Indoor Area Location System Using UWB Technology and Axis-Linear Bounding Boxes

Abstract

Due to the need of determining the localization of elderly people in indoor spaces to detect their habits and potential anomalies in their behaviour, there is a need to manage the lack of real intelligent systems capable of calculating the area localization of a person within an enclosed space. This paper proposes an architecture that uses the commercial DWM1001 system from Decawave and a Raspberry Pi, employing Ultra Wideband technology and a proposed method based on axisaligned bounding boxes to obtain the area localization where the person is. An experimentation with this system was deployed in a six-room apartment to show its effectiveness, achieving an accuracy rate exceeding 85%. Furthermore, the paper also discusses the extension of the proposed method with evolutionary algorithms. Finally, our proposal is complemented by a mobile application designed to collect and manage information from the localization sessions carried out by a person testing the system in the apartment.