Multi-function Phased Array Radar Signal Processing And Implementation Based On GPU

With the battlefield environment changes,radar must be improved in function and performance to meet the complex battlefield environment.In terms of functionality,radar is required to have search and tracking.In terms of performance,radar is required to have faster data rate,greater computational load and higher real-time performance.Phased array radar can flexibly change the direction of the antenna beam because it changes the phase of the antenna channel,which makes it more flexible than the traditional mechanical scanning radar.In addition,because of its beam agility,the search beam and the tracking beam can be separated,which makes the phased array radar widely used.Therefore,studying the signal processing of phased-array radar with multi-function and higher real-time performance is a research trend of the future radar.Under the above background,through analyzing the signal processing process of the multi-function phased array radar,it is found that there are a large number of parts that can perform parallel computation in the signal processing process,and the graphics processor(GPU)hardware platforms have a very high speedup for processing parallel computing.This paper will adopt a new signal processing form,namely CPU and GPU co-processing method to realize the key technologies in the signal processing of the multi-function phased array radar.Thesis first part introduces the research status of radar signal processing technology using GPU on the whole world.Then,the multi-function phased array radar is briefly introduced,and the basic theoretical of the digital beamforming(DBF),pulse compression(PC)and moving target detection(MTD)in signal processing of search echoes in multi-function phased array radar is elaborated.In the introduction of moving target detection(MTD),this paper proposes a design method of variable notch width MTD filter based on prior information,and gives its associated simulation analysis.The simulation results show that the filter has good clutter suppression effect.Then,two important parts in the tracking process of the multi-function phased array radar are discussed,namely the target motion model and the tracking filter.The common motion models are CV,CA and Singer models.The common tracking filters include linear Kalman filtering,alpha-beta filtering and nonlinear extended Kalman filtering algorithms,and the final simulation results are given.Finally,the hardware architecture and CUDA language of the GPU are introduced in detail.The key signal processing technology in the multi-function phased array radar is implemented by using the GPU,and the Alpha-Beta filtering algorithm is implemented using the GPU.