Stochastic Circuits for Embedded Signal Processing Applications
Mobile electronics are an essential part of our daily lives. With advances in the semiconductor and sensor technologies, wearable devices and medical implants are also becoming ubiquitous. The area/power/reliability constraints of such devices can be so demanding that they make conventional digital design approaches unsuitable, especially when energy efficiency is of concern. One potential alternative is to employ stochastic computing. This is a (re-)emerging computation technique which processes data in the form of bit-streams that denote probabilities. It can implement complex operations by means of simple logic circuits. We show how efficient stochastic circuits can be designed and employed in important signal processing applications such as medical implants. We demonstrate that the simplicity of stochastic circuits allows massively parallel processing of images in real time. We also show that stochastic circuits are very noise tolerant, a property that is becoming more important as electronic technology advances.
Bio: Armin Alaghi is a PhD candidate in the Electrical Engineering and Computer Science Department at the University of Michigan. He received his Bachelor's degree in electrical engineering and his Master's degree in computer architecture from the University of Tehran. He has received several awards and fellowships for his PhD research. His research interests include digital system design, embedded systems, VLSI circuits, and computer architecture. Host: Gu-Yeon Wei
- Tags
-