Research On Radar Multi-frame Target Association And Filtering Technology

In modern military war,the radar reflection cross-section of stealth fighters is so small that the radar is almost difficult to detect,which requires the radar to have the ability to effectively detect weak targets.The traditional Detect-Before-Track(DBT)technology loses the weak target information contained in the data due to threshold detection of the original baseband data of each frame of radar,resulting in a sharp decline in its ability to detect weak targets.Track-Before-Detect(TBD)technology can effectively improve the radar’s ability to detect weak targets.It processes multi-frame radar original baseband signals together,and enlarges the difference between the target and noise by increasing the time dimension,thereby improving the algorithm performance of target detection.The task of radar is to track the target for a long time and with high precision.However,the plot-sequences output by TBD are low in accuracy and short in duration,so it is necessary to perform tracking processing on the plot-sequences,but the existing research work is lack of attention to the filtering and correlation of the plot-sequences.This dissertation focuses on the problem of radar multi-frame target association and filtering,and has carried out theoretical analysis,method research,simulation experiment,and actual measurement data verification.The main work is as follows:1.The radar target detection and tracking model based on Bayesian estimation is studied,and the details of the TBD algorithm based on the dynamic programming(DP)algorithm are given(DP-TBD),which provides a theoretical basis for the research of radar multi frame target correlation and filtering technology.2.The structural characteristics of sliding window batch processing DP-TBD and the batch structure of sliding window are studied;then,the plot-sequences model based on the sliding window batch processing structure is established;the covariance reduction rank applicable to the plot-sequences is proposed.Based on the covariance crossover method,the decoupling multi-frame target filtering algorithm solves the filtering problem of the unknown correlation between the point and trace sequences.3.Propose a single-time hypothesis test multi-frame target association algorithm suitable for plot-sequences and target trajectories.By studying the single-time hypothesis test method,the matching problem between target trajectories and plotsequences is solved.The above models and methods are verified through theoretical analysis,simulation experiments.The analysis results show that the above methods can effectively improve the radar’s tracking accuracy and tracking capabilities for weak targets.