NLOS Classification based on RSS and Ranging Statistics Obtained from Low-Cost UWB Devices

Ultra Wide Band (UWB) devices have been largely considered for indoor location systems due to their high accu- racy. However, as in other wireless systems, such accuracy is significantly degraded under Non-Line-of-Sight (NLOS) propaga- tion conditions. Therefore, distinguishing between Line-of-Sight (LOS) and NLOS becomes essential to mitigate inaccuracies due to NLOS propagation.Most of these classification processes are based on the study of the Channel Impulse Response (CIR), something that can be hard to perform depending on some aspects like the computational cost, the time of process or the availability of data from the CIR. In this research, we present an alternative for classification based on another set of parameters easier to obtain and work with, in particular the Radio Signal Strength (RSS) and the range estimation given by low-cost UWB devices. We analyze the effect of using different statistics of these parameters as features to feed a Support Vector Machine (SVM) Classifier. To test the classifier, we use measurements taken in a real scenariowith real low-cost hardware and in both LOS and NLOS situations.