【作者】 高丙坤； 田丽； 任晶秋；

【Author】 Gao Bingkun Tian Li Ren Jingqiu (Department of automation and control engineering,Daqing petroleum Institute,Daqing 163318)

【机构】 大庆石油学院自控系；

【摘要】 针对非线性滤波中存在的"实时"问题,提出了用神经网络实现非线性滤波的解决方案。为了提高神经网络学习速度,利用遗传算法全局搜索能力强的特点,对神经网络的结构和权阵进行优化,构成了遗传神经网络,实现了全局最佳逼近。仿真结果验证了遗传神经网络实现最优非线性滤波的有效性,并说明了该方法精度高,收敛性好。更多还原

【Abstract】 Aimed at"real time"problem in nonlinear filtering,the scheme of realizing nonlinear fdtering based on neural networks is proposed.In order to enhance studying speed,the paper utilizes good ability of searching in all area,and optimizes structure and right array of neutral network,and constructs genetic and neutral network,and realizes optimum approach in all area. The simulation results verify the superiority for genetic and neutral network in optimum nonlinear filtering,and indicate this method has higher accuracy and better convengence.更多还原