Abrar Ahsan

Papers

Total Citations

H-Index

About

Abrar Ahsan is a researcher at the intersection of embedded systems, computer architecture, and robotics, with a focus on enabling efficient edge intelligence through novel hardware-software co-design. His most cited work, "Compiling CNNs with Cain: focal-plane processing for robot navigation" (2022, 4 citations), addresses the challenge of deploying convolutional neural networks on Focal-plane Sensor-processors (FPSPs)—a specialized camera technology that performs computation directly on the image sensor. This approach achieves ultra-low power and high frame rates, making it ideal for resource-constrained robotics applications. Ahsan’s key contribution lies in developing a compiler framework that maps complex neural network workloads onto FPSPs with severely limited instruction sets and memory, bridging the gap between algorithmic demands and hardware constraints. His work demonstrates practical robot navigation using this paradigm, highlighting the potential for real-time, energy-efficient perception at the edge. Though early in his career, Ahsan’s research has already garnered attention for its innovative integration of sensing and processing, offering a promising path toward autonomous systems that operate without cloud dependency. His contributions are particularly relevant for students and researchers exploring edge AI, low-power vision, and embedded robotics.