Our research focuses on the study of the magnetic field recorded in urban environement. We are particularly interested in applying novel technology to identify individual magnetic signatures seemingly undistinguishable from environmental noise.
The technique consists of using an array of magnetometer sensors and perform coherent search of magnetic transient events. Sensor-network technology has proved to be very efficient in identifying signal that are not detectable by a single instrument. A very famous example of using such technology is the discovery of gravitational waves by the LIGO collaboration where they used a network of laser interferometers.
It depends how many stations are in the network and how far away they are from each other. The dominant source of magnetic noise in a city usually comes from the public transportation system. However, once you understand that pattern and carefully subtract it from the data, you can start identifying underlying coherent signals.
Until now, the data were analyzed manually mostly using wavelet transform analysis and cross-correlation algorithms. We are currently developing a machine learning framework to automate the data analysis process and enable predictive analysis on the streaming data.
Our original sensor network was deployed in the city of Berkeley, California. We applied cross-correlation algorithms to identify and monitor daily fluctuations from the local subway system and developed a time-series feature extraction technique to subtract this dominant source of magnetic noise from our data. We have also started to built a library of magnetic signatures that will be used to train our machine learning algorithm currently under development.
"Network of sensitive magnetometers for urban studies", T. Bowen et al.
We are proud to mention that our network has found specific applications for various research projects within the UC Berkeley community. Specifically, our network was used by the Berkeley geophysics group to show that their seismometers were, in fact, picking up magnetic signals from the local subway system, thus being subject to a significant source of noise and systematics. In a separate occasion, we applied our "urban" magnetometry network to help researchers at the UC Berkeley Mechanical Engineering Department to characterize a novel rotating-electret antenna for potential use for communicating with workers trapped in underground mines.
As of October 2017, our network has been deployed in downtown Brooklyn in the Center of Urban Science & Progress (CUSP). So far, our preliminary results show distinct variations in the magnetic field compared to the data taken in Berkeley. Most of the signatures we see in our data comes from human activities which are yet poorly understood. Using our network, we aim to measure the urban dynamics due to human activity and correlate any identified signatures to other aspects of urban life, e.g. energy consumption, pollution.
Research Associate
@ UC Berkeley
Physics Professor
@ Johannes Gutenberg-Universität
@ Helmholtz Institut Mainz
@ UC Berkeley
@ Berkeley Lab
Physics Professor
@ UC Berkeley
Physics Professor
@ UC Berkeley
@ Space Science Laboratory
PhD Student
@ UC Berkeley
@ Space Science Laboratory
PhD Student
@ UC Berkeley
@ Space Science Laboratory
Research Associate
@ Johannes Gutenberg-Universität
@ Helmholtz Institut Mainz
Research Associate
@ Nortwestern University
@ CIERA
Physics Professor
@ New York University
@ CUSP
Associate Research Scientist
@ New York University
@ CUSP
Research Project Manager
@ New York University
@ CUSP
Research Associate
@ University of Wisconsin-Madison
Feel free to contact us if you are either interested to join our group or simply want to know more about our research.
The University of California at Berkeley
U.S. Department of Defense
New York University
Johannes Gutenberg-Universität
