Radio Propagation and Channel Characterization

Testing a wireless system in the field is slow and expensive; a channel emulator reproduces the radio channel in the laboratory instead — but only as faithfully as its underlying channel model allows. We develop a grid-based channel modeling technique in which multipath parameters are precomputed deterministically — by ray tracing or extracted from measurements — at the nodes of a spatial grid, and the channel at any receiver position along any trajectory is then synthesized by interpolation. This turns site-specific, ray-traced fidelity, normally far too heavy for real time, into something an emulator can play back live for a moving receiver. The framework has been validated from indoor offices to urban field measurements, and its grid database further points toward a practical way to share site-specific channel models without disclosing the underlying 3D environment.

Millimeter-wave and terahertz bands offer enormous capacity for beyond-5G networks, but their signals barely penetrate walls, leaving coverage gaps that are costly to fill with active equipment. We develop printable, flat passive reflectors that redirect an incident beam to illuminate a designated region — no power, no electronics, just engineered geometry. The designs are derived deterministically from ray-based geometrical optics: tracing how each ray must be redirected to cover the target region, then realizing that behavior on a flat, printable surface. Where a conventional flat reflector yields only a single hot spot, our reflectors spread the energy uniformly across the target area — an improvement demonstrated with prototypes at 24 GHz and 300 GHz.

Aircraft on the airport surface are located by multilateration systems that measure the arrival times of their broadcast signals — and strong multipath from terminals and hangars is a principal source of localization error. Knowing the multipath channel reveals how reflections distort the measured arrival times — the knowledge needed to correct the resulting position errors and to place receivers where distortion is least. Because active channel sounding is prohibited inside an operational airport, we estimate the channel passively, using the ADS-B surveillance signals that aircraft already transmit. From the passive measurements conducted at the Airport, we resolve the multipath structure of the airport surface with super-resolution accuracy — and, since each ADS-B message reports the aircraft's own position, it can trace strong reflections back to the structures that cause them by comparing with the ray tracing simulation built upon 3D outdoor airport environment map.

We build the testbeds that take Integrated Sensing and Communication (ISAC) from concept to real systems. The core is a software-defined-radio (SDR) node that senses entirely passively: it captures signals already in the air — unknown waveforms from commercial transmitters or known broadcast signals from a 5G base station — and uses them as opportunistic sources to detect human motion through Doppler signatures, without transmitting anything itself. Alongside it, we develop an OpenAirInterface(OAI)-based 5G base station that extracts channel state information, letting us prototype network-based ISAC in a controlled laboratory before moving to commercial networks.

Channel models for Integrated Sensing and Communication (ISAC) must describe how radio waves scatter off vehicles accurately enough for sensing, yet simply enough for real-time simulation — and extracting scattering centers from detailed 3D point clouds of every car type is impractical. We instead represent a vehicle by simple curved-cuboid geometries, built from nothing more than its length, width, and height, on which scattering centers can be assigned deterministically from the geometry itself. Validated against full-wave simulation of a realistic car, a two-cuboid model with properly chosen edge curvature closely reproduces the vehicle's bistatic scattering — including the side-scattering behavior where a plain box fails — providing a practical target model for ISAC channel simulation and standardization.