基于弹性机制的萤火虫优化粒子滤波算法
Firefly optimized particle filter algorithm based on spring mechanism
田梦楚 1柳林燕 1陈志敏 2方昱斌1
作者信息
- 1. 南京理工大学智能制造学院,南京 210094
- 2. 中国卫星海上测控部,江苏江阴 214431
- 折叠
摘要
针对标准粒子滤波重采样导致的粒子贫化问题,提出一种基于弹性机制的萤火虫优化粒子滤波算法.首先,利用萤火虫算法的吸引和移动机制,设计最优粒子引导粒子群体朝高似然区域移动的粒子运动控制策略;然后,评估粒子实时分布情况,根据每次迭代的高似然区域粒子占比值自适应控制粒子的优化强度;最后,检测最优粒子周围的粒子密度,引入弹簧的弹性机制,根据粒子密集度对判断区域内的粒子进行位置调整,使得粒子分布更加合理,提高粒子滤波的精度.实验结果表明,在粒子数目较少的情况下,改进算法滤波精度较标准粒子滤波提高12%∼25%;在同等滤波精度需求下,改进算法的运算时间比标准粒子滤波的运算时间减少20%∼30%,改进算法的综合性能更优.
Abstract
Aiming at the problem of particle impoverishment caused by the resampling process in standard particle filter,we propose a firefly optimized particle filter algorithm based on the spring mechanism.Firstly,adopting the attraction and movement mechanism of the firefly algorithm,we design the motion controlled strategy where the optimal particle is used to guide the particles moving towards the high likelihood region.Secondly,the real-time distribution of particles is evaluated,and the optimization intensity of particles is adaptive controlled by the proportion of particles in the high likelihood region.Finally,the density of the particles around the optimal particle is detected,and the elastic mechanism of spring is introduced,and the positions of particles are adjusted according to the particle density around the optimal particle,which can make the distribution of particles more reasonable and enhance the precision of the particle filter.The experimental results show that the filtering accuracy of the improved algorithm is 12%to 25%higher than that of the standard particle filter when the number of particles is small,and the operation time of the improved algorithm is about 20%to 30%less than that of the standard particle filter under the same filtering accuracy requirements,and the improved algorithm has the better comprehensive performance.
关键词
粒子滤波/萤火虫算法/粒子贫化/弹性机制/智能优化Key words
particle filter/firefly algorithm/particle impoverishment/spring mechanism/intelligent optimization引用本文复制引用
出版年
2024
NETL
NSTL
万方数据