The Research Of Parallel Signal Processing Algorithm In PD Radar

The pulse Doppler(Pulse-Doppler,PD)radar is a new radar system that utilizes the principle of Doppler effect,which not only has good ranging performance and speed measurement performance,but also has strong clutter suppression capability,so,it has been widely used.With the development of software radar,the traditional development model based on FPGA+DSP has gradually been replaced by the development model of software radar?One of the bottlenecks in software radar implementation is the real-time nature of radar signal processing.With the rapid development of the computer hardware,especially the multicore,many-core architecture and the heterogeneous computing platform,it provides the hardware foundation for the parallelization of the software radar.The parallel programming framework is a bridge for its parallelization on computing platforms.This paper presents a parallel radar signal processing algorithm based on different platforms,and research on the whole signal processing system and some modules.The main work of this paper as follows:This paper proposed a parallel pulse doppler radar signal processing algorithm based on multi-core CPU.Firstly,parallel analysis of the signal processing process is performed.It is found that there are three parallel points in the signal processing process.This paper mainly designs different parallel algorithms for these three parallel points.Firstly,aiming at the characteristics that there is no interdependence between the data of each pulse group in the radar signal processing process,a parallel idea based on data is proposed.Using the multi-core characteristics of multi-core CPU,the data of different pulse groups is distributed to different threads.That is,each thread processing a single pulse group of data;Secondly,the processing is also pipeline,that is,the output of the previous level is the current level of input,and the current level of output is the next level of input,so this paper uses pipeline parallel technology to parallelize the entire process flow,according to the principle of static load balancing,the algorithm module is assigned to different threads,through the communication mechanism between the threads to control the transmission of data.Finally,parallel analysis is performed on the inside of the algorithm module.Based on the characteristics that the processing between the pulses within the module,such as pulse compression and moving target detection,is independent of each other,this paper proposes a parallelism within the module.The idea is to assign different pulses to different threads for processing.Finally,this paper analyzes the experimental results of parallel pulse Doppler radar signal processing algorithm based on multi-core CPU.This paper proposed a parallel moving target detection algorithm based on Open CL.Open CL is a crossplatform parallel programming model that supports hardware platforms including CPU,GPU,MIC,and FPGA.This paper designs the parallel moving target detection algorithm based on the parallel analysis of moving target detection algorithm.The different distance units is mapped to the computing core for processing in parallel algorithm.The GPU-side parallel moving target detection algorithm is implemented,and then some optimization is performed.Meanwhile,transplant the parallel algorithm to the FPGA platform,Also,adopts some optimizations to optimize the algorithm,such as vector,loop unrolling and pipeline.Finally,the experimental results are analyzed.