基于双卡尔曼滤波的系统偏差消除方法

在非线性滤波问题中,一般对噪声有零均值的概率分布假设,但在实际工程中,这样的假设有时不成立,即存在一个固定的系统偏差。把该偏差当作系统待估计的参数,应用双卡尔曼滤波的基本思想,在估计状态量的同时估计出系统偏差,从而自适应地消除了其影响。以一个仅用测角信息的被动定位问题为例,经实验研究表明,提出的方法可显著改善收敛精度。

Method of systematic error elimination based on double Kalman filters

It is commonly supposed that the mean value of noise is zero in the problem of nonlinear filters, but sometimes this is not true in reality, i.e. there is a fixed systematic error. The systematic error is regarded as a parameter to be estimated, then the double Kalman filters are applied to estimate the systematic error and state variables simultaneously. Thereby the influence of the error is eliminated adaptively. As an example, passive location based on bearings-only is considered. The experiments indicate that this method can notably improve the convergence precision.