摘要
在城市道路施工场景下应用短时交通量预测对提高施工区域交通效率及安全水平至关重要。考虑到施工场景下短时交通量历史样本量小且样本呈现非线性的特点,引入灰色预测模型,构建施工场景下的灰色小波神经网络短时交通量预测模型。以行宫西大街由西向东断面的交通量数据为例,分别基于小波神经网络短时交通量预测模型、灰色小波神经网络短时交通量预测模型,利用Matlab进行训练。结果显示,灰色小波神经网络短时交通量预测结果的平均绝对误差、平均相对误差和均方误差相较于小波神经网络短时交通量预测模型,分别降低了74.14%、75.21%和92.70%,该模型对城市道路施工场景下的短时交通量预测精确度更高。
The application of short-term traffic volume prediction in urban road construction scenarios is very important to improve the traffic efficiency and safety level of construction area.Considering the small historical sample size and the nonlinear characteristics of the sample,this paper introduces the grey prediction model to build the short-term traffic volume prediction model of grey wavelet neural network under the construction scenarios.Taking the traffic volume data of the west to east section of Xinggong West Street as an example,the short-term traffic volume prediction model of wavelet neural network and the short-term traffic volume prediction model of grey wavelet neural network were respectively trained based on Matlab.The experimental results show that compared with the wavelet neural network short-term traffic volume prediction model,the mean absolute error,mean relative error and mean square error of the short-term traffic volume prediction result of the grey wavelet neural network are reduced by 74.14%,75.21%and 92.70%,respectively.The short-term traffic volume prediction model of grey wavelet neural network is more accurate for the prediction of short-term traffic volume under the urban road construction scenarios.
作者
孙瑶
李挥剑
钱哨
Sun Yao;Li Huijian;Qian Shao(Transport Management Institute Ministry of Transport of the P.R.,Beijing 101601,China)
出处
《青海交通科技》
2023年第1期25-30,共6页
Qinghai Transportation Science and Technology
关键词
城市道路
施工场景
短时交通量预测
灰色小波神经网络预测模型
urban road
construction scene
short-term traffic volume prediction
grey wavelet neural network prediction model