TrinaryMC: Monte Carlo Based Anchorless Relative Positioning for Indoor Positioning

Abstract

Identifying positions of mobile devices within indoor environments allows for the development of advanced applications with context and environmental awareness. Classic localization methods require GPS; an expensive, high power consuming and inaccurate solution for indoor situations. Relative positioning allows nodes to recognize their location in relation to neighboring nodes to develop an internal mapping of their own position compared to those around them. This enables a quick deployment of a given system in new, unknown indoor environments without requiring prerequisite human mapping steps. In this paper, we develop a Monte Carlo Localization (MCL) based anchorless, relative positioning algorithm which simplifies the problem to considering three states of interaction between devices: approaching, retreating and invisible. Considering three states contributes to existing MCL methods which so far only consider binary states of visible or invisible. Through our anchorless approach, we show by simulations that TrinaryMC can provide more accurate positioning information than existing anchor based methods without relying on GPS, hence decreasing hardware costs and energy consumption from the use of GPS modules as well as reducing communication overhead compared to state-of-the-art MCL methods.