CAJ | 학술논문

针对雷达系统的非线性目标跟踪存在精度低、滤波易发散等问题，提出一种新的非线性滤波算法———随机球面径向积分滤波算法(SSIF)。该滤波算法基于随机积分准则，利用随机球面积分准则和随机径向积分准则来近似函数积分。所提的滤波算法和传统的非线性滤波算法，例如扩展卡尔曼滤波、不敏卡尔曼滤波和容积卡尔曼滤波等相比在计算复杂度相当的情况下，不仅可以消除系统误差具有更高的跟踪精度，而且可以防止滤波发散提高滤波稳定度。通过蒙特卡洛仿真实验表明，所提出的非线性滤波算法整体性能明显优于传统的滤波算法。
침대뢰체계통적비선성목표근종존재정도저、려파역발산등문제，제출일충신적비선성려파산법———수궤구면경향적분려파산법(SSIF)。해려파산법기우수궤적분준칙，이용수궤구면적분준칙화수궤경향적분준칙래근사함수적분。소제적려파산법화전통적비선성려파산법，례여확전잡이만려파、불민잡이만려파화용적잡이만려파등상비재계산복잡도상당적정황하，불부가이소제계통오차구유경고적근종정도，이차가이방지려파발산제고려파은정도。통과몽특잡락방진실험표명，소제출적비선성려파산법정체성능명현우우전통적려파산법。
To solve the problem of low filter accuracy and easy getting divergence during target tracking in nonlinear system,a new nonlinear filter algorithm—stochastic spherical radial integration filter(SSIF)is proposed.The method uses stochastic radial rules and stochastic spherical integration rules to approximate the integral.Comparing with the traditional nonlinear filter,such as EKF,UKF and CKF,the new nonlinear filter can not only eliminate the systematic error with higher tracking accuracy,but also improve the stability of the filter,with the similar computational complexity.The Mont Carlo simulation shows that the method outperforms the traditional algorithms.