This project target to design hardware accelerators targetting neural network applications at the edge, ranging from biomedical applications to computer-vision based applications.

This project works on using noise in the sensor to for ensuring privacy-preserving data transfer in embedded systems instead of using additional compute resources for the same

The goal of Sentry-NoC is to enable side-channel resilience while maintaining high performance, energy efficiency, and low over- head compared to previous works. Sentry-NoC provides a secure platform for communication among malicious IPs and offers extensive protection against confidentiality and integrity attacks, complete protection against availability attacks. Moreover, it provides protection against side-channel attacks by applying temporal and data obfuscation techniques.

This project aims at realizing a ultra low power heterogeneous CGRA architecture which can accelerate general purpose workloads at the edge.

Laser Attack Benchmark Suite (LABS) aims to complete the security evaluation loop against laser fault injection by allowing circuit designers to test their designs agaist well-known laser fault injection attacks and automatically integrate a hardware-based redundancy technique at the early RTL design stage.

This project targeted on leveraging existing sensors available on smartwatches to support dehydration sensing and skin health sensing by integrating a real-time, low-power, highly reusable pH sensor.

This project deals with developing a fast, low-power on-skin AI compute engine (neural network accelerator) to be integrated with current artificial skins.