Research
Circuits and Systems
Broadband filters and applications
We are working on improving the performance of wideband filters and true time delay circuits. These are useful in baseband beamforming of MIMO receivers, pulsed radars, etc. Designing such filters in scaled process technologies is challenging (before you ask, we have to go to lower nodes for reaching the desired bandwidths). This is primarily due to the low output impedances along with other scaling artefacts. We are working on novel architectures as well as improved basic circuit elements like transconductors, for achieving circuits that have performances that push the state of art.
High speed circuits
High-speed circuits are designed to operate at very high frequencies, crucial for applications such as data communication, signal processing, and high-performance computing. They require meticulous design to minimize signal integrity issues like jitter, noise, and crosstalk, ensuring reliable high- frequency performance.
mm-Wave Receiver Frontend
The proliferation of machine-to-machine/IoT in the home and work locations with a variety of complementary and sometimes competing technologies has opened up value-added evolutionary opportunities ranging from healthcare, entertainment, and energy management through intelligent and automated communication between distributed devices. The challenges of integrating front-end passive components and IP onto a CMOS process for high bandwidth millimetre-wave signalling are alleviated due to advances in silicon technology with faster transistors. This opens up vast opportunities to expand connectivity to more devices and contexts (higher bandwidth needed) in our immediate vicinity.
True random number generators
Cryoptography applications need truly random number generator circuits, which need to generate completely uncorrelated stream of bits. These circuits extract entropy from natural random processes like thermal noise. However, for generating high quality random number sequences, the entropy extraction circuits should not introduce any bias or correlation in the extraction process. Theoretically it means that the entropy extraction circuit should have infinite bandwidth! We are working on practical circuits that can generate good quality, cryptographic grade random sequences.
Power Management circuits
Power management circuits play a critical role in ensuring efficient energy usage and distribution within electronic devices. They regulate voltage, manage power conversion, and ensure stable operation, especially in battery-powered systems. Effective power management is essential for extending battery life and reducing heat dissipation in modern electronics.
LED Drivers
LED drivers, another crucial component, control the current supplied to LEDs, ensuring consistent brightness and energy efficiency. They also protect the LEDs from voltage fluctuations, enhancing both performance and longevity.
Instrumentation circuits for MEMS sensors
Suspended Gate FET based MEMS sensors are sensors where the gate is suspended with a spring mass system and is free to move. The movement can be efficiently detected as a change in the transistor chaacteristics (that should give you goosebumps!). While such sensors are quite exciting, working with
them requires novel ways of modeling the device and simulating circuits with it in spice/spectre.
Often times, such sensors are operated in closed loop mode - and operating in closed loop mode
requires high voltage circuits (for us microelectronics guys, anything above 3V is high voltage 
). All things together, this is quite interdisciplinary engineering work and very engaging.
Energy Scavenging Systems
Harvesting energy from ambient sources has become an attractive and promising option for powering the IoT node. Extracting maximum energy, with a minimum loss, from an energy harvesting sources is one of the primary design goals of an energy processing circuit, and to realize it, an optimized energy processing circuitry is required. The task of an energy processing circuit is not only to provide desired voltage or current for the load circuit but also to account the source characteristic, load variation, and environmental changes such that overall system energy can be optimized.
AI and Machine Learning
AI and machine learning (ML) are emerging as powerful tools for optimizing analog circuit design, including power management and high-speed circuits. By leveraging AI/ML, designers can automate tuning, identify optimal parameters, and predict performance outcomes across varying conditions. This can significantly reduce design time and improve circuit efficiency, making the development process faster and more reliable. These technologies hold great potential in streamlining analog circuit design workflows, achieving superior results through data-driven insights.
Sensors and Devices
Flexible electronics
In recent years, non-invasive personal healthcare monitors have gained a lot of attention owing to their capability to detect human physiological signals on the go for disease management and rehabilitation. Wearable diagnostic patches with non-invasive sensors driven by AI are the future of personalized healthcare as it would help monitor real-time biological activity and enable on the spot diagnosis. The development of such a robust system emphasizes the need for novel sensor design and advanced packaging technology. The aim of the project is to develop healthcare based wearable devices for 24X7 monitoring of physiological signals such as EMG, ECG, pulse and BP. As a first step we aim to develop a surface electromyography system for detection of muscular activity on flexible substrate which can span applications all the way from medical diagnostic to personal healthcare.