基于期望最大化与容积卡尔曼平滑器的机载多平台多传感器系统误差配准算法

An airborne multi-platform, multi-sensor systematic error registration method based on expectation maximization and cubature Kalman smoother

中文关键词  系统误差配准  期望最大化算法  容积卡尔曼滤波器  容积卡尔曼平滑器

英文关键词  systematic error registration  expectation maximization  cubature Kalman filter  cubature Kalman smoother

中文摘要

      针对机载多平台多传感器系统误差配准过程中出现的系统误差参数未知问题, 本文提出了一种基于期望 最大化(EM)与容积卡尔曼平滑器(CKS)的机载多平台多传感器系统误差配准算法. 该算法将传感器的量测系统误 差视为系统待估计的未知参数, 构建了新的传感器量测方程. 引入EM算法框架, 在期望步(E–step)利用容积卡尔曼 滤波器(CKF)和CKS近似计算对数似然函数的数学期望, 在最大化步(M–step)对该数学期望进行最大化处理, 最后 通过解析更新反复迭代的方式获得各传感器系统误差的参数估计. 数值仿真验证了本文提出算法的有效性.

英文摘要

      Considering the unknown parameter of the airborne multi-platform, multi-sensor systematic error registration process, an airborne multi-platform, multi-sensor systematic error registration method based on expectation maximization (EM) and cubature Kalman smoother (CKS) is proposed. Firstly, the method regards the sensor systematic error as an unknown parameter to be estimated, and a new measurement equation is derived. Secondly, the method combines with the expectation maximization algorithm frame, which consists of expectation step (E–step) and the maximization step (M– step). In the E–step, the expectation of the complete data log-likelihood function is approximately calculated based on the cubature Kalman filter and smoother. In the M–step, the approximately calculated expectation value is maximized, and unknown parameter estimations are updated analytically. Finally, the efficiency of the proposed algorithm is illustrated in numerical simulations.