Analog-Based Edge Computing
Compute Locally. Consume Less. Enable Intelligence.
Why move data when we can compute where it is generated?
At COSMIC Lab, we explore energy-efficient analog and mixed-signal computing architectures that process information directly at the edge. By performing computation closer to the sensor, we aim to minimize data movement, reduce latency, and lower energy consumption.
Our research investigates how circuit-level computation can enable intelligent signal processing and decision-making in resource-constrained sensing and IoT systems.
Research Directions
Analog & Mixed-Signal Computing
Exploring analog circuit architectures that perform mathematical operations and signal processing directly in hardware, reducing computational overhead and energy consumption.
Real-Time Edge Intelligence
Developing low-latency, on-device signal processing and decision-making methods for intelligent sensors and connected devices.
Ultra-Low-Power Architectures
Investigating computing architectures designed to operate within tight energy budgets, extending battery life and supporting energy-constrained systems.