摘要
提出将马尔可夫蒙特卡罗方法与标准的粒子滤波算法有机结合应用于接收机自主完好性监测(RAIM)中。通过状态观测概率密度似然比方法建立一致性检验统计量进行卫星故障的检测与隔离。对算法进行了数学建模,描述了算法的流程。通过实测数据验证,结果表明,该方法在非高斯测量噪声情况下可以对状态进行精确的估计,成功检测和隔离故障卫星,克服了卡尔曼滤波的RAIM算法在处理非高斯测量噪声时性能下降的问题,从而验证了MCMC粒子滤波在接收机自主完好性监测中的有效性。
The investigation presents a new approach combining Markov Chain Monte Carlo method and standard particle filtering for GPS receiver autonomous integrity monitoring. The log likelihood ratio （LLR） test based on probability density function of state-measurement is set up. The consistency test utilizing LLR is devised for satellite fault detection and isolation （FDI）. Mathematic model and algorithm flow for FDI are described in detail. Experimental results based on real GPS data demonstrate that the algorithm can estimate the state precisely under non-Gaussian measurement noise, detect and isolate GPS satellite failures successfully and solve the performance degradation problem of RAIM algorithm based on Kalman filter. Therefore, experimental results validate the validity of MCMC particle filtering for RAIM.
出处
《仪器仪表学报》
EI
CAS
CSCD
北大核心
2009年第10期2208-2212,共5页
Chinese Journal of Scientific Instrument
基金
国家863计划(2009AA12Z312)
交通部科技教育司(200536422504)资助项目
关键词
GPS
粒子滤波
接收机自主完好性监测
马尔可夫蒙特卡罗方法
故障检测
global positioning system
particle filtering
receiver autonomous integrity monitoring （RAIM）
Markov Chain Monte Carlo （MCMC） method
fault detection
