Presentation Title

Using RSSI to Estimate the Number of Users in a Queue

Format of Presentation

15-minute lecture to be presented the Saturday of the conference

Location

IB 1015

Start Date

24-3-2018 10:15 AM

End Date

24-3-2018 10:30 AM

Abstract

Campus life is busy; with many students and very few places to get food, there are often queues that have unpredictable length. This results in students walking across campus only to find that to get food they would have to spend their entire lunch waiting in line. A side effect of this is added frustration in students’ lives. I set out to solve this problem: to be able to check the status of the queue without having to walk across campus.

The first step was to be able to measure the number of people standing in line. An earlier paper that attempted to correlate the signal strength to distance (Benkic et. al. 2008) proved unsuccessful, but gave me an idea to use similar technologies to estimate the number of people in an area.

The premise of the paper was that mobile devices constantly and consistently send out a broadcast to the immediate area in search of wireless networks (Probe Requests). This broadcast contains a way to identify the device as well as the signal strength (RSSI) of the broadcast. The paper found that there was a weak correlation between the signal strength of the broadcast and the distance between the mobile device and the receiver. The correlation was too weak for accurate tracking, however in my research I found that the correlation can be used as an estimate of whether the mobile device is within a certain distance. I found during my research that when listening to multiple probe requests simultaneously it was possible to configure the receiver to increase accuracy. These configurations depended on the number of mobile devices in the area, as many devices can congest the channel. The receiver also required configuration based the environment and the propagation of WIFI through the surrounding infrastructure.

I took these improvements to various places around campus and found that due to the building structure and layout I could estimate the number of people in the queue for Tim Hortons. These estimations were verified by logging the estimations and comparing the results to counts made by hand.

Department

Computing Science

Faculty Advisor

Kevin O'Neil

This document is currently not available here.

Share

COinS
 
Mar 24th, 10:15 AM Mar 24th, 10:30 AM

Using RSSI to Estimate the Number of Users in a Queue

IB 1015

Campus life is busy; with many students and very few places to get food, there are often queues that have unpredictable length. This results in students walking across campus only to find that to get food they would have to spend their entire lunch waiting in line. A side effect of this is added frustration in students’ lives. I set out to solve this problem: to be able to check the status of the queue without having to walk across campus.

The first step was to be able to measure the number of people standing in line. An earlier paper that attempted to correlate the signal strength to distance (Benkic et. al. 2008) proved unsuccessful, but gave me an idea to use similar technologies to estimate the number of people in an area.

The premise of the paper was that mobile devices constantly and consistently send out a broadcast to the immediate area in search of wireless networks (Probe Requests). This broadcast contains a way to identify the device as well as the signal strength (RSSI) of the broadcast. The paper found that there was a weak correlation between the signal strength of the broadcast and the distance between the mobile device and the receiver. The correlation was too weak for accurate tracking, however in my research I found that the correlation can be used as an estimate of whether the mobile device is within a certain distance. I found during my research that when listening to multiple probe requests simultaneously it was possible to configure the receiver to increase accuracy. These configurations depended on the number of mobile devices in the area, as many devices can congest the channel. The receiver also required configuration based the environment and the propagation of WIFI through the surrounding infrastructure.

I took these improvements to various places around campus and found that due to the building structure and layout I could estimate the number of people in the queue for Tim Hortons. These estimations were verified by logging the estimations and comparing the results to counts made by hand.