Communication Digital Filter Design

Digital filter is commonly used as it is flexible, stable, immune to inner noise and has low hardware cost. And digital filter becomes a very important part in the modern communication system. Whileas Software Defined Radio is a hot topic in the communication system. General digital filter and fractional sample rate converter match the central idea of SDR, and linear phase is important for the digital filters in communication systems.The central idea of SDR is that it implements all kinds of functions through software programming based on a general standard hardware platform and makes the Digital Signal Processing module and Analog to Digital Converter as close as possible to the antenna. In the digital base-band, there needs all kinds of digital filters such as Pulse Shaping filter, Noise Cancellation filter, Multi-rate filter, frequency analysis filter, etc; the above filters have different design targets in different communication protocol, the chip area will be large if design a partly general filter for each kind of filters classified by the above function, while the throughput and power will be a challenge if general processor is used. Therefore, a compromise plan is needed, the throughput must be satisfied, and all kinds of the above in all kinds of protocols can be implemented while the area and power is as small as possible. Based on the background, this paper proposed general filter architecture based on lattice wave digital filter that has high stability and low sensitivity to coefficient. Based on the analysis of all kinds of digital filters in Software Defined Radio and lattice wave digital filters, this paper proposes a general filter architecture and verify it through FPGA.Software defined radio can implement many communication protocols, and the data rate differs among different protocols, but the entire system has one analog to digital converter for data sampling. Therefore there need multi-rate filters. Based on the analysis of Farrow filters, this paper proposes a new simplifiedμ-Generator used in fractional sample rate conversion, and do the simulation.The linear phase characteristic of digital filter is important in communication system, especially the system using phase modulation. Whereas infinite impulse response filter is used commonly due to its low order and small delay, but its phase is not linear usually. Based on the existing algorithm that implements real coefficient IIR filter, this paper expands the algorithm to fit in complex coefficient digital filters, and verify it through MATLAB.