Research


Surrogate based Optimization and Verification of AMS Designs

 
Ibtissem Seghaier, Mohammed Hussein Zaki and Sofiene Tahar

Contact: seghaier@encs.concordia.ca

To design circuits, current industrial practices rely heavily on simulation based verification and knowledge based optimization techniques. However, such techniques lack mathematical rigor which is necessary to catch up with the growing design complexity besides being computationally intractable. In this research project, we propose a framework to induct the optimization and verification of Analog and Mixed Signal (AMS) designs as depicted in the figure below. First, we perform a surrogate based optimization method with an interplay between a circuit model and a behavioral model of an AMS design. Then, a surrogate based qualitative verification is conducted to ensure a non chaotic behavior of the circuit for the computed nominal parameters. Finally, a surrogate based quantitative verification is developed to prove the circuit robustness to process variation. To illustrate the usefulness of the proposed framework, we are tackling various applications including oscillators, Analog to Digital Converters (ADC) and Phase Locked Loops (PLL).



Matlab Code

 

Publications

 
  1. I. Seghaier, M. H. Zaki and S. Tahar: Cross Recurrence Verification Technique for Process Variation-Resilient Analog Circuits [IEEE International Symposium on Circuits and Systems (ISCAS'16), to appear.].
  2. I. Seghaier, M. H. Zaki and S. Tahar: A Statistical Approach to Probe Chaos from Noise in Analog and Mixed Signal Designs [IEEE Annual Symposium on VLSI (ISVLSI'15), Montpellier, France, July 2015, pp. 237-242]. [Best Paper Award]
  3. I. Seghaier, M. H. Zaki and S. Tahar: Statistically Validating the Impact of Process Variations on Analog and Mixed Signal Designs [Proc. ACM Great Lakes Symposium on VLSI (GLSVLSI'15), Pittsburgh, Pennsylvania, May 2015, pp. 99-102.].
 
 

Concordia University