Study On Real-time Signal Processing Technology For Airborne Forward-squint SAR Based On DSPs

Synthetic aperture radar(SAR)is widely used in both military and civil areas because of its all-day,all-weather and high-resolution features.Digital Signal Processor(DSP)is widely used in airborne SAR real-time signal processing system because of its excellent performances.With the increasingly higher requirements of system performance in practical application,single-chip and single-core processor cannot meet the demand,so multi-chip and multi-core processors are needed.This paper studies the real-time signal processing technology for air-borne forward-squint SAR based on multi-chip and multicore DSPs,and the study content is shown as follows:Firstly,the model for echo signals and basic imaging theory of SAR are presented.Then several conventional SAR imaging algorithms with high resolution are mentioned.By analyzing and comparing the performance of these algorithms,the theoretical basis for the selection of imaging algorithm while designing the real-time imaging algorithm is provided.In this paper,basic theory and advantage of the stepped-frequency SAR are discussed and methods to stich the stepped frequencies in time-domain or image-domain are studied.In addition,aiming at the problem of air-borne platform motion instability,the effects of the motion error on the image quality and several estimations of doppler parameters are researched.Secondly,this paper studies the multi-nodes collaborative methods for real-time processing.Some performance indicators in multi-nodes collaborative signal processing are introduced and the factors that can affect performance of the multi-nodes collaborative system are dicussed.Two factors,multi-nodes collaboration strategy and interconnect structure,are studied in this paper.And a mixing collaborative method of multi-nodes,which has better performance than traditional collaborative methods,is proposed.Thirdly,a SAR real-time imaging system based on multi-chip and multi-core DSPs is designed.The real-time imaging algorithm is presented.The real-time signal processing system and the model of signal processing based on mixing collaborating using multi DSPs are designed.Lastly,the correctness and effectiveness of the design of the system are demonstrated with processing of experimeantal data.