Study Of Node Localization Technology Based On Particle Swarm Optimization For Wireless Sensor Networks

Wireless Sensor Network, which is based on node localization, has thefunctions of data collection, processing, communication and management. Nodelocalization can not only monitor the location, but keep watching on moving andtracking of the targets, and assist network management. Generally the nodelocalization algorithm in WSNs could be divided into two categories: range-freeand range-based localization. To some extent, the range-free location simplifies thesteps of localization, but generally has low accuracy unless there is higher densityof node deployment. So it is not practical. Distance ranging methods ofRange-based localization include Time of Arrival (TOA), Time Difference ofArrival (TDOA), Angle of Arrival (AOA) and Received Signal Strength Indicator(RSSI), et al.Compared with the other ranging methods, RSSI measurement can get thedistance though the wireless channel model without additional hardware. Itachieves easily and has low cost relatively and does not require timesynchronization strictly. So this paper which was based on the RSSI ranging chosethe logarithm-distance path loss model to calculate RSSI, through comparingseveral signal transmission loss models, and selected the data screening method torevise the RSSI value, then used trilateration or maximum likelihood estimation tocalculate the unknown node’s location. For better localization accuracy, thisdissertation introduced the self-correction mechanism. To minimize the locationerror, an improved particle swarm optimization was adopted by this paper tooptimize the nodes localization for WSNs. This algorithm improved the defect ofthe particle swarm optimization algorithm which may hardly converge to theoptimal solution as falling into the local extreme area, and then got the betterpositioning accuracy.Firstly Particle Swarm Optimization (PSO) algorithm was compared with theimproved PSO algorithm in the simulation, and then compared the ImprovedParticle Swarm Optimization (IPSO) localization, PSO localization and the leastsquares method. Simulation results showed that the average optimal value of IPSOalgorithm was improved about10%, and the average iteration number was decreased by about20%than PSO algorithm. The average accuracy oflocalization algorithm based on IPSO increased by almost10%to20%than based onPSO algorithm, and compared with the least squares method the accuracy wasimproved by about20%to40%, while the improved localization algorithm shownthe high stability and better convergence.