Research For The Parallel Algorithm Of FIR Filter At IF Digitalization Of Software Radio System

In recent years, there has been an increasing interest in the research area of software radio. Software radio is one of the most important emerging technologies for the future of wireless communication services. A software radio can be regarded as a set of primitive Digital Signal Processing (DSP). There are some multilevel systems which combine the primitives into communication functions, and some other target processors on which the software radio is hosted for real-time communications. Countries in the west are paying much attention to the research of software radio, and working at all aspects for many years. However, current systems are really far from the idea architecture of software radio technology, because the technical level and condition are not perfect. Especially, IF digitalization requires a high processing speed. Furthermore, real-time processing is also important. Parallel DSP architecture has been the chief scheme to solve this problem.In this paper, based on the improved architecture of software radio, we present a new parallel algorithm for the FIR filter of IF digitalization. We also investigate this parallel processing system in terms of performance and efficiency. The implementation of the parallel algorithm is discussed and results are given to show that the performance of the parallel system is much better than a single processor system, and this parallel algorithm is more suitable for multi-processor than pipeline algorithm. FIR filter is only an example, but this idea of parallel algorithm can be used for the implement of other functions on the improved software radio platform.