Traffic Volume Forecasting for Highway Toll Stations Based on the Grey Forecasting GM (1,1) Model

高速公路收费站通行量预测是优化收费站管理、提高通行效率的关键技术。传统时间序列预测方法(如ARIMA、指数平滑等)通常需要大量历史数据，且难以适应短时交通流的非线性特征。本研究采用灰色预测GM (1, 1)模型，构建了一种适用于小样本数据的收费站通行量预测方法，并结合排队论和资源配置优化模型，为收费站闸口设计、人工收费窗口配置及机械臂自动化设备部署提供决策支持。实验数据来源于山东省某高速路收费站6期出闸通行量记录，结果表明，GM (1, 1)模型预测精度较好，为高速公路动态车道分配策略和机械臂优化配置等智慧化管理提供了理论依据和工程实践参考。Forecasting traffic volume at highway toll stations is a critical technology for optimizing toll management and improving traffic efficiency. Traditional time-series forecasting methods (e.g., ARIMA, exponential smoothing) typically require substantial historical data and often fail to adequately capture the nonlinear characteristics of short-term traffic flow. This study employs the Grey Prediction GM (1, 1) model to develop a traffic volume forecasting method suitable for small-sample datasets. By integrating queuing theory and resource allocation optimization models, the proposed approach provides decision-making support for toll gate design, manual toll booth allocation, and automated robotic arm deployment. Experimental validation using six-phase exit traffic volume records from a Shandong Province highway toll station demonstrates satisfactory prediction accuracy of the GM (1, 1) model. The findings offer both theoretical foundations and practical engineering references for intelligent management strategies, including dynamic lane allocation and robotic arm configuration optimization in highway toll systems.