Session D4: Robust Navigation Using Alternative Navigation Sensors and Solutions

Collaborative RF-SLAM for Multipath Channels
Robert M. Tenny, Todd E. Humphreys, The University of Texas at Austin; Atieh Khamesi, Thomas Cheng, Stefan Adalbjörnsson, Ericsson Research
Location: Johnson (First Floor)
Date/Time: Thursday, Sep. 11, 2:58 p.m.

This paper presents a collaborative technique for radio-frequency simultaneous localization and mapping (RF-SLAM) for positioning in indoor environments and urban canyons using signals from a single base station. Multipath effects in indoor environments and urban canyons create ambiguities in both the time of arrival (TOA) and angle of arrival (AOA) measurements, causing difficulty for traditional localization methods. Yet reflections can be exploited for positioning by serving as landmarks for RF-SLAM. Multipath components caused by specular reflections are used to aid in positioning by estimating the location of a virtual transmitter (VT) transmitting a signal with the same properties as the reflected signal. The line-of-sight path to the base station and specular reflections are used to jointly estimate the user’s position, velocity, and the position of VTs, enabling robust positioning despite multipath channels. This paper makes three primary contributions. First, it develops the unscented Poisson multi-Bernoulli mixture SLAM (UPMBM-SLAM) estimator to tightly couple TOA and AOA measurements with an inertial sensor to jointly estimate the UE and VT positions and detect walls. Second, it validates the UPMBM-SLAM estimator using data simulated by ray tracing software. Third, it leverages the VT estimates from multiple UEs to create collaborative maps of the VT locations and thereby improve localization precision.