Research Area
Memristor for Neuromorphic Computation
Why is it important? Due to the frequent data travel between memory and CPU, the Von-Neumann architecture suffers from high latency and power consumption. So, near-memory or in-memory computation is the possible solution. Neuromorphic computation architecture is based on artificial neural network (ANN) where the memory and data processing units are inseparable, like human brain. Memristors show potential to emulate various neuromorphic functionalities and they are promising for hardware implementation of ANN.
What we do? Fabricate memristive devices based on organic and/or bio-organic materials and perform different electrical characterizations of the fabricated devices to access their memory and neuromorphic performance.
How we do? Design devices by means of metal contact and material engineering. Low cost techniques like spin-coating, thermal evaporation etc. are used. Electrical characterization of the devices using a probe-station connected to source meters.
Organic material based resistive switching memory
Read More: Memristor using organic small molecules, bio-organic materials.
Read More: Realization of synaptic weight modulation through memristor conductivity modulation
Energy Harvesting for Self-Sustaining Systems
Why is it important? Internet-of-Things (IoT), mobile and wearable electronic devices are primarily powered by batteries which have limited lifetime and require frequent battery replacement. For uninterrupted operation, the next-generation IoTs and wearable systems should be energy autonomous. Energy harvesting at the point of application is the possible alternative.
What we do? Fabrication of energy harvesting devices, known as nanogenerators, which harvest electrical energy from other form of energy ubiquitously available in the ambient. The devices mostly Triboelectric Nanogenerators (TENGs) which can generate electrical power in milliwatt (mW) range.
How we do? Fabricated TENGs are characterized to access their power generation ability. Performance improvement is done by means of device structure and/or material engineering. The harvested energy is managed and stored in a storage device to drive small electronic devices.
Material engineering for performance enhancement
Read More: Molecular self-assembly for orbital energy level modulation leading to enhanced charge density
Self-Powered Smart Sensing Devices
Why is it important? According to reports, the number of IoT devices will be larger than 40 billions by the year 2025. Therefore, a significant amount of energy is consumed globally by the so large number of IoT sensors. Self-powered active sensors, that does not require external energy to operate, is the potential solution to reduce the energy budget.
What we do? Traditional sensors are passive in nature as they require external power supply to generate electrical signals by detecting any change in their electronic properties (resistance, capacitance, etc.) due to the presence of external stimuli or analytes. The investigations are focused on active sensors that generate electrical signals directly from the energy contained in the stimuli.
How we do? Device structure and the sensing materials are engineered to improve sensitivity and signal dimensionality. Sensed signal is fed to machine learning (ML) algorithms for artificial intelligence (AI).
Self-Powered sensor for real-time detection of vehicle speed
Distinct signal patterns to distinguish the biomechanical movements