COMPLEXITY REDUCTION IN FIR FILTERS USING INTEGER LINEAR PEROGRAMMING TECHNIQU

Abstract

Complexity reduction in FIR filters implies to design techniques used to optimize performance of an FIR filter circuit by reducing its structural complexity. These optimizations improve performance parameters like power consumption, total area occupied by the circuit, when fabricated on a chip and its processing speed. The process involves reducing the complexity of the filter design and consequently that of the hardware which makes it up. Filters fabricated with these techniques show improved performance parameters as compared to conventionally designed ones. This thesis covers the design of FIR filters using the Common Sub-Expression Elimination technique, further enhanced by using Linear Programming to automatically produce filter coefficients in CSD format, using algorithms which maximize commonality of sub-expressions in filter coefficients. This maximizes reduction in filter design and fabrication complexity.