Design And Implementation Of FRFT Signal Separation And Detection System Based On Parallel Computing

The radar echo signal is often used for target recognition,but in the actual environment,the received signal is often a mixed signal of multiple signal components,and it is a non-stationary signal.In response to this situation,it is necessary to use time-frequency analysis methods combined with time-domain and frequency-domain information to process signals.Fractional Fourier Transform(FRFT)is a very important method in time-frequency analysis because of many unique properties.It is often used to detect and separate radar signals.In the actual radar signal detection environment,there are often tens of thousands of sampling points generated per second.The rapid processing of these data is of great significance to its subsequent signal separation,target detection and other applications.Although in recent years,domestic and foreign scholars have proposed many fast calculation methods related to FRFT,the calculation rate has not been ideal,which limits its application in engineering practice.How to deal with a large amount of signal data is particularly important to meet its processing speed and real-time requirements.In response to the above problems,this paper combines the FRFT related calculation process and the operating characteristics of the Flink big data parallel computing framework to propose a parallel computing-based FRFT algorithm,and based on this algorithm,designs and implements a parallel computing-based FRFT signal separation detection system.The main work of this paper and the results obtained are as follows:(Ⅰ)Analyze the discrete computing process of FRFT and study the steps that can achieve parallel computing,compare and analyze the two big data parallel computing frameworks,Spark and Flink,and combine the FRFT computing process to illustrate the choice of Flink The reason why Spark is faster.Based on the sampling-type discretization FRFT calculation method,the calculation of discrete linear convolution is improved,and combined with Flink’s simultaneous processing of multiple data streams and the iterative calculation of the same data set,the design of Flink-based FRFT parallel The calculation method is optimized,and the performance of the algorithm is tested experimentally in terms of accuracy and calculation rate.The experimental results prove that the Flink-based parallel algorithm proposed in this paper has no obvious advantage in the case of small-scale data compared with the traditional FRFT algorithm,but when the amount of data increases to a certain scale,its calculation rate is greatly improved.When the volume reaches 16 million,the calculation speed is about 18 times that of the traditional algorithm.(Ⅱ)The FRFT signal separation and detection system based on parallel computing is designed and implemented,which mainly processes and analyzes the radar echo signal.It includes user management,data preprocessing,digital signal data processing,signal separation detection and data storage module,and carries out the outline design and detailed design of the above modules as well as the corresponding engineering implementation.In the process of engineering implementation,this paper compares and analyzes the two widely used big data parallel computing frameworks spark and Flink,expounds the reason why Flink is faster than spark,and selects Flink as the parallel computing processing framework of the system.(Ⅲ)The system is tested by using the weak small aircraft target detection and tracking data set in radar echo sequence,including function test,system interface display and load test,and the test results are analyzed.Finally,it is proved that the system meets the expected effect and can separate and detect the radar signal timely and effectively.