Data relationship degree–based clustering data aggregation for VANET. Issue 3 (3rd March 2016)

• Record Type: Journal Article
• Authors: Kumar, Rakesh; Dave, Mayank
• Is Part Of: International journal of electronics. Volume 103:Issue 3(2016)
• Page Start: 485
• Page End: 503
• Publication Date: 2016-03-03
• Subjects: vehicular ad hoc networks -- information dissemination -- data aggregation -- routing -- clustering; Electronics -- Periodicals; Electronic control -- Periodicals; 621.381
• DOI: 10.1080/00207217.2015.1036380
• Languages: English
• ISSNs: 0020-7217

Abstract

Abstract : Data aggregation is one of the major needs of vehicular ad hoc networks (VANETs) due to the constraints of resources. Data aggregation in VANET can reduce the data redundancy in the process of data gathering and thus conserving the bandwidth. In realistic applications, it is always important to construct an effective route strategy that optimises not only communication cost but also the aggregation cost. Data aggregation at the cluster head by individual vehicle causes flooding of the data, which results in maximum latency and bandwidth consumption. Another approach of data aggregation in VANET is sending local representative data based on spatial correlation of sampled data. In this article, we emphasise on the problem that recent spatial correlation data models of vehicles in VANET are not appropriate for measuring the correlation in a complex and composite environment. Moreover, the data represented by these models is generally inaccurate when compared to the real data. To minimise this problem, we propose a group-based data aggregation method that uses data relationship degree (DRD). In the proposed approach, DRD is a spatial relationship measurement parameter that measures the correlation between a vehicle's data and its neighbouring vehicles' data. The DRD clustering method where grouping of vehicle's data is done based on the available data and its correlation is presented in detail. Results prove that the representative data using proposed approach have a low distortion and provides an improvement in packet delivery ratio and throughput (up to of 10.84% and 24.82% respectively) as compared to the other state-of-the-art solutions like Cluster-Based Accurate Syntactic Compression of Aggregated Data in VANETs.