Research

Vision

My vision is to develop computational metamaterials that embed intelligence in the physical environment to increase spectral efficiency of communications and resolution of sensing. Currently, the physical environment is the fundamental barrier to wireless systems. Radars cannot see through objects or around corners, and communication links degrade the moment direct signal paths are blocked by walls or obstacles. For decades, researchers have improved endpoints like phones and radars with more antennas, power, and complex digital processing. However, endpoints can only shape own transmissions, not the propagation environment.

Instead of treating the environment as an adversary, my work transforms it into an active, intelligent part of the system. Strategically deployed on buildings and roadsides, my hardware maintains connectivity through walls, reflect signals around corners, and detects hidden objects by sensing scattered reflections. My devices move digital processing workload to analog, performing computation directly on physical signals and manipulating them almost instantaneously with minimal power.

RIS-assisted mmWave Networks

High-frequency millimeter-wave (mmWave) signals (24-71 GHz) provide multi-Gbit/sec data rates from their massive bandwidth. However, they cannot penetrate walls and are easily blocked by obstacles, creating dead zones in both indoor and outdoor environments. I developed programmable surfaces that mount on walls and vehicles to control how mmWave signals propagate. Think of them as smart walls that can make themselves transparent like glass or reflective like mirror to radio waves on demand. They consist of over 4,000 tiny programmable metamaterials, artificially engineered elements that manipulate radio waves. Acting like a microscopic array of adjustable mirrors, these elements instantaneously control how signals bend. By electronically configuring them, a single surface can create new paths through itself, reflect signals around obstacles, and shape beams in complex patterns.

mmWall is the first steerable metamaterial surface that can refract mmWave signals through itself or reflect them around obstacles. mmWall dynamically switches between two modes: (1) glass mode: steering outdoor signals through the surface to reach indoor users; (2) mirror mode: reflecting signals around obstacles to reach blind spots. The core innovation is a novel, see-through 3D structure. While existing metamaterial surfaces are 2D planar, mmWall has horizontally stacked ribs that allow signals to propagate through the structure itself rather than simply bouncing off a flat plane. This approach is fundamentally more efficient than digital repeaters, which receive, decode, and re-transmit entire packets. Instead, mmWall redirects the passing physical waves, bypassing the computational complexity and latency of digital processing.

This video summarizes mmWall’s key contributions in 3 minutes:

RIS-assisted Vehicular Networks

mmWave connectivity in moving vehicles is extremely unreliable. As a vehicle moves, users inside the vehicle must constantly switch between cell towers in a process called handover. To decide when to switch, each user frequently pauses data transmission to scan for nearby cells, causing severe service interruptions and battery drain. To solve this, I built Wall-Street, the first metamaterial surface that simultaneously controls two independent mmWave beams, one to maintain data transmission and another to manage handovers on behalf of in-vehicle users. Wall-Street dynamically switches between two modes: (1) scan mode: the surface acts as both a glass and a mirror. One part remains transparent, maintaining an uninterrupted data transmission between the user and the serving cell. The other part becomes a mirror, reflecting signals from nearby cells to the serving cell. This allows the network itself to measure proximity between the vehicle and nearby cells, offloading the scanning process from the user to the network; (2) handover mode: during handover, the surface combines links from both the old and new cells, delivering duplicate packets to user for a reliable transition. By performing one collective handover for all in-vehicle users, Wall-Street bypasses the complex handover process for mobile users. I integrated Wall-Street into the COSMOS testbed at Rutgers and developed custom features for real-time control. I mounted the prototype on an SUV with two in-vehicle user nodes, driving by three base station nodes

RIS-assisted Satellite Networks

Realizing cost-effective network densification remains another major challenge in 5G, especially in remote and rural areas where deploying fixed infrastructures is not economically viable. Recently, there has been much interest in complementing existing cellular networks with Low Earth Orbit (LEO) satellite networks for coverage extension. LEO networks, however, introduce new challenges: the process of aligning the wireless beam directions between user and satellite is very complex as both the satellite and user moves. Additionally, the constellation of LEO satellites collectively moves over the earth, necessitating a handoff from one satellite to another to serve each user. Such handovers result in increased round-trip-time and a significant drop in throughput.

Wall-E is an electronically tunable surface mounted on a window/skylight, either refracting the satellite link into the home directly or reflecting signals to support outdoor users. Owing to angular reciprocity, Wall-E can simultaneously steer the downlink and uplink beams at the same angle, speeding up the beam alignment process for the uplink via downlink, and vice-versa. Also, Wall-E supports soft handovers by allowing two (or multiple) satellites impinge on the surface at the same time, thus reducing network jitter during handovers. A key obstacle in designing Wall-E was the frequency duplex division (FDD) in LEO networks, which complicates operation due to the use of different frequency sub-bands in the uplink (upper Ku band) and downlink (lower Ku band) directions. Wall-E is the first of its kind to target dual channels in the Ku radio frequency band, allowing ultra-reliable downlink and uplink data transmissions across the globe.

AI-assisted massive IoT Networks

Smart infrastructure needs a brain to process wireless data into actionable intelligence. Inspired by computer vision, I developed a multi-view learning algorithm that merges wireless channels from distributed sensors to model the invisible propagation environment. This approach addresses a fundamental challenge in massive Internet-of-Things (IoT) networks where access points need channel information from each device to allocate resources efficiently, but collecting this information incurs huge overhead that scales with the number of sensors. To solve this, I developed CLCP (Best Paper in MobiHoc ’23, ToN ’25) [10, 11], which predicts channels across sensors to minimize network overhead. I made three key contributions: (1) I adapted multi-view learning from computer vision to wireless communication. Like reconstructing 3D scenes from photos, CLCP treats each transmission as an RF snapshot, combining sparse observations from a subset of devices to form a joint representation and predict channels for adjacent devices; (2) I developed an adaptive view combiner that merges multiple snapshots reliably under fluctuating traffic patterns and varying number of reporting devices; (3) I added model interpretability by mapping raw wireless signals to the learned feature space and comparing against their physical properties (signal travel time and arrival direction). Similar properties clustered together, verifying CLCP learns actual propagation rather than working as a black box.