Research On Intelligent Tunnel Lighting Control Scheme Combined With Traffic Flow Prediction

Energy conservation,as an important indicator of sustainable development,has become a global issue that has attracted much attention.Because of its special structure,the tunnel also needs to provide lighting during the day,which causes the energy consumption of the tunnel lighting system to remain high.Currently,there is a lack of effective intelligent tunnel lighting control technology.Most tunnels are still using traditional segmented tunnel lighting,and current tunnel lighting control methods,such as vehicle entrance light dimming and dynamic dimming,have their shortcomings.In response to the above problems,this article takes the tunnel lighting system as the research object,and conducts research for the purpose of ensuring tunnel traffic safety and reducing tunnel lighting energy consumption.The main work content and innovations of this article are as follows:Due to the randomness and inhomogeneity of the traffic flow data,the convergence speed of the stacked autoencoder becomes slower,and when the data set is small,the stacked autoencoder is also prone to overfitting.In response to this problem,this paper proposes to optimize it by using a greedy layer-by-layer training algorithm.This thesis uses this algorithm to predict traffic flow data at different sampling intervals,and compares the prediction results with the prediction results of Long Short-Term Memory(LSTM)and Gated Recurrent Unit(GRU)models.In contrast,the experimental results are analyzed on the Mean Absolute Error(MAE)and other predictive indicators,and finally come to the conclusion: using the greedy layer-by-layer training algorithm to optimize the stacked autoencoder can effectively improve the accuracy of traffic flow prediction.Secondly,this thesis takes lighting parameters as the optimization object,and aims at higher uniformity of tunnel illumination and human eye comfort.First,the tunnel model is established using DIALux lighting simulation software,and then three different lighting devices with the same power are selected.They are the high-pressure sodium lamp model ZQ108b-N100,the fluorescent lamp model ZQ109-T8 and the LED lamp model ZD901-3.The tunnel lighting system is modeled and simulated in DIALux software.Finally,the three lamps are compared in When the elevation angle is between,the longitudinal uniformity of the road surface and the horizontal uniformity of the road surface and other parameters,after analyzing the uniformity of the road surface,select the most suitable lighting parameters.At last,this thesis compares and analyzes the advantages and disadvantages of the tunnel dimming method combined with traffic flow prediction and the current tunnel dimming method.When the traffic flow is too high,the use of the car-in light dimming method will cause frequent dimming of the lighting equipment,resulting in higher switching losses,and affecting the life of the lighting equipment.When the traffic flow is too low,dynamic dimming to maintain a higher brightness level will result in higher energy consumption.Aiming at the shortcomings of the abovementioned vehicle entering light dimming method and dynamic dimming method,this thesis proposes to add traffic flow prediction to the tunnel lighting control as the basis for judging whether the traffic flow exceeds the traffic flow threshold,so as to determine whether to use dynamic dimming or car The timing of the light-in dimming.The advantage of traffic flow forecasting is that it can accurately predict different traffic flows according to different weather and holidays.However,the average value of historical data and manual forecasting cannot achieve such an effect.