Hardware Complexity Reduction of Parallel FIR Filter Structures Based on Fast FIR Algorithm

摘要

The main objective of the study is to reduce the hardware cost considerably. Parallel Finite Impulse Response (FIR) filters can be implemented with less hardware cost at different level of parallelism using Fast Symmetric Convolution algorithm. Multiplications are major part in FIR filter implementation. The number of required multipliers is reduced by effectively cascading the short length FIR filter using symmetric coefficients in the sub-filter section. The reduction of multipliers and adders are more advantages in terms of silicon area. For example, the proposed structure of a four-parallel 144-tap FIR filter saves 33 multipliers and 7 adders whereas for a eight-parallel 576-tap FIR filter saves 351 multipliers and 11 adders. Finally, when the length of filter is large, the proposed parallel FIR filter structure saves the hardware cost in terms of multiplications in both odd and even length.