Interval Arithmetic Multiplier for Digital Signal Processors
Yasir Saleem, Muhammad Khawar Bashir, Asma Umar, Asma Mushtaq · 2014
Interval methods offer quite new research aspects regarding Digital Signal Processing. In scenarios like finite numeric representation, limited precision sensors or the quantization process, signal processing has been practicing interval mathematics as quite a reliable gizmo for expressing the uncertainties. Some types of control systems like estimating or soft computing systems, the uncertainties are caused due to the presence of variable instability, variance of signal, or the rate of safety of some actuators. Since most DSPs are designed to work in real time environments, effective implementation of interval arithmetic for these processors asks for fast processing capabilities. Hence the aim is to develop an environment that facilitates interval arithmetic operations to be executed at the very same computational speed as is offered in many contemporary signal processors. So we are proposing the design and implementation of Interval Arithmetic multiplier that functions with IEEE 754 numbers. This design carries out the execution of an algorithm which performs better than conventional algorithm of Interval multiplier. The performance of proposed architecture is better than that of a conventional Canonic sign digit (CSD) floating-point multiplier, as it can perform interval multiplication and floating-point multiplication as well as Interval Comparisons.