RSSI-based localization in wireless sensor networks using Regression Tree

Wireless sensor networks are different from other networks; therefore it is necessary to use innovative techniques to solve some issues. Localization is a significant area of research in wireless sensor networks due to its various applications. This paper proposes and evaluates a Received Signal Strength-based localization algorithm using Regression Tree by comparing its performance with Least Squares Support Vector Regression and Multi Layers Perceptron Neural Network. The evaluation considers the localization error and the complexity of the algorithm. Simulations show that Regression Tree method is simple and efficient, even when using a small number of anchor nodes.