Research On Wireless Sensor Network Localization Algorithms Based On DV-HOP

Wireless sensor networks (WSN) has attracted greatly interests in recent years and also has been considered as one of most popular research aspects. Node localization technology, which is the key technology for application of WSN, can be classified into two categories, that are range-based localization algorithm and range-free localization algorithm. The former algorithm, the notes can be localized with higher accuracy. However, the cost and energy consumption is high, and thus it does not satisfy the practical applications. On the contrary, the later algorithm could achieve the localization with lower cost even though the accuracy is lower than former. Furthermore, the localization accuracy is enough for many practical applications. DV-Hop as the classic algorithm of range-free algorithm is the most widely adopted algorithm, and meanwhile has attracted greatly interests. At present, some research achievements have been made, but it is still to be further improved. In this thesis, the localization principle of DV-Hop and the reason of making error are investigated in detail. Most importantly, we reduce the error of DV-Hop by using following methods:(1) In conventional DV-Hop localization algorithms, the relative location between two notes is represented by one-hope simply. However, this relative location is inaccurate beceuse the range of one-hope cannot be determined. To overcome this disadvantage, the quantitative-RSSI-amendment-DV-Hop (QRA-DV-Hop) algorithm has been proposed based on a quantitative model, which is established by using RSSI (Received Signal Strength Indicator). The improved algorithm could calculate the distance between two notes by only using the intensity of RSSI instead of complex distance calculation. In addition, by employing the proposed algorithm, the neighbour notes of direct communication are quantified, and the number of quantization unit is corrected. Therefore, the distance of one-hope is reduced and the accuracy of localization is improved.(2) In DV-Hop localization algorithm, the coordinates are estimated by using least square method (LSM). However, LSM is sensitive to the error of one anchor node. It indicates that the localization error would become very large if only one anchor node error is large. Actually, the coordinate localization can be considered as minimum value calculation of error function. Therefore, to overcome the disadvantage of DV-Hop, an improved algorithm has been proposed based on SA-CLS-PSO, which combines with simulated annealing (SAX chaos local search (CLS) and particle swarm optimization(PSO) with compression factor. In proposed algorithm, the localization is achieved by founding the minimum value of error function instead of using LSM algorithm. Simulation results show that such algorithm could improve the localization accuracy effectively and mitigate the superposition error.