CAJ | 학술논문

针对粒子滤波在非线性目标跟踪中存在粒子退化的问题，提出一种迭代积分粒子滤波的目标跟踪算法。该算法从改进重要性函数的角度入手，在积分卡尔曼滤波的基础上，通过高斯牛顿迭代的方法进行量测更新，并对粒子集合中的粒子进行迭代积分卡尔曼滤波，使得构造的重要性函数更加贴近真实后验分布。仿真结果表明，与粒子滤波算法、积分粒子滤波算法相比，该算法在有效改善非线性目标跟踪中粒子退化的同时，提高了跟踪精度。
침대입자려파재비선성목표근종중존재입자퇴화적문제，제출일충질대적분입자려파적목표근종산법。해산법종개진중요성함수적각도입수，재적분잡이만려파적기출상，통과고사우돈질대적방법진행량측경신，병대입자집합중적입자진행질대적분잡이만려파，사득구조적중요성함수경가첩근진실후험분포。방진결과표명，여입자려파산법、적분입자려파산법상비，해산법재유효개선비선성목표근종중입자퇴화적동시，제고료근종정도。
In view of particle degradation problem of nonlinear target tracking, an iterative Quadrature particle filter ( IQPF) target tracking algorithm was proposed based on the improved importance function. The algorithm is the basis of the Quadrature Kalman filter ( QKF) , the Gauss-Newton iteration method is used to make iterative processing, which can take the measurement into account by means of the iterative Quadrature Kalman filter, to make the importance function be closer to the true posterior. The computer simulation results indicate that compared with the particle filter ( PF) algorithm and the Quadrature particle filter ( QPF) algorithm, this algorithm has better performance in accuracy and effectively solves the particle degradation problem in nonlinear tracking.