Deep Learning Approach for LOS and NLOS Identification in the Indoor Environment

Abstract

Due to confined spaces and various obstacles e.g. walls, furniture, indoor environment may be considered as a harsh and disturbing in terms of the indoor radiocommunication services operation. The given paper presents FNN (Feedforward Neural Network) method for LOS (Line-Of-Sight) and NLOS (Non-Line-Of-Sight) identification which may support mitigation of such a negative influence. Described FNN architecture was evaluated based on a real indoor measurements collected with the use of the UWB (Ultra Wideband) radio modules.