Mobile location based on SVM in MIMO communication systems

Abstract

In this paper, a novel method for the positioning issue in MIMO wireless communication system is proposed, which is based on least squares support vector machine (LS-SVM). The experimental environment is established by using ray-tracing channel model, which can get three channel characteristics: delay of arrival (DOA), angle of arrival (AOA) and angle of departure (AOD) in the experimental environment. Thousands of channel characteristics make up of training set for least square support vector machine to fit the function from characteristics to coordinates. Finally, channel parameters of unknown point are used to test the function to estimate the point coordinates. The comparison results of location errors of support vector machine method, K nearest neighbor algorithm and artificial neural networks show the effectiveness and superiority of this method.