Research On Geomagnetic Matching Algorithm Based On Improved PSO

Geomagnetic matching navigation is an important navigation method to realize the autonomous navigation.Geomagnetic field measurement is the basis of geomagnetic navigation.To restrain the random magnetic interference with high-amplitude magnetic from the environment and vehicles in the fluxgate sensor,a noise reduction method based on the combination of multi-scale interference detection and complete integrated empirical mode decomposition is proposed.Based on the technology of aircraft network communication,a geomagnetic matching model for multi-track coordinated navigation is established,and a geomagnetic matching algorithm based on Particle Swarm Optimization(PSO)is proposed.Experiments on the effects of multi-factors on multi-track geomagnetic matching positioning are designed to verify the multi-dimensional geomagnetic matching algorithm Effectiveness;Traditional particle initialization methods will have problems such as particle density reduction and low algorithm convergence efficiency when the matching initial positioning error is large.A Constrained Effective Positioning Particle Swarm Optimization(CEPSO)is proposed.The main research contents of the thesis are as follows:(1)Analyze the sources of geomagnetic measurement errors,and proposed a de-noising algorithm based on the combination of multi-scale Interference Detection and Complete Ensemble Empirical Mode Decomposition(ID-CEEMD)focusing on the random highamplitude pulse magnetic interference from environment and vehicle in geomagnetic navigation fluxgate sensor.Firstly,according to the strong correlation of the geomagnetic signals,the autocorrelation function of each Intrinsic Mode Functions(IMF)is calculated and the effective IMFs are selected adaptively.Secondly,permutation entropy is applied to perform a multi-scale analysis of each effective IMF to mark the IMF interference signal periods.Then,low frequency effective components are extracted through the secondary decomposition performed on each interference period to filter out high amplitude pulse interference information in the IMF.Finally,the filtered signal is obtained by performing signal reconstruction on effective IMF after interference suppression through superimposing.Experiment results show that signal de-noising method based on interference detection and CEEMD can effectively remove magnetic interference,and the Signal-to-Noise Ratio(SNR)is improved 3 times.(2)In the context of aircraft cooperative navigation,a multi-track PSO geomagnetic matching algorithm is proposed.First,the multi-track space constraint model is analyzed.On the basis of the single-track geomagnetic matching model,a master-slave network geomagnetic matching model is constructed.Then,the multi-track space constraint is used to extend the average absolute value error rule from one-track calculation.Matching success rate,matching positioning error and algorithm adaptability are used to evaluate the positioning effectiveness in high-track matching.Simulation results show that multi-track geomagnetic matching can obtain higher matching probability and matching positioning accuracy in a short period.In the local geomagnetic anomaly environment,multi-track geomagnetic matching can still maintain a high matching success rate and matching positioning accuracy;under the interference of the same geomagnetic measurement noise,multi-track geomagnetic matching can obtain a higher matching success rate.(3)In order to speed up the convergence of PSO and increase the geomagnetic matching success ratio,CEPPSO is proposed for the geomagnetic matching.The algorithm uses the geomagnetic measurement redundant information,combining with the geomagnetic prior map,to expand the geomagnetic sequence of the match point,calculates the confidence probability that the matching point is the real position,the algorithm defines the effective positioning area by constraining the confidence density function of the search area,and initialize the particles in the effective positioning area,which effectively increases the probability of particles near the true value and reduces the number of particles.In order to balance the global search and local search capabilities,the inertia weight update rules are optimized.The simulation results show that when the initial positioning deviation is kind of large,the effective localization area under the constraint of geomagnetic redundancy information contains the real trajectory.The particle initialization range is 1/3 of the conventional search range.Under the same number of particles,the matching success rate of CEPPSO is nearly twice higher than that of the conventional PSO algorithm.