基于空频域信息的改进UKF算法研究

摘要

固定单站无源定位系统面临着可观测性弱、初始误差大的问题,为了实现稳定高精度定位,在定位模型中引入角度变化率和多普勒频率变化率信息,并在此基础上提出了一种基于空频域信息的改进不敏卡尔曼滤波(UKF)算法.该算法利用两次观测时刻之间的间隔,根据当前时刻定位结果,通过后向平滑算法平滑估计前一时刻状态向量和协方差矩阵的估计值,为前向UKF算法提供较高精度的起始值,从而得到对当前时刻状态更为准确的估计;仿真结果表明,增加高精度的角度变化率和多普勒频率变化率观测量,能够显著改善定位性能,改进的UKF算法在保证实时性的基础上提高了定位性能.%The single non-moving passive location system is confronted with problems as low observability and large initial error. In order to realize stable and high-precision passive locating by a single non-moving observer,the angle changing rate and Doppler changing rate information were introduced into the location model. An improved Unscented Kalman Filter (UKF) algorithm based on spatial-frequency domain information was presented. The proposed algorithm smoothed the previous state vector and covariance matrix by the backward-smoothing algorithm using the current location results, then an initial value with higher precision was obtained to get more precise locating results. Simulation results indicated that the location performance was improved significantly by adding the angle changing rate and Dopple changing rate information, and the improved UKF algorithm can improve the locating performance while keeping the realtime characteristic.