Skip to main content

LUP Student Papers

LUND UNIVERSITY LIBRARIES

TDOA-Based Multilateration Using Networked Devices for Real-Time Positioning and Visualisation

Bergendorff, Oliver LU and Bergman, Martin LU (2024) EITM01 20241
Department of Electrical and Information Technology
Abstract
Every day more devices all over the world are connected to networks. Some of these devices have microphones that could potentially be used in a networked array. This opens up the possibility to determine where sounds in the environment are produced, given the positions of the microphones are known. This thesis investigates the feasibility of using a number of network connected speakers with built-in microphones to track where a sound came from. The proposed system tries to position sounds by synchronising the clocks between devices and sending the audio streams from their microphones to a server to be processed. The server estimates the difference in arrival time of the signal between the devices. This results in a series of differences... (More)
Every day more devices all over the world are connected to networks. Some of these devices have microphones that could potentially be used in a networked array. This opens up the possibility to determine where sounds in the environment are produced, given the positions of the microphones are known. This thesis investigates the feasibility of using a number of network connected speakers with built-in microphones to track where a sound came from. The proposed system tries to position sounds by synchronising the clocks between devices and sending the audio streams from their microphones to a server to be processed. The server estimates the difference in arrival time of the signal between the devices. This results in a series of differences that can be used in multilateration to estimate the location of origin. The results are then visualised in a graph. The investigated system displays decent results when positioning sounds in a controlled environment, with differences in appropriate configurations for different types of sounds. The results show that more reliable synchronisation methods would allow for greater precision with the suggested approach. (Less)
Popular Abstract
Using a setup of networked devices with microphones we visualised the position of a speaking or snapping human to around a metre of accuracy, improving to approximately 10 centimetres when synchronised manually. By determining the difference in time of a sound arriving at several devices along with their locations we are able to estimate the position the sound originated from. These positions could be displayed on screen with a negligible latency, alongside a precision estimate.
Please use this url to cite or link to this publication:
author
Bergendorff, Oliver LU and Bergman, Martin LU
supervisor
organization
course
EITM01 20241
year
type
H2 - Master's Degree (Two Years)
subject
keywords
Positioning, Multilateration, Time Difference of Arrival, TDOA, Synchronisation, Visualisation
report number
LU/LTH-EIT 2024-976
language
English
id
9158228
date added to LUP
2024-06-17 10:39:30
date last changed
2024-06-17 10:39:30
@misc{9158228,
  abstract     = {{Every day more devices all over the world are connected to networks. Some of these devices have microphones that could potentially be used in a networked array. This opens up the possibility to determine where sounds in the environment are produced, given the positions of the microphones are known. This thesis investigates the feasibility of using a number of network connected speakers with built-in microphones to track where a sound came from. The proposed system tries to position sounds by synchronising the clocks between devices and sending the audio streams from their microphones to a server to be processed. The server estimates the difference in arrival time of the signal between the devices. This results in a series of differences that can be used in multilateration to estimate the location of origin. The results are then visualised in a graph. The investigated system displays decent results when positioning sounds in a controlled environment, with differences in appropriate configurations for different types of sounds. The results show that more reliable synchronisation methods would allow for greater precision with the suggested approach.}},
  author       = {{Bergendorff, Oliver and Bergman, Martin}},
  language     = {{eng}},
  note         = {{Student Paper}},
  title        = {{TDOA-Based Multilateration Using Networked Devices for Real-Time Positioning and Visualisation}},
  year         = {{2024}},
}