摘要
研究交通高峰车辆拥挤问题,通过准确预测车辆的平均行程速度,进而进行交通状况判别对于有效缓解交通拥挤,提高道路的使用效率。采用BP神经网络存在收敛速度慢、易陷入局部最优解等缺陷,预测效果差,难以应用于高速公路交通状况判别。为了克服BP神经网络这些缺陷,提出基于RBF神经网络的高速公路交通状况判别方法,RBF神经网络结构简单,收敛速度快,具有很强的非线性函数逼近能力,在高速公路交通状况判别中具有广泛的应用前景。从实验结果可以看出,RBF神经网络比BP神经网络有着更高平均行程速度预测精度,更适合于高速公路交通状况判别。
It is very significant to improve the service efficiency of roads by forecasting the average travel speed of vehicle exactly,and then discriminating traffic situations.As BP neural network has some inherent drawbacks,such as falling into local optimization and low convergence rate,in order to overcome the drawbacks of BP neural network,RBF neural network（RBFNN） is applied to discriminate traffic situations of highway.RBF neural network has simple structure,high convergence rate,and strong approximation ability of nonlinear function,which has widespread practical applications in discriminating traffic situation of highway.The experimental results show that RBF neural network has higher forecasting accuracy of average travel speed than BP neural network,and is more suitable to discriminate traffic situation.
出处
《计算机仿真》
CSCD
北大核心
2011年第2期350-353,共4页
Computer Simulation
