CAJ | 학술논문

针对弹道导弹再入阶段飞行受力情况复杂多变，状态噪声未知时变的非线性跟踪问题，提出基于极大后验估计的STUKF非线性滤波跟踪算法。该算法采用最小偏度单行采样策略，在保证跟踪精度的同时，提高实时性；引入带有多重次优渐消因子的强跟踪算法，在线调整状态一步预测均方误差阵，提高系统对突发机动跟踪的稳定性；采用指数加权的方法，利用次优无偏MAP时变噪声统计估计器，在线估计未知系统过程噪声的统计特性，提高系统应对噪声变化的能力。仿真结果表明：该算法具有比不敏卡尔曼滤波算法(UKF)和扩展卡尔曼滤波算法(EKF)更好的跟踪性能。
침대탄도도탄재입계단비행수력정황복잡다변，상태조성미지시변적비선성근종문제，제출기우겁대후험고계적STUKF비선성려파근종산법。해산법채용최소편도단행채양책략，재보증근종정도적동시，제고실시성；인입대유다중차우점소인자적강근종산법，재선조정상태일보예측균방오차진，제고계통대돌발궤동근종적은정성；채용지수가권적방법，이용차우무편MAP시변조성통계고계기，재선고계미지계통과정조성적통계특성，제고계통응대조성변화적능력。방진결과표명：해산법구유비불민잡이만려파산법(UKF)화확전잡이만려파산법(EKF)경호적근종성능。
In order to solve the nonlinear tracking problems,a new nonlinear filter algorithm,i.e.stong tracking unscented kalman filter,on maximum posterior estimation is presented.The new algorithm a-dopts minimal skew simplex sampling strategy to reduce the computation time and insures the accuracy as well.Unexpected maneuvering is tracked stabely by using strong tracking filter to calculate single-step forecast mean square error.The recursive equations of time-varying noise statistic estimator are given through exponential weight of the constant noise statistic estimator to calculate statistical property of sys-tem condition noise.For this reason,the capability of dealing with variable noise statistic is improved.The simulation results show that the tracking performance of the new method is better than that of the un-scented kalman filter(UKF)and that of the extended kalman filter (EKF).