Research Of High-technology Decimation Filter In Modern Communication Systems

Digital filters are commonly used DSP system. They play a key role in the processing of ommunication signal in SDR system. However, the modern communication system's requirements of real time processing of large throughput of information, being handy and low power consumption, put forwards very stringent requirements on the efficiency of digital filters, which can be concluded as the highest operation speed, the smallest circuit area and the lowest power consumption. The three requirements mentioned above are, however, mutually contradictory, hence the designers usually make a tradeoff in order to satisfy the application requirement.There should be two ways as following aimed at raising the efficiency of digital filters.Firstly, study transfer function of digital filter to seek a simplest mathematical expression form. The cascaded integrator comb filter (CIC) is a sort of interpolation or decimation filter with an extremely reduced transfer function and very high operation efficiency. The another sort of filter is halfband (HB) filter which is a special linearly-phased filter, the tap coefficients of which are greatly reduced.Secondly, study the realization structure, i.e. the algorithm to realize the transfer function. Distribution Algorithm forms the corresponding binary bits of the sampled signal sequence in MAC operation into address, which maps into a table, the content of which are the accumulation of coefficients of various taps, thus avoiding the multiplication and raising the operation speed sharply. The polyphase structure decomposes the high-speeded filter into several parallel low-speeded subfilters, thus decreasing the system's power consumption.Firstly the cascaded integrator comb filter are studied. Secondly the DA-based HB-filter, which is widely used in the SDR system, are studied, including transfer function and frequency characteristics. Finally the polyphase structure of digital filter are researched. Several instances of the filters and algorithms mentioned above are verified via Matlab-based theoretical simulation, Modelsim-based Logic simulation and Quartus-based EDA simulation of FPGA implementation, obtaining the simulation results of waveform ,spectrum, circuit consumption and power consumption, all of which fit the corresponding theory well.