Project information

  • Team Member: Alan Hsiao, Chloe Wu,
  • Emily Wang, Grace Tan
  • Category: VLSI Design
  • Project date: Spring 2020
  • Project Documentation

Unsigned 8-Bit Discrete-Time Convolution

Overview:

The 8-bit discrete-time convolution circuit was a final project for ECE4740: Digital VLSI Design. I worked in a group of four on this project.

Convolution is an essential technique employed in digital signal processing. It is commonly used for common everyday applications such as image or audio processing, as well as Convolutional Neural Networks (CNN), a class of deep learning.

In order to make our system more efficient (meaning to have the smallest area delay product) while consuming as less power as possible, we chose to implement our system using a modification of the replication architecture. Our system can take two inputs of length three. Each number of the inputs can be up to 8-bit. The design, simulation, and layout of the system was done in Cadence, and the testing was done in both Cadence and MatLab.