Presentation Title

Single-stage Shadow Matching Localization Using GPS Pseudorange Measurements

Presenter Information

Anthony M. Brice, CSUCIFollow

Start Date

November 2016

End Date

November 2016

Location

HUB 302-105

Type of Presentation

Poster

Abstract

In urban areas where nearby buildings either block or reflect Global Positioning System (GPS) signals, localization quality is often degraded because the receiver must rely too heavily on unblocked satellites within a small slice of the sky above the receiver. Measurements from these satellites create high uncertainty in the predicted position of the GPS receiver due to the linear relationship the unblocked satellites share with each other. A traditional GPS receiver either throws out information from blocked signals or considers its noisy signals as good as the unblocked, less noisy ones. A procedure known in the literature as Shadow Matching performs a ray tracing operation between the estimated position of the GPS receiver and the known position of each GPS satellite to assign a probability of blockage based on signal strength measured by the GPS receiver. These additional measurements can be fused with typical GPS estimates and a dynamical model of receiver motion in a particle filter framework to produce an improved localization estimate. In previous iterations of these shadow matching particle filters, some of the inputs to the particle filter were themselves outputs of a previous filter stage. A common problem with cascade filters is correlations in inputs to the later stage filter result in biased output estimates. We present a method of improving the above approach by performing all filtering operations in a single stage thus alleviating the output bias. In order to evaluate the effectiveness of this new approach we first needed to design a testbed that enabled gathering all the relevant measurements. We wrote software to allow an off-the-shelf microcontroller equipped with a GPS receiver to collect, parse, and log to a database standard GPS messages. We then designed and implemented additional software to take the collected data to prototype the particle filter presented herein.

This document is currently not available here.

Share

COinS
 
Nov 12th, 1:00 PM Nov 12th, 2:00 PM

Single-stage Shadow Matching Localization Using GPS Pseudorange Measurements

HUB 302-105

In urban areas where nearby buildings either block or reflect Global Positioning System (GPS) signals, localization quality is often degraded because the receiver must rely too heavily on unblocked satellites within a small slice of the sky above the receiver. Measurements from these satellites create high uncertainty in the predicted position of the GPS receiver due to the linear relationship the unblocked satellites share with each other. A traditional GPS receiver either throws out information from blocked signals or considers its noisy signals as good as the unblocked, less noisy ones. A procedure known in the literature as Shadow Matching performs a ray tracing operation between the estimated position of the GPS receiver and the known position of each GPS satellite to assign a probability of blockage based on signal strength measured by the GPS receiver. These additional measurements can be fused with typical GPS estimates and a dynamical model of receiver motion in a particle filter framework to produce an improved localization estimate. In previous iterations of these shadow matching particle filters, some of the inputs to the particle filter were themselves outputs of a previous filter stage. A common problem with cascade filters is correlations in inputs to the later stage filter result in biased output estimates. We present a method of improving the above approach by performing all filtering operations in a single stage thus alleviating the output bias. In order to evaluate the effectiveness of this new approach we first needed to design a testbed that enabled gathering all the relevant measurements. We wrote software to allow an off-the-shelf microcontroller equipped with a GPS receiver to collect, parse, and log to a database standard GPS messages. We then designed and implemented additional software to take the collected data to prototype the particle filter presented herein.