CAJ | 학술논문

针对单站无源定位系统可观测性弱、滤波稳定性差、收敛速度慢和滤波精度低的问题，提出了一种基于迭代的无迹H∞滤波算法。该算法将UT变换引入到扩展H∞滤波中，避免了计算线性化的雅克比矩阵，增强了滤波效果。在迭代判决准则的约束下，重复利用观测信息对状态向量和协方差矩阵进行估计使其更趋于真实值。仿真结果表明，在不同的观测精度环境下，该算法在稳定性、收敛速度和定位精度方面表现更好。
침대단참무원정위계통가관측성약、려파은정성차、수렴속도만화려파정도저적문제，제출료일충기우질대적무적H∞려파산법。해산법장UT변환인입도확전H∞려파중，피면료계산선성화적아극비구진，증강료려파효과。재질대판결준칙적약속하，중복이용관측신식대상태향량화협방차구진진행고계사기경추우진실치。방진결과표명，재불동적관측정도배경하，해산법재은정성、수렴속도화정위정도방면표현경호。
Aiming at such defects of the single-observer passive localization system as weak observability, inferior filtering stability, slow convergence speed and low filtering precision, a kind of iterative unscented H∞( IUH∞) fil-tering algorithm was presented. In the algorithm, the unscented transformation ( UT) was introduced into the ex-tended H∞ filtering, so as to avoid the calculation on linearized Jacobian matrix and enhance the filtering effect. Under the constraint condition of iterative judgment criterion, the state vector and the covariance matrix were esti-mated by repeatedly utilizing the observation information to make them more tend to the true value. The simulation results show that, under different observation precisions, this algorithm has the following merits: better stability, convergence speed and positioning precision.