Evolutionary approaches for minimising error in localisation of wireless sensor networks

TLDR: Simulation results show that the evolutionary algorithms greatly improve the localisation accuracy when it is used over the traditional mathematical approach, and GSA is the most efficient in bringing down the localised error.

Abstract: Localisation in wireless sensor networks WSNs is one of the important fundamental requisite that needs to be resolved efficiently for the deployment of sensor nodes and its operation. Localisation is a challenging issue in applications such as routing and target tracking which is all location dependent. Hence, this work aims at determining the location of the sensor nodes with high precision. This work is initially based on localisation using Mobile Anchors, a range-free localisation method used for localising the nodes. When the anchors move through the network, they broadcast their location as beacon packets. The sensor nodes after collecting enough beacon packets from mobile anchors and location packets from neighbouring nodes are able to calculate their location. To improve the localisation accuracy, evolutionary algorithms such as genetic algorithm GA, particle swarm optimisation PSO and genetic simulated annealing GSA have been used. Detailed study of localisation accuracy, root mean square error RMSE and comparison among the evolutionary approaches has been done. Simulation results show that the evolutionary algorithms greatly improve the localisation accuracy when it is used over the traditional mathematical approach. The results also show that among the evolutionary algorithms, GSA is the most efficient in bringing down the localisation error.