摘要
对飞行中风场测量值含连续野值较多的问题,提出了将连续野值当作噪声处理的方法。噪声设置为随机游走模型并在状态方程中引入时变系数,利用辅助粒子滤波(APF)处理。与当前的自适应Kalman方法进行了比较,在含10个连续野值的模拟数据处理中,Kalman方法发生了跳变,而APF方法成功地处理了连续野值;APF方法和Kalman方法的平均均方误差分别为0.8313和1.0021。最后,将APF方法用于飞行测量数据处理。结果表明,APF方法能处理更多的连续野值,具有更好的精度和稳定性,适合工程应用。
To solve the continuous outliers eliminating problem for wind estimation in flight, a method was proposed by viewing the outliers as noises. The noises were assumed to be Gaussian random walk and the coefficient of the state equation was set to be time-varying. The auxiliary particle filter （APF） was chosen for its filtering. This method was compared with previous adaptive Kalman filtering. In synthetic simulations with 10 continuous outliers, this method successfully eliminated the continuous outliers while the Kalman method made a jumping estimation. The average mean squared error of this method and the Kalman method was 0.8313 and 1.0021. Finally, this method was used for flight measurement data processing. Processing results show that this method can eliminate more continuous outliers, has better precision and stabilization than the adaptive Kalman method, and should be applicable to wind estimation in flight.
出处
《系统仿真学报》
EI
CAS
CSCD
北大核心
2008年第22期6248-6251,共4页
Journal of System Simulation
