Research On Effective Mechanism Of Traffic Signs For Long Freeway Tunnel

In the past ten years,with the increasing of highway mileage in China,the number of long highway tunnels(in the range of 1km-3km)and extra-long tunnels(longer than 3km)has doubled or tripled.According to the traffic accident data,the long tunnel section often has more traffic accidents and higher severity,and has higher traffic safety risk than other sections.Relevant research results,as well as China’s actual road reconstruction experience fully show that the scientific setting of traffic signs can effectively improve the level of traffic safety in this kind of road.Based on the project of National Science and Technology Support Program in 2012 "Research on Traffic Sign Setting Technology of Mountain Expressway" and the project of standardization research in 2014 "Research on Night Visibility Standardization of Highway Traffic Signs",this paper focus on exploring the effectiveness mechanism of traffic signs on this section,calibrating of targeted traffic sign setting parameter threshold,constructing a scientific evaluation theory system,verifing and proposing optimized system technical solutions.Considering traffic safety requirements for long tunnel sections of highways,this research obtained field investigation materials,traffic accidents and traffic operation data,and established vehicle test,simulated driving,theoretical deduction and statistical prediction.In this paper,the characteristics of traffic accident aggregation in long tunnel sections,the disposal strategy of traffic signs and the setting of key parameters,the effect evaluation model and the optimal systematic setting scheme are studied.The research has been carried out in the following aspects:(1)Based on the analysis of the main factors affecting the traffic safety of tunnel sections,the traffic accident distribution characteristics of long tunnel sections are studied by analyzing the traffic accident data of 9 extra-long tunnels and 6 long tunnels for 3 years.In the aspect of spatial distribution law,based on relative length of tunnel section,the traffic accident distribution analysis method is proposed,which realizes the normalization of traffic accident spatial distribution of different length of tunnel,and provides the basis for theoretical research.Statistical analysis of traffic accident data shows that the traffic accident rate of entrance section,exit section and middle position of long tunnel and extra-long tunnel is relatively high.What’s more,The traffic accident rate at the exit of the super long tunnel is obviously higher than that of the long tunnel.The traffic accident time distribution of extra-long tunnel and long tunnel has significant difference,and the traffic accident rate of extra-long tunnel at night is higher.Improper operation and speeding are the main causes of traffic accidents on this section.Rear end collision is the main accident type.According to the results of vehicle test and traffic flow observation,there is a significant deceleration process at the entrance of the tunnel,and the driver’s heart rate is greatly affected by the environmental changes in the entrance and the radius of the horizontal curve.Vehicle speed in the middle of the tunnel gradually stabilized,after the exit gradually increased to the general road speed limit or running speed.(2)Based on normalization of spatial distribution of traffic accidents,this paper carries out spatial clustering analysis by using DBSCAN density clustering algorithm and K-means clustering algorithm.Machine learning and genetic algorithm are used to establish traffic accident accumulation area in tunnel sections.The results show that there is not only uniformity but also dissimilarity in the rule of accident accumulation of long tunnel and extra-long tunnel sections.Traffic accidents are clustered into four categories,namely,the entrance section,the exit section,the middle section of the tunnel and the middle and rear section of the tunnel.According to the aggregation degree,long tunnels are mainly concentrated at the entrance of the tunnel,while the extra-long tunnels are mainly concentrated at the exit.The previous methods mainly focus on statistical analysis of accident data.Using unsupervised machine learning algorithm,this paper determined number and range of traffic accident aggregation sections which are more accurate and comprehensive to express the characteristics of accident aggregation.(3)Based on the previous research on the behavioral characteristics of drivers from real vehicle experiments,this paper proposed the disposal plan of traffic sign system for the key disposal section.Considering of visual barriers space constraints on the setting of traffic signs in tunnel sections,the disposal plans were studied with system.This paper analyzed the variable deceleration operation process of traffic signs recognized by drivers,and obtained the deceleration distance model of the whole process of vehicle deceleration in long tunnel.The simulation results showed that compared with the previous uniform deceleration calculation method,the application of this model has achieved more close to the actual results.The visual obstacle distance of tunnel section is calibrated by vehicle test,and the threshold value of the distance between traffic signs and the tunnel entrance is determined.Furthermore,aiming at the problem that the space in the tunnel is limited and the conditions of setting the traffic signs are more severe,the variable weighted grey clustering algorithm is applied to establish the quantitative classification method of the guideline demands of the tourist attractions,and the selection conditions of the guideline signs are defined.(4)Based on the characteristics of tunnel dark environment,an evaluation model of traffic sign visibility level is proposed,in which the brightness is taken as the calculation parameter.Theoretical deduction and a large number of real vehicle experiments confirm the minimum recognition distance of traffic signs under dark environment.The threshold values of luminance threshold and recognition level of different traffic signs were obtained.The visual recognition evaluation of traffic signs in dark environment truly reflects the visual recognition effect of human eyes,and published the standard "Night Visual Recognition Evaluation of Highway Traffic Signs".(5)Finally,based on the guiding ideology of improving the effectiveness of traffic guidance facilities and reducing the incidence of traffic accidents,combining the research methods of ergonomics,cognitive psychology,traffic engineering,computer simulation and other fields,relying on theoretical analysis and driving simulation,a comprehensive driving comfort,tension,safety can be constructed.The utility evaluation index system of traffic signs for reliability effect is established.In the index system,the index of driving discomfort is added for the first time.This paper quantitatively analyzed the effect of traffic sign setting on each index parameter,and established a matter-element comprehensive evaluation model of traffic sign utility based on extension theory and entropy weight.According to the subjective and objective analysis,the optimal setting plan for traffic signs on long highway tunnel is put forward.