Vireshwar Kumar

Problem Statement

The applications of cyber-physical systems (CPSs), such as smart homes, intelligent transportation, and smart cities, are proliferating and transforming our everyday lives. However, unlike traditional enterprise computing environments, these systems operate under stringent computational, communication, memory, timing, and energy constraints, which makes the direct adoption of advanced security and privacy mechanisms impractical.

As a result, CPSs rely on two complementary lines of defense: tailored cryptographic techniques to prevent attacks while fitting the fundamental system constraints, and artificial intelligence (AI)-based security systems to detect the attacks that cannot be prevented by cryptographic mechanisms alone.

For such systems, I investigate a fundamental scientific challenge: how can strong security and privacy guarantees be provided for highly resource-constrained, safety-critical, and intelligent CPSs without sacrificing performance, latency, scalability, or usability? Resolving this challenge is critical for these systems to grow sustainably and reach their full potential.

Research Approach

I address the aforementioned challenge through a systematic three-pronged approach: first, emulating these systems for attack and defense experiments; second, analyzing them for security and privacy vulnerabilities; and third, mitigating the discovered vulnerabilities with innovative defense mechanisms.

Concretely, to conduct attack and defense experiments, I develop the software and hardware mechanisms needed to implement core system components in testbeds, and instrument them with domain-specific security, privacy, and performance analysis tools. To discover vulnerabilities, I analyze the limitations of the cryptographic protocols these systems employ using protocol verification tools, and I explore the weaknesses of their AI components against adversarial threats.

To secure these systems, my work moves beyond incremental defenses: it identifies fundamental limitations in existing approaches, develops new abstractions that reformulate the security-performance trade-off, and validates them through rigorous theoretical and experimental analysis. Together, these contributions aim to advance both the scientific foundations and the engineering practice of modern cybersecurity.

Research Areas

I pursue this research with my talented team in the Security and Privacy Research Innovation Group (SPRING). Building on the approach above, the group’s activities are organized along three complementary research verticals, described below.

Robust Intrusion Detection

We develop intrusion detection systems that safeguard diverse environments, including smart home networks, automotive systems, drone communications, speech recognition systems, and edge computing. By analyzing network traffic, device behavior, and user interactions to detect signs of malicious activity, we aim for resilient detectors that produce few false positives, add minimal performance overhead, and withstand adversarial evasion.

Smart Home Intrusion Detection In-Vehicle Network Intrusion Detection 5G Cellular Network Security UAV Security and Regulation Bluetooth Network Security

Lightweight Data Authentication

We develop efficient and secure authentication mechanisms for resource-constrained environments. Drawing on data from 5G networks, in-vehicle and inter-vehicle communication, UAV navigation signals, and IoT devices, we design lightweight cryptographic methods that protect the integrity and authenticity of data and of the devices that produce it, without degrading performance, so that secure communication remains possible even where computation, energy, and bandwidth are scarce.

5G Cellular Network Security UAV Security and Regulation IoT Device Attestation In-Vehicle Network Authentication Inter-Vehicular Network Authentication

Resilient Privacy Preservation

We study robust privacy-preserving techniques for sensitive data, such as the storage and authentication of biometric data, while preserving its integrity and availability. We also explore privacy-enhancing technologies for mobile user data, aiming for systems that withstand sophisticated privacy attacks.

Biometric Data Privacy Smartphone User Data Privacy Inter-Vehicular Network Authentication

My lab, SPRING, has openings for motivated candidates at several levels: MS (Research) and PhD students, postdoctoral researchers, and project staff (research associates and project assistants) on sponsored research projects. B.Tech. and M.Tech. students looking for a project are also welcome.

If you are interested, please send me an email with a brief description of your background and research interests, how they align with mine, and your CV.

Room 214, Khosla School of Information Technology
Indian Institute of Technology Delhi
Hauz Khas, New Delhi 110016, India

Email: (please enable JavaScript to view)
Phone: (please enable JavaScript to view)