Research Overview

My research is focused on the full stack of a communication network, starting from PHY, MAC, and network layer, with an extensive focus on prototyping and validating the solutions on an experimental testbed. I have used tools from online learning, causal inference, and signal processing for my research.

In the following, I have put an overview of the published projects.

Graduate Research

Automation and learning in cellular network

Proposed a Bandit learning-based approach to user clustering based on the similarity in channel distribution and the associated rate regions.
Designed a heuristic-based approach to an online user selection to optimize the network throughput.


Defense against eavesdropping attack on low-resolution phased arrays [Project page]

Developed a low-complexity physical layer defense against eavesdropping attacks on low-resolution phased arrays.
Designed experiments on the mmWave testbed to verify the defense mechanism.


Side information-aided mmWave beam alignment designs

Designed a beamforming mechanism for mmWave communications based on the side information.
Setup mmWave testbed with SiBeam phased arrays and designed calibration procedure.

Undergraduate Research

Analytical Modelling of Bluetooth Low Energy Protocol based Sensor Networks

Supervisor: Dr. Debashis Ghosh, Indian Institute of Technology Roorkee (July 2016 - May 2017)
Developed an analytical model for analysis of Bluetooth Low Energy protocol-based sensor networks.

NS3-based simulator for Bluetooth Low Energy protocol-based sensor networks

Supervisor: Dr. Neelesh Mehta, Indian Institute of Science, Bengaluru (May - July 2016)
Developed a BLE protocol module on NS3 for end-to-end protocol simulation.

Distribution-free Spectrum Sensing Technique for Full-Duplex Cognitive Radio

Used likelihood ratio-based statistic for spectrum sensing for Full-Duplex Radios.

Hobby projects

gr-bokehgui: Web-based display for GNURadio [Link]
GNU Radio, Google Summer of Code 2017

MIPS Single Cycle Microprocessor using VHDL [Link]
Implemented a processor capable of running 7 basic instructions.
Added Python script to act as a compiler for converting assembly code to ModelSim simulation script (binary instructions).