Hybrid Localization: A Low Cost, Low Complexity Approach Based on Wi-Fi and Odometry

TL;DR

This paper proposes a hybrid indoor localization method combining Wi-Fi signals and odometry sensors, demonstrating improved accuracy in real-world tests for mobile robots in diverse environments.

Contribution

It introduces a novel algorithm that fuses Wi-Fi and odometry data, enhancing localization accuracy over traditional methods in indoor environments.

Findings

Reduced localization error in real-world tests

Effective fusion of Wi-Fi and odometry data

Improved accuracy for mobile robot localization

Abstract

Localization in indoor environments is essential to further support automation in a wide array of scenarios. Moreover, direction-of-arrival knowledge is essential to supporting high speed millimeter-wave (mmWave) links in indoor environments, since most mmWave links are of a line-of-sight nature to combat the high pathloss in this band. Accurate wireless localization in indoor environments, however, has proved a challenging task due to multi-path fading. Additionally, due to the effects of multi-path fading, methods such as trilateration alone do not result in accurate localization. As such, in this paper we propose to combine the knowledge of wireless localization methods with that of odometry sensors to track the location of a mobile robot. This paper presents significant real-world localization measurement results for both Wi-Fi and odometry in diverse environments at the Boise State University campus. Using these results, we devise an algorithm to combine data from both odometry and wireless localization. This algorithm is shown in hardware testing to reduce the localization error for a mobile robot