Research And Implementation Of Radar Data Processing Algorithm Based On DSP

Radar data processing is an important part of radar systems,and its task is to quickly and accurately obtain the tracks of multiple targets in clutter and noise backgrounds.This article takes a low-altitude detection radar as the background of the project,systematically studies the data processing algorithm,explores the method that is most suitable for the radar and the application environment,and completes the software implementation of the algorithm on the high-performance multi-core DSP TMS320C6678.Good performance.This article first explains the overall process of radar data processing and simulates the Kalman filter used for parameter estimation.Then,the following four key links were discussed in depth in conjunction with the project requirements to provide theoretical support for the implementation of the project:(1)During preprocessing,the target may be scanned by multiple consecutive beams in the same period.The sliding window maximum algorithm designed in this paper can condense the points of the same target.In addition,In this paper,the biased estimation problem is derived.In this paper,the compensation formula is derived,which effectively ensures that the converted measurement data after compensation is unbiased.(2)During track association,solutions are provided for the best matching problem of the nearest neighbor algorithm(NNSF)and the clutter density calculation problem of the probabilistic data interconnection algorithm(PDA),and various environments are set up for NNSF,PDA,and joint data.The interconnection algorithm(JPDA)is used for simulation and comparison,and the PDA algorithm is comprehensively considered to consider the association accuracy,the correct tracking rate,and the operation speed.(3)During filtering and tracking,the tracking results of three models of Singer,Jerk,and CS on different maneuvering targets are given.For the problem that the CS model is difficult to select the acceleration extreme,this paper combines the idea of interactive multi-model(IMM)to analyze The model is improved,and simulations show that compared with the CS model with a single acceleration extreme,the improved CS model has higher tracking accuracy for maneuvering targets.(4)During track starting,the basic principles of the intuitive method and the logical method were introduced.Because the radar data processing module in this project receives the measurement set in units of beams instead of scanning cycles,and the logic method is difficult to start when the target turns to motion To solve these two problems,this paper proposes an improved track starting algorithm based on logic based on engineering.Finally,based on the above theories,and around the data structure and the implementation flow of each functional module,the design process of data processing based on the hardware platform is fully explained,and the algorithm optimization method is given,which mainly shows how to efficiently implement matrix multiplication Mathematical operations such as inversion,and reasonable allocation of data storage space using DSP resources.In this paper,Matlab is used to simulate an environment where multiple targets and clutter coexist,and an improved algorithm is used to perform a complete radar data processing simulation.Compared with the real track,the output track meets the design requirements.The comparison between DSP-based running results and Matlab simulation results verifies the correctness of the hardware implementation.After multiple sets of experiments under actual scenarios,the radar system can stably generate a track,and the single processing time meets the real-time requirements,verifying the feasibility of the algorithm and providing a basis for the successful development of the project’s radar.