My research
My work is on wireless sensing and localization. It has four parts.
What I work on
Wireless signals are sent to carry messages. They can also tell us where something is, how it is moving, and whether something unusual has happened. I work on the limits of what can be measured and on the engineering that makes a measurement useful outside the lab.
Where is it, and how sure are we?
GPS does not reach indoors. A drone, a headset, or a healthcare robot there needs its own positioning system: it measures distances to beacons, Wi-Fi routers, or 5G and 6G small cells, and turns them into a position. For indoor drones I worked on this in RoLaTIN, RAIL, and PILOT.
A precise distance measurement is not enough, because the layout of the beacons or cells can dominate the error. Much of my work uses optimization and learning to decide where to put them: sensors and beacons (SPIN, iDROP, and OPTILOD), Wi-Fi routers in 5G and 6G networks (Wi-Five and Wi-Six), and small cells (5GPS and OFDRA). I also worked with reconfigurable intelligent surfaces (RIS), panels whose reflection of radio waves can be programmed, and with retroreflective tags. Other papers are about positioning for headsets, healthcare robots, and vehicles. Wi-Six and TACTIC each received a best paper award.
Selected papers
Can satellite signals detect aircraft?
A radar normally sends out a signal and listens for the echo. A passive radar sends nothing. It listens to signals that are already in the air and watches how a target changes them.
Satellites in low Earth orbit are attractive transmitters for this: there are thousands of them, and they transmit around the clock. But their signals were not designed for radar. The geometry changes quickly, the satellite and the receiver share no clock, the waveform is only partly known, and what a target adds is weak next to the direct signal. The papers below look at how densely the receivers must be spread and where to place them (CARTA and STARMAP).
This became the main work of WayWave. The WayWave chapter →
Selected papers
How do systems stay right when the measurements are wrong?
Measurements go wrong. There is noise, signals bounce off walls, data goes missing, and sometimes somebody feeds the system a false signal on purpose.
I am interested in systems that say how sure they are, notice when measurements contradict each other, and adapt when conditions change. This began at Lehigh with spoofed GPS signals in the power grid, continued at Virginia Tech with 5G as a second check on GPS, and now includes reinforcement learning for tracking and placement.
Spoofed GPS in the power grid
Phasor measurement units report the state of a power grid and take their clock from GPS. I worked on detecting a GPS spoofing attack from their data. My master's thesis is a reliability test of the detection scheme.
Spoofed GPS, checked against 5G
The positioning that 5G offers can expose a spoofed GPS position, with only a single base station.
Detecting trespassing drones
At Virginia Tech I started with the security of drones and ways to detect them. A later journal article is about drones that trespass a protected area.
Geofencing systems
A geofence is a boundary in software around a real place. EGO-6 asks where to deploy 6G transmission and reception points (TRPs) to make one stronger. Two later papers are about precise tracking in such systems.
Location-based authentication
One paper uses deep Q-learning, a form of reinforcement learning, for authentication based on location.
Reinforcement learning for tracking and placement
Deep and meta-reinforcement learning for tracking in geofencing systems and for placing sensors, routers, cells, and receivers (GREEN, RAPID, RADIO, and STARMAP).
Selected papers
How finely can we measure time?
A range measurement is a time measurement. A radio wave covers one millimeter in about three trillionths of a second, so positioning to a fraction of a millimeter needs timing at least that fine.
An oscillator that is pushed off its steady rhythm returns to it early or late. The curves that sort its states by that timing are called isochrons. With G. Himona, Y. Kominis, and V. Kovanis, I studied the isochrons of photonic oscillators and of nonlinear electronic oscillators, and how they can turn an oscillator into a time sensor for ranging and positioning. The aim is localization and tracking below a millimeter, for augmented and virtual reality.
Modeldr/dt = r (1 − r²)dθ/dt = ω + b (1 − r²)phase Φ = θ − b ln rkick r → k r: every Φ shifts by −b ln k
Selected papers
Networks
With T. O. Atalay as first author, I have coauthored measurement studies of 5G core networks that run in the cloud. One of them, 5G-MAP, received the Best Community Paper Award at the ACM International Conference on Mobile Computing and Networking (MobiCom) in 2025. Other papers cover service meshes for the 5G core and relay networks carried by drones.