| Software-defined radar technology is scalable,reconfigurable and upgradable,which makes the modularization and standardization of radar processing algorithm being particularly important.This thesis is based on the software-defined radar system architecture,with multicore high-performance CPU as hardware platform,using VSIPL standard and MKL efficient scientific computing library to design and implement algorithm components.The logical correctness and real-time processing efficiency of the algorithm component are verified by the sampling data of a ground-based radar system,and the versality and cross-platform portability of the algorithm component are analyzed.Aiming at the processing requirement of a real radar system,the operation index of the system is tested and analyzed by accomplishing software-defined radar system.Firstly,the development and current situation of software-defined radar technology and radar signal processing platform at home and abroad are analysed,the advanced technology achievements and research trends at home and abroad are introduced.The necessity of developing algorithm components,and the advantages of multi-core high performance CPU as a hardware processing platform are presented.Secondly,on the basis of literature research,the hierarchical structure of software-defined radar system is studied,and the structure,interface and functional behavior of hardware layer,middleware layer,core framework layer and radar application layer are elaborated.In order to improve the reusability and reconfigurability of components,a standard algorithm component design structure is proposed.To simplify the component development process,algorithm component development templates and description files are given.To shield the differences between different hardware platforms,development the VSIPL computing middleware with unified specification API is used.In order to improve the processing efficiency of the algorithm component,the MKL library is called at the bottom of VSIPL to realize the calculation part which takes more time.Based on the CPU processing platform,the pulse compression,moving target indicator and clutter map algorithm components are realized.Thirdly,performance testing and result analysis on algorithm components are performed in this thesis.Compared with the simulation results of MATLAB,the logic correctness of the algorithm component is verified.The processing time of components developed based on different methods is tested and compared,it is verified that the processing efficiency of the underlying calling MKL efficient scientific computing function in the VSIPL standard development is higher,and the system real-time is satisfied.At the same time,the algorithm components are analyzed to show versatility and cross-platform portability.Finally,to fulfill the processing requirement,working mode and signal processing algorithm flow of a mechanical scanning two-coordinate radar,a software-defined radar system with CPU as the computing node and integrated the developed algorithm components is accomplished.The communication bandwidth,signal processing delay and packet loss of the algorithm components are tested as the system runs stably,showing that the software-defined radar system realized in this thesis can meet the requirements of the design goals,and has a wide application prospect in engineering. |
PDF Full Text Request