神经网络自组织模糊控制器的设计与仿真  

The Design and Simulation Based on Neural Network Self-organizing Fuzzy Logic Controller

作　　者：杨涛[1] 杨洪森 陈梅[3] 

机构地区：[1]江苏大学计算机科学与通信工程学院,江苏镇江212013 [2]山东省武城县武城镇中学,山东武城253312 [3]苏州大学计算机科学与技术学院,江苏苏州215006

出　　处：《苏州大学学报（工科版）》2008年第1期73-77,共5页Journal of Soochow University Engineering Science Edition (Bimonthly)

摘　　要：研究和设计了神经网络自组织模糊控制器,用于实现车辆自动驾驶系统。该控制器基于MATLAB软件提供的自适应神经模糊推理系统,通过网络对道路位置坐标的离线学习,获取模糊控制规则库。训练好的控制器,在对被控车辆模拟控制的过程中,在线学习和调整控制参数,优化控制器的性能,以适应在不同道路轨迹变化下控制车辆自动转向的目的。最后通过MATLAB软件的仿真试验,验证该控制器对于被控车辆的有效性。The paper studied and designed a neural network self-organizing fuzzy logic controller for vehicle auto driving system. Based on adaptive neuro-fuzzy inference system provided by the MATLAB,the controller used radial basis function neural network studying the road track data sample and gained the rules library, of fuzzy control arithmetic. When the trained controller was working in field, it could online learn and adjust the parameters to optimize the action performance for fitting different road track conditions. Finally, the MATLAB simulation results shows, the neural network self-organizing fuzzy logic controller is effective for our vehicle system.

关 键 词：径向基神经网络 自组织模糊控制器 车辆自动驾驶系统 

分 类 号：TP18[自动化与计算机技术—控制理论与控制工程]