Research On Multi-target TBD Detection Techniques Based On Background Statistics Information For High Frequency Radar

HF Surface Wave Radar, is also known as surface wave over-the-horizon radar(High Frequency Surface Wave Radar, HFSWR). Its working principle is thatmaking the use of vertical polarization electromagnetic wave can diffract on theearth’s surface, and on the way to the moving target and the marine environment ofthe scope of the400km can be monitored over the over-the-horizon with real-time.However, just like other types of radar, high frequency surface wave radar also hasits own problems and limitations. The complex clutter background with ionizedlayer clutter and sea clutter as the main body to ship target detection and trackinghas brought a great challenge, and is also been focused on by researchers.Therefore, research on such a complex electromagnetic, time-varying sea stateenvironment, how to improve the dim target detection and tracking performance,becomes very challenging and necessity. In view of the above questions, this papertries to seek a kind of tracking and detection algorithm, which can improve thestatistical properties of complex, homogeneous heterogeneous clutter coexistingdetection background of weak target detection performance of tracking. This processincludes the traditional TBD algorithm, multiple targets with complex clutterbackground of the F-Viterbi algorithm, the new algorithm BD-Viterbi algorithmdemonstration effect, a variety of algorithms in practical tracking results based oncloud model of the fuzzy comprehensive evaluation of four main modules. The fourmodules the main contents are as follows:1. Sorting the existing track-before-detect algorithm, classification, and givingthe traditional Viterbi algorithm process, the process and the advantages anddisadvantages are analyzed and summarized. For complex environment of the targetdetection and tracking, to improve the signal to noise ratio is a very important step,while the track before detect technology to good use to detect target multiple framesof information, increasing the signal-to-noise ratio, and then improve the low SNRtarget detection probability. But want to further refinement, first of all should fullyunderstand the method aspects, in the understanding of where the advantage, itsshortcoming in where, the direction of improvement will probably determine the. Itshould be said that the work was behind a new improved algorithm based andnecessary steps and work.2. After the traditional track before detect technology research, and further tothe F-Viterbi algorithm for the study of comparative. Choose the algorithm, becauseit is a recently proposed for complex environment of the single target trackingdetection method, it not only can obtain the very good detection effect, but also can keep less computation, in the detection and computation at the same time the twohave improved, so the algorithm is indeed it is worth learning and research. Thealgorithm for single target algorithm, this paper applies the multiple objectives, theshow effect of comparative study, so as to provide the data support the follow-upalgorithm improvement and breakthrough. The traditional Viterbi algorithm usingreverse search mode, can make good use of the historical frame information, but as aresult of this process for each of the data units are involved in the computation, sothe calculation amount is very large, real-time poor effect. The F-Viterbi algorithmbased on the forward search mode, and in the trajectory matching before, the firstconducted CFAR processing, the interference data unit to filter, reduce the amountof calculation, the calculation speed is improved and the real-time processingability.3. After the above two aspects of the research, integrated traditional Viterbialgorithm and F-Viterbi algorithm, and in the light of their respective faults,BD-Viterbi algorithm is put forward. The traditional Viterbi algorithm has theadvantage of search direction, to make good use of the historical frame information,and can greatly improve the target SNR. Disadvantages are arbitrary data unit to seecomputing, resulting in excessive computation, affect the real-time; F-Viterbialgorithm has the advantages of being in each frame prior to treatment, will be firststep of CFAR processing, excluding those interference data unit, reduce the quantityof data involved in the computation of the unit, so can have limited computingresources more and more applied to the effective data unit processing. TheBD-Viterbi algorithm is very good with the above two kinds of algorithm. It uses abidirectional search mode, in order to make better use of the historical frameinformation, at the same time for each frame of data in advance one stepOS-CA-CFAR processing, in order to obtain better computation efficiency. Throughthe improvement, the detection performance and computation two respects obtainbetter performance.4. Finally, new improved algorithm has a more comprehensive evaluation basedon cloud model, using the fuzzy comprehensive evaluation method to evaluate theeffect of tracking algorithm. That is that in the field of evaluation, means of fuzzyhas been applied more and more, although it can be very good for the ambiguity ofthe concept, but it cannot be well expressed for the randomness of the concept. Andcloud model in recent years begans to be recognized as a new means, it will be avery good means of combining the concept fuzziness and randomness, so it couldgive a more comprehensive, more reliable evaluation. So this paper applies thefuzzy comprehensive evaluation method based on the cloud model, to make betteruse of the advantages of the two algorithms, in order to give a comprehensiveevaluation and comparison of the tracking effect. This paper presents a new TBD algorithm can well adapt to the HFSWR lowSNR target detection and tracking, and also be generalized and applied to othermultiple domains.