基于R阵动态估计的自适应滤波算法

针对组合导航系统中观测噪声特性复杂多变、难于准确估计的问题,基于不同测量系统的测量互补特性,提出了针对单次历元的观测噪声特性动态估计方法。在此基础上,以预设滤波精度为指标,提出了通过构造自适应因子对估计观测噪声进行适当调节的自适应卡尔曼滤波算法。该算法通过构造相对测量关系,避免了直接对测量噪声真值求解的难题,并且在滤波过程中采用序贯处理方法进行实时解算,有效降低了计算量。在GPS/DR实际系统中的应用结果表明,同改进的sage-husa算法及MAKF算法相比,基于R阵动态估计的自适应滤波算法能够自适应地跟踪GPS测量噪声特性的变化,定位结果光滑可靠,具有明显的优越性。

An Adaptive Kalman Filtering Algorithm Based on the Dynamic Estimation of Measurement Noises

To solve the problem that accurate measurement noises are difficult to estimate in integrated navigation systema dynamic method for estimating measurement noises of every single epoch was proposed based on the complementary measuring characteristics of different measurement systems.To achieve the expected filtering precisionan adaptive filtering algorithm which can regulate the estimated measurement noises appropriately with an adaptive factor was put forward.By constructing relative measurement relationthis algorithm avoided the direct solution of true measurement value;and with sequential processing algorithmit reduced the computation cost effectively.The experimental results in GPS/DR system showed that: compared with sage husa algorithm and MAKF algorithmthis algorithm can track the time varying measurement noises of GPS adaptivelyand has smooth and accurate locating result.