CBL - Campus del Baix Llobregat

Projecte llegit

Títol: Implementation and Analysis of the Passive Wi-Fi RTT Positioning in the NS-3 Simulator

Estudiants que han llegit aquest projecte:

Director/a: ZOLA, ENRICA VALERIA

Departament: ENTEL

Títol: Implementation and Analysis of the Passive Wi-Fi RTT Positioning in the NS-3 Simulator

Data inici oferta: 13-10-2025     Data finalització oferta: 13-05-2026


Estudis d'assignació del projecte:
    MU AI4CI
    MU EM CODAS 1
    MU EM CODAS 2
    MU MASTEAM 2015
Tipus: Individual
 
Lloc de realització: EETAC
 
Segon director/a (UPC): MARTIN ESCALONA, ISRAEL
 
Paraules clau:
IEEE 802.11, TDOA, indoor positioning, Wi-Fi RTT, network simulation, ns-3, python
 
Descripció del contingut i pla d'activitats:
A passive solution for indoor positioning of Wi-Fi devices has been recently proposed in the literature [1], which merges a time-difference of arrival (TDOA) algorithm with the novel fine time measurements (FTM) introduced in IEEE 802.11mc (aka Wi-Fi RTT, Round-Trip-Time). The approach leverages the scalability limitation of the original Wi-Fi RTT procedure. By exploiting other IEEE 802.11mc devices performing their own location with neighbour APs, passive Wi-Fi RTT is able to locate all devices on a network at once without injecting additional location traffic. This same idea has been also adopted in Next Generation Positioning (NGP), the new IEEE 802.11az released in 2023.
A proof of concept of the Wi-Fi passive TDOA algorithm was simulated in Matlab [1], and the impact of the error on the ranging estimations was assessed. However, the recent release of the FTM-ns3 module, an extension for the widely used ns3 network simulator to support the Wi-Fi RTT protocol, paves the way for a deeper analysis of the Wi-Fi RTT passive approach.
The idea of this master thesis is to implement the passive Wi-Fi RTT approach in ns3. The goal is to analyse the impact of several sources of error on the precision of the final position estimate of the passive user in more realistic scenarios (e.g., line of sight, multipath, etc.).

[1] Martin-Escalona, I.; Zola, E. Passive Round-Trip-Time Positioning in Dense IEEE 802.11 Networks. Electronics 2020, 9, 1193. https://doi.org/10.3390/electronics9081193


Objectives:
1) Update the FTM-ns3 extension to the last ns-3 version, if required
2) Understand the error model implemented in the FTM-ns3 module for Wi-Fi RTT
3) Derive an error model for Wi-Fi Passive TDOA
4) Implement the error model proposed
5) Analyse the impact of the error on the position estimate
6) Write a research paper to present the results in an international conference
 
Overview (resum en anglès):
Wi-Fi Fine Timing Measurement, introduced by the IEEE 802.11mc amendment, allows a station to estimate its distance from an access point through the round trip time of dedicated management frames. The procedure is active. Each station that wants to be positioned must generate its own ranging traffic, which limits scalability in dense deployments. Passive Wi-Fi RTT avoids this cost: a device overhears FTM exchanges between other stations and derives a timing metric without transmitting any frame.
This thesis implements passive Wi-Fi RTT observation inside the ns-3 network simulator, and uses it to estimate the distance of the passive device from another device. The reference FTM framework available for ns-3 supports active ranging only, so the passive observer is designed and built from scratch. A Passive FTM Listener is integrated into the ns-3 Wi-Fi module. It acquires frames directly from the physical layer, associates each FTM Response with its acknowledgement, reconstructs the four timestamps of the exchange, and computes the passive metric. The ranging error models of the reference FTM framework have been reused without modification.
The implementation is validated over 104,029,800 independent scenarios. Without any error model, the passive metric is reconstructed to numerical precision. With the error models enabled, each one reproduces its expected statistics. Once the passive Wi-Fi RTT ranging estimation is validated, the achievable position accuracy from several passive observations is analysed. The position of the observer is estimated using a nonlinear least squares solver. With well-spread anchors the mean position error stays below 0.61 m. With collinear or clustered anchors the estimation degrades sharply, showing that anchor geometry, rather than measurement accuracy, is the dominant limitation of this approach.

© CBLTIC Campus del Baix Llobregat - UPC