NLOS identification for UWB localization based on import vector machine

Abstract Ultra-Wide Band (UWB) based localization is one of the most promising techniques for high accuracy localization. The crucial factor that aggravates the localization precision is None-Line-of-Sight (NLOS) propagation. To address this issue, we propose a novel NLOS identification algorithm with feature selection strategy and a localization algorithm based on Import Vector Machine (IVM) with high accuracy and low complexity. The feature selection strategy further meliorates the classification accuracy. The probability outputs of IVM is employed by the localization algorithm and yields higher positioning accuracy than its counterpart methods – Support Vector Machine (SVM) and Relevance Vector Machine (RVM). Simulation results prove that IVM is a robust and efficient method for NLOS identification and localization.