摘要
针对传统遗传算法粒子滤波容易因遗传操作参数恒定不变而陷入局部最优的问题,在遗传算法粒子滤波中引入自适应方法,提出自适应遗传算法粒子滤波。根据粒子适应度的大小,动态调节遗传操作的交叉、突变概率,从而在尽可能多地保留优势粒子的同时更有效地产生新的优势粒子,跳出局部最优。将自适应遗传算法粒子滤波应用于动态目标定位模型,并将其与遗传算法粒子滤波的性能进行比较。结果表明,自适应方法的引入可以增加算法有效粒子数,有效解决算法早熟问题,改善滤波精度,对于提高动态目标定位精度是有效的。
In order to solve the problem that traditional genetic algorithm particle filter is easy to fall into local optimization due to constant genetic operation parameters,an adaptive method is introduced into genetic algorithm particle filter,and an adap-tive genetic algorithm particle filter is proposed.The principle of dominant inheritance changes the probability of crossover and muta-tion with the change of particle fitness,so as to retain the dominant particles as much as possible while generating new dominant par-ticles more effectively,jumping out of local optimum.The adaptive genetic algorithm particle filter is applied to the established dy-namic state space model,compared with the performance of the genetic algorithm particle filter by simulation.The results show that the introduction of adaptive method can increase the effective particle number of the algorithm,effectively solve the problem of pre-mature algorithm and improve the filtering accuracy,which is very effective for improving the accuracy of dynamic target position-ing.
作者
蔡明
李国华
季茜
李培德
CAI Ming;LI Guohua;JI Qian;LI Peide(Hubei Meteorological Information and Technology Support Center,Wuhan 430074;Hubei Key Laboratory for Heavy Rain Monitoring and Warning Research,Wuhan 430074;Huanggang Meteorological Bureau,Huanggang 438000)
出处
《计算机与数字工程》
2024年第3期841-845,891,共6页
Computer & Digital Engineering
基金
湖北省气象局重点科研项目(编号:2022Z04)资助。
关键词
动态状态空间模型
自适应
目标定位
遗传算法
粒子滤波
dynamic state space model
adaptive
target positioning
genetic algorithm
particle filter
