拟线性最优平滑滤波及其应用

QUASI─LINEAR OPTIMUM SMOOTHING FILTERING ALGORITHM AND IT5 APPLICATION

在普通推广卡尔曼滤波（Ｅ．Ｋ．Ｆ．）的基础上，提出了一种更加适用于非线性状态方程和观测方程的改进的滤波算法─—拟线性最优平滑滤波（Ｑ．Ｌ．Ｏ．Ｓ．Ｆ．）算法，在求该滤波算法的状态转移矩阵时，加进了二阶导数项，同时线性化的标称值取为单步平滑值，降低了非线性对滤波的影响，提高了滤波精度。通过对某歼击机的仿真飞行试验数据及实测数据的处理，表明该算法比普通推广卡尔曼滤波具有更好的收敛性、精确性。

In this paper, an improved filtering algorithm───Quas─Linear OptimumSmoothing Filtering(Q.L.O.S.F. ) is proposed on the base of common Extended KalmanFiItering(E.K.F.).The algorithm is more suitable than E.K.F.for nonlinear dynamicsystems. Second order derivative item is added in the seek of the state transition matrix andone─step smoothing value is used as the nominal value of system─linearization , so thenonlinear effect on filtering is reduced and the filtering─accuracy is increased. This filteringalgorithm is successfully applied to estimating state variables in the aircraft six degreefreedom nonlinear dynamic equations and reconstructing angle of attack , sideslip angle. Theresult of processing some aircraft simulation and measured flight data shows that the methodhas better convergence proprety, accuracy and stability than E.K.F..