Find My Car: Simple RSS-based UWB Localization Algorithms for Single and Multiple Transmitters

Abstract

This paper analyzes a scenario where single or multiple sensors placed on a parked car help the owner to find the car in a parking lot. It is assumed that the sensors communicate with the owner device over ultra-wide-band (UWB) and the owner device only records the received signal strength (RSS) values. Based on the RSS data, which are obtained from a field test, two algorithms are proposed, first, when there is only one UWB sensor transmitting from the car, and second, when there are three such transmitters available. For the single transmitter case, we used a multi-slope method for RSS-based ranging which yielded an average improvement of 2.5% in ranging accuracy over the basic linear regression model. For the three transmitter case, we are estimating distance of the receiver from the transmitters in an individual manner and then enhance the accuracy of the distance estimates with the knowledge of the other distances. The algorithm is simple, can act with heterogeneous sensor types and do not require time synchronization among transmitters, yet achieve an 8% improvement in ranging accuracy.