保持粒子多样性的非退化粒子滤波方法研究  

Non-Degeneracy Particle Filtering Method Research for Particle Diversity Preserving

作　　者：吴昊[1,2] 孙晓燕[1] 郭玉堂[2] 刘路路[2] 沈晶[2] 

机构地区：[1]中国矿业大学信息与电气工程学院,江苏徐州221008 [2]合肥师范学院计算机学院,安徽合肥230601

出　　处：《电子学报》2016年第7期1734-1741,共8页Acta Electronica Sinica

基　　金：国家自然科学基金(No.61301062;No.61503116;No.61375067);中央高校基本科研业务经费(No.2012QNA58);安徽省高校自然基金(No.KJ2013A217);安徽省质量工程提升计划(No.2013zy058);安徽省教育厅高等学校省级优秀青年人才基金(No.SQRL129ZD)

摘　　要：针对现有粒子滤波算法中的粒子退化问题以及重采样所引起的粒子多样性减弱问题,将自适应遗传算法与粒子滤波结合设计一种新的非退化粒子滤波算法.该算法通过对粒子使用遗传算子操作以保证粒子的多样性和有效性,根据粒子在前一时刻计算出来的先验信息自适应地实时调节当前时刻的遗传操作概率,有效增加了粒子对系统状态变化的适应性.实验结果表明,该算法可有效提高非线性系统状态的估计精度,尤其在系统状态发生突变的时候,可以得到较好的估计精度.To cope with the degeneracy in the existing particle filter algorithm and the diversity weakening caused by re-sampling,a new non-degeneracy algorithm is proposed in this paper by incorporating adaptive genetic algorithm into parti-cle filter.By using genetic operators to generate new particles,the algorithm can adjust the current probability of genetic ma-nipulation adaptively based on the previously calculated information so that the diversity and effectiveness of the particle can be ensured.It effectively improves the adaptability of particle to the changes of the system state.Experimental results show that this algorithm can effectively improve the estimation accuracy of the nonlinear system state.In particular,the algorithm can guarantee good estimation accuracy when the system state changes abruptly.

关 键 词：粒子滤波 遗传算法 粒子退化 自适应 粒子多样性 

分 类 号：TP18[自动化与计算机技术—控制理论与控制工程]