Evaluation of error estimation in DV-Hop for Wireless Sensor Networks

Abstract

The Wireless Sensor Networks (WSNs) consist of spatially distributed wireless sensor nodes that cooperate with each other in order to monitor and collect data pertaining to physical or environmental conditions such as temperature, pressure, motion, sound, and other phenomena. The locations of the sensor nodes are not predetermined as they are usually randomly deployed in the region of interest. Therefore, Localization algorithms that can compute the location of sensor nodes within a WSN are needed. There are mainly two types of localization algorithms. The Range-based localization algorithm has strict requirements on hardware, thus is expensive to be implemented in practice. The Range-free localization algorithm reduces the hardware cost. However, it can only achieve high accuracy in ideal scenarios. Distance Vector Hop (DV-Hop) is the mostly range-free algorithm used. In this thesis, i review and estimate the DV-Hop performance limit by changing positioning method. The most unsighted parameter is mean absolute error which has been demonstrated in the Simulation results. Simulation results show that the localization mean absolute error for both multilateration and bounding box methods are different. However, the localization accuracy in bounding box is far smaller than that of the multilateration. Therefore, multilateration method can be applied at closer applications which need more accurate positioning estimation.