Research On Application Of Particle Filter Algorithm In GPS

Global Positioning System (GPS), as the most mature technology and the most widelyused positioning and navigation solution, has been widely used in military, marine, aviation,surveying, transportation and many other fields currently. With the development of globalnavigation satellite systems and the increasing requirements for satellite navigation andpositioning performance, the accuracy and reliability of satellite navigation and positioninghas been paid more and more attention. This paper has done the research on improvingpositioning accuracy of the GPS receiver and GPS receiver autonomous integrity monitoring,and gained certain research achievements.Considering the distribution characteristics of GPS receivers’ measurement noise, thispaper utilized the great advantages of the particle filter in processing nonlinear system withnon-Gaussian noise issues, made use of particle filter for processing GPS positioning data. Toobtain the approximate optimal estimation of system state in non-linear and non-Gaussiandynamic system model, and proposed the improved particle filter algorithm based on Markovchain Monte Carlo (MCMC) methods to solve the issues of particles degradation and samplesdiversity loss, established the dynamic state space model and the observation model todescribe the movement of system. Combined the proposed algorithm with measured data andthe basic particle filter for comparison in filtering effect, the results show that, the proposedMCMC particle filter algorithm can effectively reduce the GPS positioning errors in positionand velocity, it has a better filtering performance than the basic particle filter algorithm.In consideration of the satellite fault detection and isolation issues in RAIM algorithm,this paper proposed a GPS receiver autonomous integrity monitoring (RAIM) method basedon genetic algorithm re-sampling particle filter and the likelihood ratio test method. Improvethe accuracy of particle filter estimation by applying selection, crossover and mutation ofgenetic algorithm into the particle filter, and established the satellite fault detection andisolation (FDI) model, detected and isolated the faulty satellite through testing the particleweights-composed test statistics by applying the log-likelihood ratio (log-likelihood ratio) method. Validated by experimental data, the results show that the proposed FDI algorithmsuccessfully detected and isolated the faulty satellite in the case of non-Gaussianmeasurement noise, and the introduction of genetic algorithm improves the accuracy of faultdetection and isolation, thus verifying the feasibility and effectiveness of applying geneticalgorithm particle filter algorithm in the GPS receiver autonomous integrity monitoring.This paper has done a certain amount of research on the particle filter algorithm and itsapplications in GPS Receiver Autonomous Integrity Monitoring and GPS receiver positioningaccuracy improvement, its findings have a certain reference value in positioning data filteringand receiver autonomous integrity monitoring for the Beidou-2satellite navigation receiver.