Abstract In order to solve the problem where the actual pavement brightness in highway tunnel does not automati? cally adapt to the traffic flow in dynamic traffic conditions and the traffic safety problem and lighting energy consumption problem arising from that problem, a dynamic lighting control method for middle section of tunnel has been devised based on traffic flow prediction. First, the traffic flow data measured at the Chatiaoling Tunnel in a time period is obtained and then preprocessed to generate the data sample required for training of the prediction model. Second, given the time sequence of the data sample, the traffic flow prediction model is devised based on the long short-term memory (LSTM) neural network, and the hyper-parameters of the LSTM network are optimized by using the grey wolf algorithm, so as to enhance the prediction accuracy of the network. Third, the parametric analysis of the key factors such as driving speed and traffic flow in the middle section of tunnel indicates that there is a strong correlation between traffic flow and pavement brightness. Based on that correlation, the lamp brightness regulation model is created. By using the LSTM traffic flow prediction model, advanced regulation inside the tunnel is achieved, so that the internal lighting can respond to the traffic conditions in a timely and effective manner. At last,tests are conducted to verify the effectiveness of the traffic flow prediction model and brightness regulation method that have been devised. As the results indicate, the method devised herein will guarantee adequate lighting in tunnel and also help with energy conservation.
Abstract：
In order to solve the problem where the actual pavement brightness in highway tunnel does not automati? cally adapt to the traffic flow in dynamic traffic conditions and the traffic safety problem and lighting energy consumption problem arising from that problem, a dynamic lighting control method for middle section of tunnel has been devised based on traffic flow prediction. First, the traffic flow data measured at the Chatiaoling Tunnel in a time period is obtained and then preprocessed to generate the data sample required for training of the prediction model. Second, given the time sequence of the data sample, the traffic flow prediction model is devised based on the long short-term memory (LSTM) neural network, and the hyper-parameters of the LSTM network are optimized by using the grey wolf algorithm, so as to enhance the prediction accuracy of the network. Third, the parametric analysis of the key factors such as driving speed and traffic flow in the middle section of tunnel indicates that there is a strong correlation between traffic flow and pavement brightness. Based on that correlation, the lamp brightness regulation model is created. By using the LSTM traffic flow prediction model, advanced regulation inside the tunnel is achieved, so that the internal lighting can respond to the traffic conditions in a timely and effective manner. At last,tests are conducted to verify the effectiveness of the traffic flow prediction model and brightness regulation method that have been devised. As the results indicate, the method devised herein will guarantee adequate lighting in tunnel and also help with energy conservation.
DUAN Zhongxing DU Wanxin
.Study on Tunnel Traffic Flow Prediction and Lighting Regulation Based on GWO-LSTM Model[J] MODERN TUNNELLING TECHNOLOGY, 2024,V61(3): 157-165
2024, Vol. 61 