04-Aug-2010

Open Study / Bachelor Theses

Comparison of Fixed-Point and Floating-Point Arithmetics in Signal Processing Algorithms

Background: In a joint research project with the Gesellschaft für Schwerionenforschung (GSI), we are working on digital signal processing for real-time control systems. Different signal processing algorithms have been implemented in software on Digital Signal Processors (DSPs) using floating-point computation. Since we need to improve the system throughput, these algorithms are to be implemented in hardware on a Field-Programmable Gate Array (FPGA) without degrading its accuracy.

Objectives: Before switching to a hardware implementation, a comparison of different numeric representations (in terms of data width, value range, precision, …) has to be performed. The influence of these design alternatives on system characteristics such as resource consumption, throughput, noise figures etc. are to be analyzed.

Prerequisites: No particular previous knowledge is required. Basic knowledge in communications engineering, signal processing and digital systems design is desirable, but not mandatory.

Contact: Christopher Spies, Surapong Pongyupinpanich

02-May-2009

02-May-2009