基于改进粒子滤波的GPS接收机自主完好性监测算法

GPS Receiver Autonomous Integrity Monitoring Algorithm Based on Improved Particle Filter

针对基本粒子滤波重采样过程中粒子权值退化和多样性丧失的问题,将遗传算法引入基于神经网络的权值调整粒子滤波算法中,结合了遗传算法全局寻优的收敛性与神经网络局部寻优的快速性优点。将提出的算法与对数似然比方法结合用于GPS接收机自主完好性监测,通过建立一致性检验统计量实现对故障卫星的检测与隔离。通过采集实测数据进行验证,结果表明:该算法可以成功检测和隔离故障卫星,其性能优于基于基本粒子滤波的接收机自主完好性监测,验证了该算法应用于GPS接收机自主完好性监测的可行性和有效性。

In order to overcome the degeneracy phenomenon and the sample impoverishment problem in parti-cle filter （PF） during resampling process, an algorithm combining the genetic algorithm（GA） with neural network-based weights adjustment particle filter（ NNWA-PF） is presented,which possesses the advantages of the convergence for global optimization of genetic algorithm and the rapidity for local optimization of neural network. The proposed algorithm combined with the log-likelihood ratio（LLR） method is applied for GPS re-ceiver autonomous integrity monitoring（ RAIM） . Through setting up the consistency of the test statistic to a-chieve fault detection,the proposed method is tested and verified by GPS measured data. The results show that this method can successfully detect and isolate the fault satellite and improve the performance of fault detection. Therefore,the improved particle filter algorithm is feasible and effective for RAIM.