Complex-Domain FIR Filter Design for Signal Processing Applications

The finite impulse response (FIR) filters play a vital role in a wide range of applications including audio, image, and video processing as well as wireless communication systems. Due to the high computational complexity and memory requirement, designing an efficient and high performance FIR filter is challenging, especially for high-order filters. This paper introduces a new approach for implementation of high-throughput, fully-parallel, and reconfigurable FIR filters through the utilization of processing-in-memory (PIM) technique. Within this framework, the filter coefficients are programmed into memristive devices, which can perform the essential multiply-accumulate operations. This innovative method not only enhances processing efficiency but also underscores the flexibility achieved by leveraging the unique capabilities of PIM in programming filter coefficients within memristors. Moreover, the proposed scheme supports arbitrary filter frequency responses and filter lengths.