NLOS identification and mitigation in a real time indoor Ultra Wide Band localization system

The aim of the thesis is to support the Home Robot project with a reliable navigation system and a positioning algorithm able to locate both the robot and the patient in any harsh home environment. The main problem faced in this work is the No-Line-Of-Sight (NLOS) problem: NLOS situations occur when there is an obstacle between the tag to be localized and one or more fixed anchors. In those cases, the final localization of the tag will be wrong thanks to the interaction between the obstruction and radio signals. The overall proposed algorithm is composed by a classification algorithm to identify NLOS ranges and a subsequent mitigation algorithm to correct the final position estimation affected by NLOS positive bias. Results for both static and dynamic data collections will be discussed at the end.