor Network, Its Applications and Issues. J. Phys. Conf. Ser. 2021, 1969, 1–10. [Google Scholar] [CrossRef]
- Wang, X.; Chen, H. A Survey of Compressive Data Gathering in WSNs for IoTs. Wirel. Commun. Mob. Comput. 2022, 2022, 1–14. [Google Scholar] [CrossRef]
- Hilmani, A.; Maizate, A.; Hassouni, L. Automated Real-Time Intelligent Traffic Control System for Smart Cities using Wireless Sensor Networks. Wirel. Commun. Mob. Comput. 2020, 2020, 1–28. [Google Scholar] [CrossRef]
- Durai, S.K.S.; Duraisamy, B.; Thirukrishna, J.T. Certain Investigation on Healthcare Monitoring for Enhancing Data Transmission in WSN. Int. J. Wirel. Inf. Netw. 2021, 1–8. [Google Scholar]
- Majid, M.; Habib, S.; Javed, A.R.; Rizwan, M.; Srivastava, G.; Gadekallu, T.R.; Lin, C.-W. Applications of Wireless Sensor Networks and Internet of Things Frameworks in the Industry Revolution 4.0: A Systematic Literature Review. Sensors 2022, 22, 2087. [Google Scholar] [CrossRef] [PubMed]
- Braginsky, D.; Estrin, D. Rumor Routing Algorithm for Sensor Network. In Proceedings of the 1st ACM International Workshop on Wireless Sensor Networks and Applications, Atlanta, GA, USA, 28 September 2002; pp. 22–31. [Google Scholar]
- Lindsey, S.; Raghavendra, C.S. PEGASIS: Power-Efficient Gathering in Sensor Information System. In Proceedings of the IEEE Aerospace Conference, Big Sky, MT, USA, 9–16 March 2002; Volume 3, pp. 1125–1130. [Google Scholar]
- Jung, S.M.; Han, Y.J.; Chung, T.M. The Concentric Clustering Scheme for Efficient Energy Consumption in the PEGASIS. In Proceedings of the 9th International Conference on Advanced Communication Technology, Phoenix Park, Korea, 12–14 February 2007; Volume 1, pp. 260–265. [Google Scholar]
- Ye, F.; Haiyun, L.; Jerry, C.; Songwu, L.; Zhang, L. A Two-Tier Data Dissemination Model for Large-Scale Wireless Sensor Networks. In Proceedings of the ACM International Conference on Mobile Computing and Networking, London, UK, 21–25 September 2002; pp. 148–159. [Google Scholar]
- Nagesh, R.; Raga, S.; Mishra, S. Elimination of Redundant Data to Enhance Wireless Sensor Network Performance Using Multi Level Data Aggregation Technique. In Proceedings of the 2019 International Conference on Computing, Communication and Networking Technologies, Kanpur, India, 6–8 July 2019; pp. 1–5. [Google Scholar]
- Deepakraj, D.; Raja, K. Hybrid Data Aggregation Algorithm for Energy Efficient Wireless Sensor Networks. In Proceedings of the 2021 International Conference on Intelligent Communication Technologies and Virtual Mobile Networks, Tirunelveli, India, 4–6 February 2021; pp. 7–12. [Google Scholar]
- Lee, W.M.; Wong, V.W.S. E-Span and LPT for Data Aggregation in Wireless Sensor Networks. In Proceedings of the Computer Communications, Arlington, VA, USA, 9–11 October 2006; Volume 29, pp. 2506–2520. [Google Scholar]
- Bandral, M.S.; Jain, S. Energy Efficient Protocol for Wireless Sensor Network. In Proceedings of the Recent Advances and Innovations in Engineering, Jaipur, India, 9–11 May 2014; pp. 477–482. [Google Scholar]
- Zhou, L.; Ge, C.; Hu, S.; Su, C. Energy-Efficient and Privacy-Preserving Data Aggregation Algorithm for Wireless Sensor Networks. IEEE Internet Things J. 2020, 7, 3948–3957. [Google Scholar] [CrossRef]
- Tan, H.O.; Korpeoglu, I. Power Efficient Data Gathering and Aggregation in Wireless Sensor Networks. ACM SIGMOD Rec. 2003, 32, 66–71. [Google Scholar] [CrossRef]
- Wang, N.-C.; Chen, Y.-L.; Huang, Y.-F.; Chen, C.-M.; Lin, W.-C.; Lee, C.-Y. An Energy Aware Grid-Based Clustering Power Efficient Data Aggregation Protocol for Wireless Sensor Networks. Appl. Sci. 2022, 12, 1. [Google Scholar] [CrossRef]
- Albowitz, J.; Chen, A.; Shang, L. Recursive Position Estimation in Sensor Networks. In Proceedings of the International Conference on Network Protocols, Riverside, CA, USA, 11–14 November 2001; pp. 35–41. [Google Scholar]
- Kaplan, E.D. Understanding GPS: Principles and Applications; Artech Hourse: Boston, MA, USA, 1996. [Google Scholar]
- Pan, M.-S.; Tseng, Y.-C. The Orphan Problem in ZigBee-Based Wireless Sensor Networks. In Proceedings of the Tenth ACM Symposium on Modeling, Analysis, and Simulation of Wireless and Mobile Systems, Crete Island, Greece, 22–26 October 2007; pp. 95–98. [Google Scholar]
- ZigBee Alliance. ZigBee Specification, Version 1.0; ZigBee Alliance: Davis, CA, USA, 2004. [Google Scholar]
- ZigBee Alliance. ZigBee—2007 Specification; ZigBee Alliance: Davis, CA, USA, 2007. [Google Scholar]
- Heinzelman, W.R.; Chandrakasan, A.; Balakrishnan, H. Energy-Efficient Communication Protocol for Wireless Microsensor Networks. In Proceedings of the IEEE Annual Hawaii International Conference on Systems Sciences, Maui, HI, USA, 6 January 2000; pp. 3005–3014. [Google Scholar]
- Andreou, P.; Pamboris, A.; Zeinalipour-Yazti, D.; Chrysanthis, P.K.; Samaras, G. ETC: Energy-Driven Tree Construction in Wireless Sensor Networks. In Proceedings of the International Conference on Mobile Data Management: Systems, Services and Middleware, Taipei, Taiwan, 18–20 May 2009; pp. 35–40. [Google Scholar]
Parameters | Values |
---|---|
Network area | 100 m × 100 m |
Location of BS | (50, 150) |
Initial energy | 0.25 J/node |
Number of cells | 10 × 10 |
Number of sensor nodes | 100–400 |
Packet size | 512 bits |
Network depth (Lm) | 5–10 |
Maximum number of child nodes of the node (Cm) | 4, 7 |
Protocol | LB-TBDAS | GB-PEDAP | PEDAP |
---|---|---|---|
Hierarchical architecture | two layers | two layers | single layer |
Data transmission structure | direct and tree | direct and tree | tree |
Type of energy consumption | load balancing | uniform | general |
Energy efficient | very high | high | low |
Publisher’s Note: MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations. |
© 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/).
Share and Cite
Wang, N.-C.; Lee, C.-Y.; Chen, Y.-L.; Chen, C.-M.; Chen, Z.-Z. An Energy Efficient Load Balancing Tree-Based Data Aggregation Scheme for Grid-Based Wireless Sensor Networks. Sensors 2022, 22, 9303. https://doi.org/10.3390/s22239303
Wang N-C, Lee C-Y, Chen Y-L, Chen C-M, Chen Z-Z. An Energy Efficient Load Balancing Tree-Based Data Aggregation Scheme for Grid-Based Wireless Sensor Networks. Sensors. 2022; 22(23):9303. https://doi.org/10.3390/s22239303
Chicago/Turabian StyleWang, Neng-Chung, Chao-Yang Lee, Young-Long Chen, Ching-Mu Chen, and Zi-Zhen Chen. 2022. "An Energy Efficient Load Balancing Tree-Based Data Aggregation Scheme for Grid-Based Wireless Sensor Networks" Sensors 22, no. 23: 9303. https://doi.org/10.3390/s22239303
APA StyleWang, N. -C., Lee, C. -Y., Chen, Y. -L., Chen, C. -M., & Chen, Z. -Z. (2022). An Energy Efficient Load Balancing Tree-Based Data Aggregation Scheme for Grid-Based Wireless Sensor Networks. Sensors, 22(23), 9303. https://doi.org/10.3390/s22239303