Research Areas

Secure and Efficient From green automobiles to malware resistance, researchers in embedded and secure systems are making our devices more efficient while keeping our information secure.

Current Research

Lightweight Cryptography for Internet of Things

As the Internet of Things grows, the majority of devices will be extremely resource- and energy-constrained. Some devices will be used in critical infrastructures, such as transportation or power plants, and must be protected from malicious attacks while using minimal power. We are developing systems that rely on lightweight cryptography, which is as trustworthy as normal cryptography, but uses a fraction of the requirements. Furthermore, we use energy-harvesting techniques to power all its operations.

Tamper-Resistant Design

We are building a microprocessor called FAME, which is resistant to active tampering by hackers, or fault attacks.

Cyberphysical System Reliability

Safety-critical embedded systems must be both predictable and efficient. We are developing a unified analysis and optimization framework, and automatically generating cyberphysical system designs that are efficient, while still satisfying critical timing requirements and avoiding concurrence defects. Our approaches sacrifice accuracy only if necessary, by leveraging the model-based design process with synchronous reactive models to enforce deterministic concurrency behavior. We are developing mechanisms that can be used to construct a correct implementation from the functional synchronous reactive model.

Malware Resilience

Conventional malware defenses are insufficient for critical systems, since such protections are primarily reactive rather than preventative. The NSF-funded Trust Enhancement of Critical Embedded Processes (TECEP) project integrates malware resilience into the control system development process, rather than treating it as an independent problem. We use conventional control system development outputs (control algorithm software), as well as process specifications and models, and apply malware resilience considerations to partition these outputs across the resources available in a programmable system-on-chip. This allows formal analysis to focus on a small subset of the code in software-inaccessible hardware that can detect when the system is about to become unstable. This method can be used in many areas, including Unmanned Aerial Vehicles (UAVs).

Anti-Counterfeiting Technology

We are integrating Physical Unclonable Functions (PUFs), which are the equivalent of a fingerprint for chips, into common electronic components such as RAM, non-volatile memory chips, and programmable hardware chips. This allows the chips to be identified or authenticated. We are demonstrating how OEM circuit boards can use PUFs as counterfeiting protection. The work includes the PUF technology, as well as the protocols needed to integrate it into a networked, embedded system. Our protocol is provably secure, and we have demonstrated the first comprehensive demonstration of an embedded device that authenticates itself to a host PC.