变异率自适应的差分进化粒子滤波算法研究
|
摘要
利用差分进化思想解决粒子滤波算法中存在的粒子退化具有一定的可行性,针对差分进化变异率固定,导致经过若干次迭代之后,后代样本间差异性变小,样本多样性减少,引入变异率自适应的差分进化算法,自适应变异过程中大权值粒子变化不大而小权值粒子变化较大,小权值粒子通过变异寻得较优状态,优化了粒子采样集合,使粒子集合向接近后验概率密度分布的取值运动,缓解了粒子滤波算法中的粒子退化缺陷。仿真结果表明,变异率自适应的差分进化粒子滤波算法较差分进化改进的粒子滤波,因此能够以较少的粒子数达到较高的准确度,由此算法运行时间较短,具有较好的实时性。
To solve the particle degeneration of particle filter algorithm,bringing the differential evolution algorithm to particle filter algorithm. For the mutation rate of differential evolution is constant,the difference between the offspring is smaller and the diversity of the sample is reduced after several iterations,introducing the differential evolution algorithm of adaptive mutation rate,it makes the particle of high weight change little but the particle of low weight change greatly,low weight particles can get a better state by variation,which increases the diversity of particles and optimizes the process of particle sampling. It makes particles closely move to posterior probability density distribution and overcomes the particle degradation defects. The simulation results show that the differential evolution particle filter algorithm of adaptive mutation rate can achieve a higher accuracy with fewer particles,so it runs for a short time and it has a real-time performance,it has a higher accuracy in target tracking.
To solve the particle degeneration of particle filter algorithm,bringing the differential evolution algorithm to particle filter algorithm. For the mutation rate of differential evolution is constant,the difference between the offspring is smaller and the diversity of the sample is reduced after several iterations,introducing the differential evolution algorithm of adaptive mutation rate,it makes the particle of high weight change little but the particle of low weight change greatly,low weight particles can get a better state by variation,which increases the diversity of particles and optimizes the process of particle sampling. It makes particles closely move to posterior probability density distribution and overcomes the particle degradation defects. The simulation results show that the differential evolution particle filter algorithm of adaptive mutation rate can achieve a higher accuracy with fewer particles,so it runs for a short time and it has a real-time performance,it has a higher accuracy in target tracking.
引文
[1]YU L,WEI C,JIAJ,etal.Compressive sensing for cluster structured sparse signals:variational Bayes approach[J].LET Signal Process,2016,10(7):770-779.
[2]HONG J,XU L,ZHAN K,etal.Joint tracking and classification based on aerodynamic model and radar cross section[J].Elsevier Ltd:Parttern Recognition,2014,47(9):3096-3105.
[3]XIE X,LI H P,HU F P,etal.An improved tracking algorithm of floc based on particle filter[J].International Journal of Digital Content Technology and Its Applications,2013,7(8):84-91.
[4]WALIA G S,KAPOOR R.Intelligent video target tracking using an evolutionary particle filter based upon improved cuckoo search[J].Expert System with Applications,2014,41(14):6315-6326.
[5]SERRATOSA F,ALQUEZAR R,AMEZQUITA N.A probabilistic integrated object recognition and tracking framework[J].Expert Systems with Applications,2012,39(8):7302-7318.
[6]马兆南,裴腾达,张浩.一种改进的粒子滤波目标跟踪算法[J].辽宁工程技术大学学报(自然科学版),2016,35(9):978-982.MA ZH N,PEI T D,ZHANG H.An improved target tracking algorithm combined with particle filter[J].Journal of Liaoning Technical University(Natural Science),2016,35(9):978-982.
[7]杜正聪,冯大海,牛高远.粒子群优化人工免疫粒子滤波器[J].四川大学学报(工程科学版),2013,45(1):146-151.DU ZH C,FENG D H,NIU G Y.Particle swarm optimization artificial immune particle filter[J].Journal of Sichuan University(Engineering Science Edition),2013,45(1):146-151.
[8]MALARVEZHI P,KUMAR R.Particle filter with novel resampling algorithm:A diversity enhanced particle filter[J].Wireless Personal Communications,2015,84(4):3171-3177.
[9]余春超,杨智雄,夏宗泽,等.采用GPU并行架构的基于互信息和粒子群算法的异源图像配准[J].红外技术,2016,38(11):938-946.YV CH CH,YANG ZH X,XIA Z Z.Heterogeneous images registration based on mutual information and particle swarm optimization algorithm using GPU parallel architecture[J].Infrared Technology,2016,38(11):938-946.
[10]HAN H,DING Y S,HAN K G.A new immune particle filter algorithm for tracking a moving target[C].Sixth International Conference on Natural Computation.Piscataway,IEEE,2010:3248-3252.
[11]PENG Z Y,LIU D R,JIN N,etal.Neural network strategy for sampling of particle filters on the tracking problem[C].International Joint Conference on Neural Networks.Piscataway,IEEE,2007:254-259.
[12]ZHOU Y F,BENOIS-PINEAU J,NICOLAS H.Multiresolution tracking of a non-rigid target with particle filters for low and variable frame-rate videos[C].International Workshop on Image Analysis for Multimedia Interactive Services,IEEE,2009:109-112.
[13]SARKER R A.Differential evolution with dynamic parameters selection for optimization problems[J].IEEE Transaction on Evolution Compution,2014,18(5):689-707.
[14]王亚楠,陈杰,甘明刚.基于差分进化的改进粒子滤波目标跟踪算法[C].第30届中国控制会议,2011.WANG Y N,CHEN J,GAN M G.Improved differential evolution-based particle filter algorithm for target tracking[C].Proceedings of the 30th Chinese Control Conference,2011.
[15]王龙,夏厚培.改进粒子滤波算法及其在目标跟踪中的应用[J].信息技术,2013,37(2):121-123.WANG L,XIA H P.Improved particle filter algorithm and its application in the target tracking[J].Information Technology,2013,37(2):121-123.
[2]HONG J,XU L,ZHAN K,etal.Joint tracking and classification based on aerodynamic model and radar cross section[J].Elsevier Ltd:Parttern Recognition,2014,47(9):3096-3105.
[3]XIE X,LI H P,HU F P,etal.An improved tracking algorithm of floc based on particle filter[J].International Journal of Digital Content Technology and Its Applications,2013,7(8):84-91.
[4]WALIA G S,KAPOOR R.Intelligent video target tracking using an evolutionary particle filter based upon improved cuckoo search[J].Expert System with Applications,2014,41(14):6315-6326.
[5]SERRATOSA F,ALQUEZAR R,AMEZQUITA N.A probabilistic integrated object recognition and tracking framework[J].Expert Systems with Applications,2012,39(8):7302-7318.
[6]马兆南,裴腾达,张浩.一种改进的粒子滤波目标跟踪算法[J].辽宁工程技术大学学报(自然科学版),2016,35(9):978-982.MA ZH N,PEI T D,ZHANG H.An improved target tracking algorithm combined with particle filter[J].Journal of Liaoning Technical University(Natural Science),2016,35(9):978-982.
[7]杜正聪,冯大海,牛高远.粒子群优化人工免疫粒子滤波器[J].四川大学学报(工程科学版),2013,45(1):146-151.DU ZH C,FENG D H,NIU G Y.Particle swarm optimization artificial immune particle filter[J].Journal of Sichuan University(Engineering Science Edition),2013,45(1):146-151.
[8]MALARVEZHI P,KUMAR R.Particle filter with novel resampling algorithm:A diversity enhanced particle filter[J].Wireless Personal Communications,2015,84(4):3171-3177.
[9]余春超,杨智雄,夏宗泽,等.采用GPU并行架构的基于互信息和粒子群算法的异源图像配准[J].红外技术,2016,38(11):938-946.YV CH CH,YANG ZH X,XIA Z Z.Heterogeneous images registration based on mutual information and particle swarm optimization algorithm using GPU parallel architecture[J].Infrared Technology,2016,38(11):938-946.
[10]HAN H,DING Y S,HAN K G.A new immune particle filter algorithm for tracking a moving target[C].Sixth International Conference on Natural Computation.Piscataway,IEEE,2010:3248-3252.
[11]PENG Z Y,LIU D R,JIN N,etal.Neural network strategy for sampling of particle filters on the tracking problem[C].International Joint Conference on Neural Networks.Piscataway,IEEE,2007:254-259.
[12]ZHOU Y F,BENOIS-PINEAU J,NICOLAS H.Multiresolution tracking of a non-rigid target with particle filters for low and variable frame-rate videos[C].International Workshop on Image Analysis for Multimedia Interactive Services,IEEE,2009:109-112.
[13]SARKER R A.Differential evolution with dynamic parameters selection for optimization problems[J].IEEE Transaction on Evolution Compution,2014,18(5):689-707.
[14]王亚楠,陈杰,甘明刚.基于差分进化的改进粒子滤波目标跟踪算法[C].第30届中国控制会议,2011.WANG Y N,CHEN J,GAN M G.Improved differential evolution-based particle filter algorithm for target tracking[C].Proceedings of the 30th Chinese Control Conference,2011.
[15]王龙,夏厚培.改进粒子滤波算法及其在目标跟踪中的应用[J].信息技术,2013,37(2):121-123.WANG L,XIA H P.Improved particle filter algorithm and its application in the target tracking[J].Information Technology,2013,37(2):121-123.