基于高斯混合分布及EM算法的抗干扰定位方法

Anti-interferences method based on Gaussian mixture distribution and EM algorithm

针对当前导航接收机自主完好性监测算法(RAIM)在欺骗式干扰环境下,存在多故障识别能力弱、计算复杂度高等问题,提出一种基于高斯混合分布及EM算法的抗干扰定位方法,用于存在生成式、转发式欺骗干扰的复杂电子对抗环境。该方法提出一种最小截尾平方和(LTS)估计的快速算法,从而得到高可靠的迭代初值,首次采用高斯混合分布建立极大似然估计模型,并使用EM算法对其进行求解得到可靠的定位结果,最终计算实时定位误差并实现多个干扰的识别。仿真结果表明,所提方法在干扰数量为10时,与MM估计方法相比定位精度提高53%,计算时间缩短到原先的1/5000,具有更高的定位精度以及更短的计算时间,可为电子对抗环境下的高可靠卫星应用提供支撑。

Navigation receivers mostly use the receiver autonomous integrity monitoring(RAIM)algorithm to monitor single-interference and double-interferences.This algorithm has high computational complexity for three-interference and above modes,so it is rarely studied in practical applications.Based on the Gaussian mixture distribution and the EM algorithm,a fast algorithm for least trimmed squared error(LTS)estimation is innovatively proposed to obtain a highly reliable iterative initial value.The Gaussian mixture distribution is used to establish the maximum likelihood estimation model,and the EM algorithm is used to solve it to obtain reliable positioning results.Finally,the real-time positioning error is calculated and multiple interference stars are identified.The simulation results showed that,when the number of interference satellites is 10,the positioning error and running time of the proposed method are 47%and 1/5000 of the MM estimation method respectively which can provide support for positioning in electronic countermeasure environment.