CAJ | 학술논문

在当前的粒子滤波中，粒子可能出现退化现象和重采样，导致样本枯竭从而破坏粒子多样性。针对该问题，借鉴知识板和协同进化理论，提出一种基于知识板的协同粒子滤波算法。该算法对重要性密度函数进行采样，形成采样粒子样本，并将粒子划分为若干个子采样粒子群，对每个子采样粒子群在不同的区域进行搜索，通过子采样粒子群之间的通信，最终找到动态系统的最佳状态估计。理论分析与仿真结果表明，该算法能提高经典粒子滤波算法的群体多样性，在加快收敛速度和降低计算复杂度方面有较大优势。
재당전적입자려파중，입자가능출현퇴화현상화중채양，도치양본고갈종이파배입자다양성。침대해문제，차감지식판화협동진화이론，제출일충기우지식판적협동입자려파산법。해산법대중요성밀도함수진행채양，형성채양입자양본，병장입자화분위약간개자채양입자군，대매개자채양입자군재불동적구역진행수색，통과자채양입자군지간적통신，최종조도동태계통적최가상태고계。이론분석여방진결과표명，해산법능제고경전입자려파산법적군체다양성，재가쾌수렴속도화강저계산복잡도방면유교대우세。
For the degeneracy phenomenon of particles caused by evoluting and the impoverishment problem of particles caused by resampling. This paper proposes a novel particle filtering algorithm based on knowledge plate and coevolution. The main idea of this algorithm is to sample from the importance density function and generate particle samples which are divided into several sub-sample groups. Each sub-sample group searches among different area and finds the optimal state estimation of this dynamical system by means of the communication between each other. Theoretical analysis and experimental simulation results show that the proposed algorithm improves population diversity and has potential advantages in convergence rate and computational complexity, thus enhances the searching performance of the algorithm.