CAJ | 학술논문

针对单站无源定位精度低、收敛速度慢和滤波性能不稳定甚至不能工作的问题，提出了后向平滑平方根容积卡尔曼滤波算法。该算法采用 Q-R分解的形式，使用误差协方差的平方根进行递推运算，同时其通过后向平滑与量测更新为二次前向滤波提供更为精确的初始值，改善了算法的总体性能。仿真实验表明，该算法在满足实时性要求的基础上，提高了单站无源定位的精度、收敛速度以及稳定性。
침대단참무원정위정도저、수렴속도만화려파성능불은정심지불능공작적문제，제출료후향평활평방근용적잡이만려파산법。해산법채용 Q-R분해적형식，사용오차협방차적평방근진행체추운산，동시기통과후향평활여량측경신위이차전향려파제공경위정학적초시치，개선료산법적총체성능。방진실험표명，해산법재만족실시성요구적기출상，제고료단참무원정위적정도、수렴속도이급은정성。
In order to solve the problems of low location accuracy,low convergence speed and poor filtering sta-bility in the single observer passive location,a novel backward smoothing square-root cubature Kalman filter was presented.The covariance square-root matrix was adopted based on Q-R decomposition.Meanwhile,the more accurate initial value for the next filtering process through the backward smoothing method and the measurement update process.Simulation results indicated that the novel algorithm improved the single observer passive loca-tion performance of the location accuracy,convergence velocity and stability,while keeping the real-time charac-teristic.