Abstract:

         We collect signal strength at different positions in wireless sensor networks covering areas.We also employ Radial Basis Function (RBF) neural network to construct a mapping model between signal strength and the coordinate of a node. We then train such a neural network with the collected signal strength value (an input vector) before use the trained neural network to localize an unknown node.Experimental results show that this model has higher accuracy, whose average localization error is less than 10%.

Key words: wireless localization, Radial Basis Function (RBF), neural network