改进的北方苍鹰算法优化粒子滤波算法
An improved northern goshawk algorithm optimizing particle filter algorithm
李广军 1徐祥书1
作者信息
- 1. 江苏理工学院 汽车与交通工程学院,江苏 常州 213001
- 折叠
摘要
针对标准粒子滤波过程的权值退化和样本贫化问题,提出一种改进的北方苍鹰算法优化粒子滤波算法INGOPF(Improved Northern Goshawk Optimization for Particle Filter).首先,利用透镜成像学习策略增加种群多样性,在优化初始解的同时增加种群搜索范围,使算法尽可能搜索到潜在的最优解,增加算法的搜索能力.其次,将改进的北方苍鹰位置更新公式用于优化迭代更新,然后将最优最差学习策略与透镜成像学习策略结合,克服算法陷入局部最优和易早熟的情况,提高算法的收敛精度.最后,将INGOPF应用于锂电池的寿命预测.仿真结果表明:与标准粒子滤波以及粒子群算法优化的粒子滤波方法相比,INGOPF有效提升了粒子多样性、系统状态估计精度、滤波稳定性和实际运用能力.
Abstract
An improved northern hawk algorithm optimized particle filtering algorithm INGOPF(improved northern goshawk optimization for particle filtering)is proposed for the problem of weight degradation and sample depletion of the standard particle filtering process.Firstly,a lensing imaging learning strategy is used to increase the population diversity and increase the population search range while optimizing the initial solution,so that the algorithm searches for the poten-tial optimal solution as much as possible and increases the algorithm search capability.Secondly,the improved northern hawk position update formula is used to optimize the iterative update,and then the optimal worst-case learning strategy is combined with the lensing imaging learning strategy to overcome the situation that the algorithm falls into local opti-mum and is prone to premature aging,and to improve the convergence accuracy of the algorithm.Finally,INGOPF is ap-plied to the lifetime prediction of lithium batteries.Simulation results show that INGOPF effectively improves particle di-versity,system state estimation accuracy,filtering stability and practical application compared with standard particle filter-ing as well as particle swarm algorithm particle filtering methods.
关键词
粒子滤波/北方苍鹰算法/透镜成像学习策略/状态估计Key words
particle filtering/northern goshawk algorithm/lens imaging learning strategies/state estimation引用本文复制引用
出版年
2024
NETL
NSTL
万方数据