基于自适应遗传算法的粒子滤波器
发布时间:2018-09-12 17:11
【摘要】:针对重采样导致的权值退化问题,应用遗传算法的进化思想来优化重采样算法,将粒子权值作为适应度值,合理设定阈值,利用最佳个体保存法保存高适应度粒子,利用自适应交叉、变异操作对低适应度粒子进行进化,将高适应度粒子与进化粒子组合成新的粒子集进行状态估计。仿真实验表明,该算法具有良好的实时性和估计精度,其状态估计精度比标准粒子滤波提高近24倍,比无迹卡尔曼粒子滤波提高近4倍,耗时约为无迹卡尔曼粒子滤波的1/10。
[Abstract]:In order to solve the problem of weight degradation caused by resampling, the evolutionary idea of genetic algorithm is applied to optimize the resampling algorithm. The particle weight is taken as the fitness value, the threshold is set reasonably, and the best individual preservation method is used to preserve the high-fitness particle. Using adaptive crossover and mutation operation, the low fitness particles are evolved, and the high fitness particles and the evolutionary particles are combined to form a new particle set for state estimation. The simulation results show that the proposed algorithm has good real-time and estimation accuracy. Its state estimation accuracy is nearly 24 times higher than that of standard particle filter and nearly 4 times higher than that of unscented Kalman particle filter. The time consuming of this algorithm is about 1 / 10 of that of unscented Kalman particle filter.
【作者单位】: 攀枝花学院;
【基金】:四川省应用基础研究项目(2011JY0115)
【分类号】:TN713;TP18
本文编号:2239672
[Abstract]:In order to solve the problem of weight degradation caused by resampling, the evolutionary idea of genetic algorithm is applied to optimize the resampling algorithm. The particle weight is taken as the fitness value, the threshold is set reasonably, and the best individual preservation method is used to preserve the high-fitness particle. Using adaptive crossover and mutation operation, the low fitness particles are evolved, and the high fitness particles and the evolutionary particles are combined to form a new particle set for state estimation. The simulation results show that the proposed algorithm has good real-time and estimation accuracy. Its state estimation accuracy is nearly 24 times higher than that of standard particle filter and nearly 4 times higher than that of unscented Kalman particle filter. The time consuming of this algorithm is about 1 / 10 of that of unscented Kalman particle filter.
【作者单位】: 攀枝花学院;
【基金】:四川省应用基础研究项目(2011JY0115)
【分类号】:TN713;TP18
【相似文献】
相关期刊论文 前8条
1 张耀镭;王友仁;;快速实现数字仿生电路设计的自适应遗传算法[J];计算机测量与控制;2007年10期
2 官伯林;贾建援;朱应敏;;基于自适应遗传算法的三轴光电跟踪策略[J];仪器仪表学报;2012年08期
3 苏琳琳;张晓林;;利用自适应遗传算法的芯片功能验证自动测试[J];应用科学学报;2011年06期
4 金力;刘桥;;基于自适应遗传算法的运放的电路级综合[J];西华大学学报(自然科学版);2006年02期
5 兰海;汪宇涵;张利军;;基于改进自适应遗传算法的船舶电力系统滤波装置优化配置[J];船电技术;2009年06期
6 许川佩;陈征南;任智新;胡聪;;基于云自适应遗传算法的NoC路径分配研究[J];计算机测量与控制;2012年09期
7 许川佩;陈征南;任智新;;基于云自适应遗传算法的NoC映射研究[J];计算机工程与应用;2012年36期
8 赵曙光,刘贵喜,杨万海;利用自适应遗传算法实现模拟电路自动设计[J];西安电子科技大学学报;2003年03期
相关硕士学位论文 前3条
1 柯家伟;面向订单快速交付的生产过程管控技术研究与系统实现[D];北京理工大学;2016年
2 陈殿夏;桥式网络系统可靠性分析和优化[D];沈阳工业大学;2005年
3 金力;基于改进自适应遗传算法的CMOS运放的电路级综合方法的研究[D];贵州大学;2006年
,本文编号:2239672
本文链接:https://www.wllwen.com/kejilunwen/dianzigongchenglunwen/2239672.html