Variable Speed Limit Control Strategy Based On Short-term Traffic Flow Prediction

Vehicle holdings and traffic demand increase rapidly recently due to the rapid development of economic.As the aorta of the urban road network,the urban expressway has direct impact on the traffic flow of its adjacent roads.Variable speed limit(VSL)do has positive effect on the improvement of performance of urban expressways.However,VSL is mainly based on measured data instead of forecast data recently.So,I propose a VSL control strategy based on predictive data in this paper.And two main aspects of VSL control strategy based on forecast data are studied.The necessity of short-term traffic flow prediction is analyzed in this paper.A short-term traffic flow prediction method,which based on dynamic weights,is also proposed.The toplevel prediction results are obtained by combining the results of the bottom-level prediction with dynamic weights in this model.The bottom-level prediction methods include: prediction methods based on multiple linear regression,based on discrete Fourier transform,and based on BP neural network.The weight operator parameters are calibrated in this paper by conducting series of experimental analysis.Numerical experiments show that the choice of weight operator affects the accuracy of the prediction result,but no matter which weight operator is selected,it has better prediction accuracy than a single prediction method,which is reflected in the maximum error,minimum error and error variance of the prediction result have been greatly reduced.An optimization model for VSL control,which can be applied to the main road of urban expressways,is constructed based on the basic principles of VSL.Taking the congestion dissipation index under the condition of no variable speed limit control as a comparison benchmark,the optimization effect of variable speed limit not based on forecast data and the optimization effect of variable speed limit based on forecast data are studied respectively.The VSL control method that is not based on predictive data has a good optimization effect on alleviating congestion in the buffer area of the case.The average queue length is reduced from297.05 meters to 129.24 meters.The optimization effect of variable speed limit control based on predictive data is more obvious,The average queue length has been reduced from 297.05 meters to 92.08 meters.Numerical experiments show that the VSL can alleviate the queuing situation in the upstream of the congested area significantly.The VSL based on short-term traffic flow prediction proposed in this paper has a better optimization performance.