变频变异粒子滤波算法
|
摘要
针对粒子滤波算法中的粒子退化及重采样所引起的粒子多样性减弱问题,将自适应变频策略应用于免疫理论的变异操作中,并与粒子滤波相结合设计了一种新的变频变异粒子滤波算法.算法引入自适应变频算子实时调节当前时刻的变异频率,控制了变异粒子的数量;再采取不同策略对粒子进行变异操作,以提高粒子对系统状态变化的适应性;最后,对新产生粒子进行权值计算,选择权值较大粒子构成新粒子集,以提高滤波精度.研究结果表明,该方法能够用更少的粒子完成高精度的估计任务,具有更高的滤波精度、粒子多样性、运算速度综合性价比.同时粒子分布更合理,在高似然区外仍然存在一定数量的粒子,为系统发生突变时保持较好的估计精度提供了条件.
Aimed at the problem of particle diversity reduction in particle filtration algorithm induced by particle degeneracy and resampling, a strategy of adaptive frequency conversion was employed in mutative operation of immune theory and, cooperated with particle filtration, a novel particle filtration algorithm with variable-frequency mutation was designed. An adaptive variable-frequency operator was introduced into the algorithm to regulate the present mutation frequency in the real time and control the quantity of the mutational particles. Further, various strategies were adopted to carry out the mutational operation for the particles and improve the adaptability of the particles to the change of the system state. Finally, the weight value calculation of the regenerated particles was conducted and a new particle set was composed from the particles with greater weight value to improve filtration accuracy. It was shown by investigation result that this approach could fulfill the high-accurate estimation task with less quantity of particles and would have higher filtration precision, particle variety, and comprehensive ratio of credit to cost of calculation speed. Meantime, the particle distribution would be more rational and a certain quantity of particles would still exist outside of the high likelihood region, so that providing a condition for keeping a better estimation accuracy in case of system mutation.
Aimed at the problem of particle diversity reduction in particle filtration algorithm induced by particle degeneracy and resampling, a strategy of adaptive frequency conversion was employed in mutative operation of immune theory and, cooperated with particle filtration, a novel particle filtration algorithm with variable-frequency mutation was designed. An adaptive variable-frequency operator was introduced into the algorithm to regulate the present mutation frequency in the real time and control the quantity of the mutational particles. Further, various strategies were adopted to carry out the mutational operation for the particles and improve the adaptability of the particles to the change of the system state. Finally, the weight value calculation of the regenerated particles was conducted and a new particle set was composed from the particles with greater weight value to improve filtration accuracy. It was shown by investigation result that this approach could fulfill the high-accurate estimation task with less quantity of particles and would have higher filtration precision, particle variety, and comprehensive ratio of credit to cost of calculation speed. Meantime, the particle distribution would be more rational and a certain quantity of particles would still exist outside of the high likelihood region, so that providing a condition for keeping a better estimation accuracy in case of system mutation.
引文
[1] 杨小军,潘泉,王睿,等.粒子滤波进展与展望 [J].控制理论与应用,2006,23(2):261-267.
[2] 程水英,张剑云.粒子滤波评述 [J].宇航学报,2008,29(4):1109-1111.
[3] 王法胜,鲁明羽,赵清杰,等.粒子滤波算法 [J].计算机学报,2014,37(8):1679-1694.
[4] LI Tiancheng,SUN Shudong,SATTAR T P,et al.Fight sample degeneracy and impoverishment in particle filters:A review of intelligent approaches [J].Expert Systems with Applications,2014,41(8):3944-3954.
[5] ARULAMPALAM M S,MASKELL S,GORDON N,et al.A tutorial on particle filters for online nonlinear/non-Gaussian Bayesian tracking [J].IEEE Transactions on Signal Processing,2002,50(2):174-188.
[6] DU M,NAN X M,GUAN L.Monocular human motion tracking by using DE-MC particle filter [J].IEEE Transactions on Image Processing,2013,22(10):3852-3865.
[7] 吴昊,孙晓燕,郭玉堂,等.保持粒子多样性的非退化粒子滤波方法研究 [J].电子学报,2016,44(7):1734-1741.
[8] 张琪,乔玉坤,孔祥玉,等.随机摄动强跟踪粒子滤波算法 [J].物理学报,2014,63(11):102-110.
[9] 王晓,韩崇昭.基于混合采样的多模型机动目标跟踪算法 [J].自动化学报,2013,39(7):1152-1156.
[10] 周卫东,孙天,储敏,等.交互式多模型粒子滤波优化重采样算法 [J].北京航空航天大学学报,2017,43(5):865-871.
[11] KONG A,LIU J S,WONG W H.Sequential imputations and Bayesian missing data problems [J].Journal of the American Statistical Association,1994,89(425):278-288.
[12] 危璋,冯新喜,毛少锋.裂变优选重采样粒子滤波算法 [J].空军工程大学学报(自然科学版),2014,15(6):82-86.
[13] AXEL G,DINHOUN L.The chopthin algorithm for resampling [J].IEEE Transactions on Signal Processing,2016,64(16):4273-4281.
[14] MALARVEZHI P,KUMAR R.Particle filter with novel resampling algorithm:A diversity enhanced particle filter [J].Wireless Personal Communications:An International Journal,2015,84(4):3171-3177.
[15] 田梦楚,薄煜明,陈志敏,等.萤火虫算法智能优化粒子滤波 [J].自动化学报,2016,42(1):89-97.
[16] 王进花,曹洁,李伟.智能优化的代价评估粒子滤波算法 [J].系统工程与电子技术,2017,39(12):2857-2862.
[17] GORDON N,SALMOND D,SMITH A F M.Novel approaches to non-linear/non-gaussian bayesian state estimation [J].Radar & Signal Processing IEE Proceedings F,1993,140(2):107-113.
[18] SPEEKENBRINK M.A tutorial on particle filters [J].Journal of Mathematical Psychology,2016,73:140-152.
[19] CASTRO L N,TIMMIS J.Artificial immune systems:A new computational intelligence approach [M].New York:Springer-Verlag,2002.
[20] ULUTAS B H,KULTUREL-KONAK S.A review of clonal selection algorithm and its applications [J].Artificial Intelligence Review,2011,36(2):117-138.
[21] 陈志敏,田梦楚,吴盘龙,等.基于蝙蝠算法的粒子滤波法研究[J].物理学报,2017,66(5):41-50.
[22] 叶龙,王京玲,张勤.遗传重采样粒子滤波器 [J].自动化学报,2007,33(8):885-887.
[2] 程水英,张剑云.粒子滤波评述 [J].宇航学报,2008,29(4):1109-1111.
[3] 王法胜,鲁明羽,赵清杰,等.粒子滤波算法 [J].计算机学报,2014,37(8):1679-1694.
[4] LI Tiancheng,SUN Shudong,SATTAR T P,et al.Fight sample degeneracy and impoverishment in particle filters:A review of intelligent approaches [J].Expert Systems with Applications,2014,41(8):3944-3954.
[5] ARULAMPALAM M S,MASKELL S,GORDON N,et al.A tutorial on particle filters for online nonlinear/non-Gaussian Bayesian tracking [J].IEEE Transactions on Signal Processing,2002,50(2):174-188.
[6] DU M,NAN X M,GUAN L.Monocular human motion tracking by using DE-MC particle filter [J].IEEE Transactions on Image Processing,2013,22(10):3852-3865.
[7] 吴昊,孙晓燕,郭玉堂,等.保持粒子多样性的非退化粒子滤波方法研究 [J].电子学报,2016,44(7):1734-1741.
[8] 张琪,乔玉坤,孔祥玉,等.随机摄动强跟踪粒子滤波算法 [J].物理学报,2014,63(11):102-110.
[9] 王晓,韩崇昭.基于混合采样的多模型机动目标跟踪算法 [J].自动化学报,2013,39(7):1152-1156.
[10] 周卫东,孙天,储敏,等.交互式多模型粒子滤波优化重采样算法 [J].北京航空航天大学学报,2017,43(5):865-871.
[11] KONG A,LIU J S,WONG W H.Sequential imputations and Bayesian missing data problems [J].Journal of the American Statistical Association,1994,89(425):278-288.
[12] 危璋,冯新喜,毛少锋.裂变优选重采样粒子滤波算法 [J].空军工程大学学报(自然科学版),2014,15(6):82-86.
[13] AXEL G,DINHOUN L.The chopthin algorithm for resampling [J].IEEE Transactions on Signal Processing,2016,64(16):4273-4281.
[14] MALARVEZHI P,KUMAR R.Particle filter with novel resampling algorithm:A diversity enhanced particle filter [J].Wireless Personal Communications:An International Journal,2015,84(4):3171-3177.
[15] 田梦楚,薄煜明,陈志敏,等.萤火虫算法智能优化粒子滤波 [J].自动化学报,2016,42(1):89-97.
[16] 王进花,曹洁,李伟.智能优化的代价评估粒子滤波算法 [J].系统工程与电子技术,2017,39(12):2857-2862.
[17] GORDON N,SALMOND D,SMITH A F M.Novel approaches to non-linear/non-gaussian bayesian state estimation [J].Radar & Signal Processing IEE Proceedings F,1993,140(2):107-113.
[18] SPEEKENBRINK M.A tutorial on particle filters [J].Journal of Mathematical Psychology,2016,73:140-152.
[19] CASTRO L N,TIMMIS J.Artificial immune systems:A new computational intelligence approach [M].New York:Springer-Verlag,2002.
[20] ULUTAS B H,KULTUREL-KONAK S.A review of clonal selection algorithm and its applications [J].Artificial Intelligence Review,2011,36(2):117-138.
[21] 陈志敏,田梦楚,吴盘龙,等.基于蝙蝠算法的粒子滤波法研究[J].物理学报,2017,66(5):41-50.
[22] 叶龙,王京玲,张勤.遗传重采样粒子滤波器 [J].自动化学报,2007,33(8):885-887.