rWiFiSLAM: Effective WiFi Ranging based SLAM System in Ambient Environments

TL;DR

This paper introduces rWiFiSLAM, a WiFi-based SLAM system for indoor localization that achieves sub-meter accuracy without needing access point location knowledge, significantly outperforming IMU-only methods.

Contribution

The paper presents a novel WiFi SLAM approach that removes the need for access point location knowledge, enhancing indoor localization in dynamic environments.

Findings

Achieves sub-meter localization accuracy.

Improves localization performance by over 90% compared to IMU-based PDR.

Effective loop closure mechanism using WiFi ranging measurements.

Abstract

In this paper, we propose rWiFiSLAM, an indoor localisation system based on WiFi ranging measurements. Indoor localisation techniques play an important role in mobile robots when they cannot access good quality GPS signals in indoor environments. Indoor localisation also has many other applications, such as rescue, smart buildings, etc. Inertial Measurement Units (IMU) have been used for Pedestrian Dead Reckoning (PDR) to provide localisation services in the indoor environment as it does not rely on any other signals. Although PDR is a promising technique, it still suffers from unavoidable noise and bias from IMUs in mobile devices. Loop closure is necessary for these scenarios. In this paper, we design an efficient loop closure mechanism based on WiFi ranging measurements along with IMU measurements in a robust pose graph SLAM framework for indoor localisation. One novelty of the proposed method is that we remove the requirement of the full knowledge of the WiFi access point locations, which makes our proposed method feasible for new and/or dynamic environments. We evaluate our designed system in real environments and show the proposed method can achieve sub-meter localisation accuracy and improve the localisation performance by more than 90\% compared with the IMU based PDR.