Research and Realization of Software Radar Signal Processing Based on Intel MKL

Base on Intel Math Kernel Library (MKL), the paper mainly focus on the research and realization of software radar signal processing, which includes magnitude, accumulation and Constant-False-Alarm-Rate(CFAR). Firstly, paper introduces simple signal processing flow of conventional search radar based on ship, then gives the detailed analysis on magnitude, accumulation and CFAR processing module´s computing complexity, then simulates computing time of magnitude, accumulation and CFAR processing module in using conventional algorithms and MKL algorithms, the result indicates that using MKL way to compute the echo samples is much quicker than using conventional way to compute. If the number of samples is no more than 2^19, the time to carry the processing of magnitude, accumulation and CFAR in using MKL is no more than 10 ms, which can be possible to meet the demand of radar project. At last, on the foundation of above conclusion, paper completes the software´s design and realization of the real-time magnitude, accumulation and CFAR processing with 500 echo samples.