Tracking Before Detection Based On Dynamic Programming

The detection and tracking of weak targets under complex background clutter is one of the challenges facing the current radar.The track-before-detect technology uses the inter-frame correlation of the target signal to accumulate in the frames and jointly declare the presence of a target and its track,which is currently considered an effective way to detect weak targets.The TBD based on dynamic programming has the advantages of easy implementation and wide applicability,which is studied widely all over the world.In this thesis,supported by the national ministry project,the DP-TBD method is analyzed in detail.For the weak target of different motion characteristics,two improved algorithms are proposed:A CT-DP-TBD algorithm based on clutter test is proposed to improve the performance of traditional DP-TBD algorithm in non-Gaussian background for non-maneuvering or weak maneuvering target.In this paper,the clutter distribution type test method based on the probability density function transformation is introduced into the DP-TBD process.Firstly,the clutter type of the target environment is tested,and then the mixed log likelihood ratio is constructed for different clutter distribution types Function,as the stage value function for energy accumulation,and finally get the target test results and its track.Compared with the traditional DP-TBD algorithm,the SNR requirement of the algorithm is reduced by about 1-2d B,and the average position RMSE index is reduced by about 12 resolution unit,algorithm detection performance improved significantly.For the maneuvering target,the ACS-DP-TBD algorithm based on the current statistical model is studied under the pulse Doppler radar system to correct the target motion model and reduce the computational complexity of the algorithm.The algorithm introduces the current statistical model to modify the target state model,limits the range of acceleration of the maneuvering target,and uses the Doppler information of the target to derive the hidden target radial acceleration state vector under the pulse Doppler radar system.In the extended target state transition area.In order to reduce the computational complexity in the DP-TBD process,a data preprocessing technique is introduced in the ACS-DP-TBD algorithm to reduce the computational effort by setting the initial threshold to filter the original observations.The results show that the proposed algorithm achieves the balance between the algorithm performance and the computational complexity by adjusting the initial threshold in the data preprocessing.The detection performance of the algorithm is better than that of the traditional DP-TBD algorithm,with the increase in the number of accumulated frames,the effective tracking of the target is achieved.