基于RBF神经网络的车辆安全换道时机决策模型研究

Research on Modelling Vehicle Safety Lane-Changing Timing Decision Based on RBF Neural Network

为了帮助驾驶人正确决策车辆换道时机,使用了一种模式识别方法——RBF神经网络,建立了车辆换道时机决策模型。模型可以预测车辆换道的安全性,从而保证驾驶人和车辆的安全。对车辆换道时机决策的影响因素进行了分析,提出了11个现代传感器容易获取的影响参数,并作为RBF神经网络的输入变量。模型的学习和测试运用了大量的车辆行驶数据,实验结果显示:11个参数的RBF神经网络模型预测精度较高,可以达到87.9%,高于7个参数模型的81.8%;随着模型精度的不断提高,在驾驶主动安全系统和智能车辆无人驾驶系统中,本文模型也可以起到关键的作用。

In order to help drivers make correct decisions on choosing lane-changing timing,this paper uses a pattern recognition method,RBF neural network,to establish a lane-changing timing model.The model can predict the safety of lane changing,thus ensuring the security of drivers and vehicles.This paper analyzes the influencing factors of vehicle lane-changing timing decision,and presents eleven parameters that are easily obtained by modern sensors.These parameters are also used as input variables of RBF neural network.A large number of vehicle driving data are used in the study and test of the model.The experimental results show that the RBF neural network model has a high prediction accuracy,which can reach 87.9%.This result is higher than that of the 7-parameter model(81.8%).With the continuous improvement of the model accuracy,the model can also play a key role in driving the active safety system and the unmanned driving system of intelligent vehicle.