本文選題：無跡粒子濾波　切入點：高斯粒子群優(yōu)化　出處：《測繪通報》2017年04期

【摘要】：針對粒子濾波算法中存在的粒子退化與粒子匱乏的缺陷,提出了利用高斯粒子群優(yōu)化無跡粒子濾波的新算法。算法使用無跡粒子濾波進行重要性采樣,并將高斯粒子群優(yōu)化算法融入重采樣過程中。該算法選取的概率密度更加接近系統(tǒng)真實狀態(tài),有效增加了粒子的多樣性,提高了抽樣效率,降低了粒子退化程度,緩解了粒子匱乏現(xiàn)象。試驗結(jié)果表明,該算法的濾波精度明顯優(yōu)于粒子濾波與無跡粒子濾波算法所得到的濾波精度。
[Abstract]:Aiming at the defects of particle degradation and particle scarcity in particle filter algorithm, a new unscented particle filter algorithm based on Gao Si particle swarm optimization is proposed.The algorithm uses unscented particle filter for importance sampling, and Gao Si particle swarm optimization algorithm is incorporated into the resampling process.The probability density selected by the algorithm is closer to the real state of the system, which effectively increases the diversity of particles, improves the sampling efficiency, reduces the degree of particle degradation and alleviates the phenomenon of particle scarcity.The experimental results show that the filtering accuracy of this algorithm is better than that of particle filter and unscented particle filter.
【作者單位】： 武漢大學衛(wèi)星導航定位技術研究中心;
【基金】：國家重點基礎研究發(fā)展計劃(2013CB733205) 武漢市科技局項目(2015011701011639)
【分類號】：P207

【相似文獻】

相關期刊論文 前2條

1 宮軼松;李保利;張錦軍;張靈敏;;基于MKLD準則的PSOPF算法設計及其應用[J];武漢大學學報(信息科學版);2011年10期

2 ;[J];;年期

，

本文編號：1710759