The Research Of Technology And Tracking Algorithms For Air-borne Passive Location

With the development of the electronic countermeasures technology and the missiledefense technology, the traditional Air-borne radar systems have encountered more and moreserious threats in modern high-tech warfare, due to its weak electromagnetic invisibility, weakanti-reconnaissance capability and poor anti-jamming capability. Previous single radar seekerwith Bearings-Only Tracking system can not meet the needs of modern warfare, As theconsummation and complement of active detection systems, passive location and trackingsystem is becoming an important part of the multi-mode radar seeker and has been paid muchmore attention by many countries. The passive observation mode by single observer has manyadvantages, such as excellent invisibility, simplicity in the facility, large effective radius andwide applicability. The subject is focused on Air-borne passive location techniques and itspassive tracking algorithm. It elaborates on the principle of GPS, system model, observablecondition, analysis on the single locating error and tracking algorithm; it also puts forwardsome solutions and applies simulation testing, which shows the validity and practicability ofthe method.Positioning method, observable condition and the location error are the preconditions ofpassive location, the second chapter introduces location theory based on Particle kinematicsfirstly, and on that basis we separately use the Positioning method based on the Dopplerchanging rate and the Complex Positioning method based on the phase and the Dopplerfrequency to model system combining with the features of Air-borne passive location system,particularly analyze the positioning principles of two kinds of positioning methods and alsowe respectively discuss their own suit conditions. Finally, we adopt single analysis on locatingerror to these two kinds of positioning methods and some significant conclusions are obtained,which provides the foundation for studying on the proper passive tracking algorithm ofAir-borne passive location system.Air-borne passive location is essentially a complex issue about nonlinear filtering, whichmakes use of the data of mixed with noises to gauge the target state. In the third chapter weanalyze the characteristics of Single Observer Passive Location with the Information ofSpatial-Frequency Domain and the reason for poor performance of directly using traditionalpassive tracking algorithm, according to the poor observability of the system, oversizeestimated error in initial state and poor convergence accuracy, we put forward Measure SpaceSigma Point Kalman Filter to improve stability and positioning accuracy of algorithm; then we come up with Iterated Sigma Point Kalman Filter by which we enhance the convergencerate and positioning accuracy obviously; at last we put forward Strong Tracking Sigma PointKalman Filter which enhances the adaptbility of algorithm to observation error.In recent years Particle Filter based on Monte Carlo integration technology is proposed,which puts aside the restriction that considering quantity of state as gauss distribution, it usesposterior distribution of monte-carlo random sample. The fourth chapter mainly studies onpassive tracking algorithm based on Monte-Carlo random sample, as the deviation betweencollecting samples by tradition PF algorithm is much larger, a novel Particle Filter algorithmwhich develops out three order Cubature Kalman Filter to work out importance densityfunction is proposed, comparing to Quadrature Kalman Particle Filter the new algorithmenhances its real time performance on the conditions that guaranteeing stability andpositioning accuracy.We come up with Merging Resampling Particle Filter from the standpoint ofResampling technology, which reduces dilution and degradation of particle effectively.When observing noise is gaussian noise, its performance is close to Gaussian Particle Filterand Regularized Particle Filter. When noise distribution is non-Gaussian distribution, itsperformance is better than Gaussian Resampling and Regularized Resampling, so theaccommodated conditions are more broad.As Quasi Monte Carlo (QMC) integration technology can avoid the phenomenon of "clusters" and the" gap" of sampling particles, so it can obtain higher approximate precisionthan MC integral. The fifth chapter mainly studies on passive tracking algorithm based onsequential Monte Carlo particle filter. According to huge workload of Quasi-Monte-CarloGaussian Particle Filter algorithm, combined with the characteristic of Air-borne passivepositioning itself, we separately propose Quasi Monte Carlo Adaptive Gaussian ParticleFilter and Quasi Monte Carlo Merging Resampling Particle Filter,The simulation resultsshow that: two kinds of new algorithms ensure filtering accuracy, and at the same time theyeffectively reduce the QMC sampling particle number, and improve the efficiency of thealgorithm. In addition Quasi Monte Carlo Merging Resampling Particle Filter centers aroundthe Merged particle, which predicts the mean of boundary spaces to carry onquasi-monte-carlo merging resampling, it can effectively reduce QMC sampling scale whiledirecting the particles moving to high likelihood area,it can also optimize space distributionof particle, and then improve the filtering precision, under conditions of the Gauss noise itsfiltering performance is similar to Quasi-Monte-Carlo Gaussian Particle Filter,when thenoise distribution is not Gauss distribution, its performance is superior to Quasi-Monte-Carlo Gaussian Particle Filter, so its applicability is much broader.