A coalition model for efficient indexing in wireless sensor network with random mobility / Hazem Jihad Ali Badarneh

Hazem Jihad , Ali Badarneh (2021) A coalition model for efficient indexing in wireless sensor network with random mobility / Hazem Jihad Ali Badarneh. PhD thesis, Universiti Malaya.

	PDF (The Candidate's Agreement) Restricted to Repository staff only Download (224Kb)
	PDF (Thesis PhD) Download (4Mb)

Abstract

Wireless Sensor Networks grow rapidly due to the sensing and scanning capabilities, in addition to the capabilities of Wireless Sensor Networks nodes to move randomly and send periodic reports. In such environments, the periodic reports of sensors' nodes contain the transmission attributes that represent the transmission locations and time. Because of these attributes, it requires to be efficiently allocated and effectively processed. In the case of random mobile sensors, a huge number of packets are generated, which exacerbated the burden of indexing the mobile sensors' reports, because of dependability on packets' attributes. However, several studies have attempted to improve the Grid-based index-trees by dividing the network area into smaller areas called grids. The fundamental principle underlying the division is to construct the index structure by determining the mobile sensors that belong to each grid, then construct the index according to the attributes of the packet. Nevertheless, it is more practical to improve the indexing procedures of random mobile sensors, because of the following challenges. First, the high dependability on multi-attributes (location and time) of packets in random mobile sensors. Second, in a random mobile environment, it is required for registering the current grid location of each sensor and assign its destination periodically. Specifically, this thesis proposes an efficient model called Dynamic Static Coalition-based (DySta-Coalition for short) to improve the indexing procedures of random mobile sensors. The proposed model consists of Dynamic-Coalition framework, Static-Coalition algorithm, and Coalition-Based Index-Tree framework. Dynamic-Coalition framework establishes a relevance between sensors and gateways to construct dynamic blocks called dynamic-coalitions. The proposed framework aims to improve the indexing packets and mitigate the disseminated data. Static-Coalition algorithm rearranges the packets in each gateway, by source (location attribute). Accordingly, it arranges the packets, step-by-step before accumulated in the main server for indexing. Coalition-Based Index-Tree framework is an indexing process for random mobile sensors. Finally, the performance of DySta-Coalition model is comprehensively evaluated on various synthetic datasets with various parameters (number of mobile sensors, clusters, and dimensions) and metrics (packets delivery ratio, routing overhead, space cost overhead, index building time, and number of traversed nodes). First, evaluate Dynamic-Coalition framework and Static-Coalition algorithm, in terms of Average Routing Overhead and Packet Delivery Ratio. As well as, evaluate Coalition-Based Index-Tree framework in terms of the number of traversed nodes, space-cost overhead, and index building-time. The second evaluation divides into three scenarios. The first one evaluates Coalition-Based Index-Tree framework independently, without any effect from Dynamic-Coalition framework and Static-Coalition algorithm. The second and third scenarios evaluate the performance of Coalition-Based Index-Tree by examining the effects of Dynamic-Coalition framework and Static-Coalition algorithm on Coalition-Based Index-Tree framework. The evaluation results demonstrate that DySta-Coalition model can mitigate the disseminated packets, which results in minimizing dependability on the attributes of the packet. Furthermore, it is supreme according to the results of index building time and space cost overhead.

Actions (For repository staff only : Login required)