| The multi-target tracking needs of the AWACS are related to early warning of large batches of non-cooperative,highly mobile enemy aircraft.According to statistics,in the modern battlefield airspace,the enemy can implement so-called saturation attacks.The number of targets can reach an order of magnitude of 1,000 on average.This poses a severe challenge to the short-term processing capabilities of the AEWACS airborne data system,and it is necessary to find more reasonable and efficient data.Processing algorithm.The traditional multi-target tracking algorithm relies on data correlation technology to deal with multiobjective problems,but the amount of computation of data association increases exponentially.At the same time,a large number of data association operations make the algorithm unable to obtain accurate target state estimation.Moreover,errors in both data association and state estimation can interact and degrade each other.The large amount of calculation associated with the steps has always been one of the problems that cannot be overcome by the existing data processing algorithms,causing the performance bottleneck of the algorithm.In view of the above situation,this paper studies stochastic finite set theory to solve the shortcomings of data association.Random finite set algorithm removes the disadvantages of large amount of computation in the traditional algorithm's data association steps at the theoretical level,so that the multi-target tracking problem has a revolutionary progress.The stochastic finite set theory has the following advantages: When the number of targets is large and the total number changes in real time,the random set algorithm can still stably and continuously track the target when there are more clutters in a complex background.The track initiation,merging,and screening processes in the tracking process of the random set algorithm are all performed automatically and are easier to implement in engineering.This paper combines the application of stochastic finite set of engineering applications,doing the following aspects: First,combined with the interactive multi-model algorithm can effectively describe the characteristics of the target maneuver state,and proposes a multimodel target based on random finite sets.The tracking algorithm significantly improves tracking of high maneuvering targets.Secondly,the tracking and optimization of stealth targets are optimized.Because the RCS of stealthy targets is small and fluctuations fluctuate,the loss rate of points is serious.I propose an L-GMPHD filtering algorithm that preserves the target state at the time of prediction for a period of time,complements the target posterior state given by the filter,and effectively improves the continuity of the track to the stealth target.Third,according to the fact that the new type of early-warning aircraft uses a huge amount of real-time data from multiple sources and heterogeneous sensors,a random finite set algorithm is used for data fusion processing.This paper explores the advantage of using infrared photoelectric sensors with high azimuth accuracy to reduce radar data processing.Iteratively calculates the step error and uses the infrared sensor to provide the starting value of the target so that the track start time of the target tracking is greatly reduced.In summary,the application research of these random set algorithms has advanced the research of a new generation of early-warning aircraft data processing algorithms to make them in complex battlefield environments(such as: strong clutter,low signal-to-noise ratio,strong maneuvering,large batches of targets),Satisfy the performance of airborne multisource sensor data processing for early warning aircraft and achieve effective target detection and fusion tracking in the electromagnetic space battlefield. |
PDF Full Text Request