基于神经网络的智能车辆导航路径识别模型  

Neural network based navigation path recognition model of intelligent vehicle

作　　者：陈诚[1] CHEN Cheng(Shanghai Jian Qiao University,Shanghai 201306,China)

机构地区：[1]上海建桥学院,上海201306

出　　处：《现代电子技术》2018年第11期124-128,共5页Modern Electronics Technique

摘　　要：传统基于PID的车辆导航路径识别模型,基于精准的数学模型实现智能车辆路径控制,在高速情况下具有较低的鲁棒性,智能控制性能差。因此,基于智能车辆运动学模型,设计基于神经网络智能车辆导航路径识别模型结构,通过神经网络对车辆行驶方向进行控制,实现对智能车辆路径导航的控制。将多层前馈型神经网络作为基础结构对T-S模糊系统进行模拟,通过多次训练对神经网络的权值实施调控,完成基于神经网络智能车辆导航路径识别模型的设计。对识别模型实施训练,降低外界的干扰,提高识别模型的控制精度,实现对智能车辆路径导航的控制。实验结果说明,设计的基于神经网络的智能车辆路径导航识别模型控制精度高且鲁棒性较强,智能控制效果佳。The traditional vehicle navigation path recognition model based on PID,used to realize the intelligent vehicle path control based on precise mathematical model,has low robustness and poor intelligent control performance while driving at a high speed. On the basis of kinematics model of intelligent vehicle,the structure of intelligent vehicle navigation path recognition model based on neural network was designed. The neural network is used to control the driving direction of the vehicle to realize the control of the intelligent vehicle path navigation. The multilayer feed-forward neural network is taken as the basic structure to simulate the T-S fuzzy system. The weight of the neural network is regulated after several trainings to design the neural network based navigation path recognition model of intelligent vehicle. The recognition model is trained to reduce the external interference,improve the control accuracy of the recognition model,and realize the control of the intelligent vehicle path navigation. The experimental results show that the designed intelligent vehicle path navigation recognition model based on neural network has high control accuracy,strong robustness,and perfect intelligent control effect.

关 键 词：神经网络 智能车辆 路径导航 模糊控制 识别模型 控制精度 鲁棒性 

分 类 号：TN96-34[电子电信—信号与信息处理] TP391.41[电子电信—信息与通信工程]