Research On Operational Evaluation Index Of Signal Intersections In Mountain Cities Based On Trajectory Data

The signalized intersection is a natural bottleneck of the urban road network.The oversaturated traffic demand,the unreasonable design of organisation,and the improper timing of signal phase will lead to congestion at the intersection,even to the urban arterial and regional road networks.In order to minimize or avoid the waste of space and time resources at the intersection,the evaluation of intersection operation is one of the indispensable technical links.In the past,China mainly used delays to evaluate the efficiency of intersection operations.The estimation method of which mainly relied on model method or traffic simulation.In some scenarios,it can not reflect the actual traffic operation status,and usually needs to collect a large number of traffic parameters such as traffic flow and capacity.At this stage,relying on multi-source traffic data for cross-port operation evaluation has attracted internal attention in the industry,and has some research results;but how to construct appropriate evaluation indicators and evaluation methods to better serve the overall optimization of signalized intersections,still further research is pending.Based on the multi-source vehicle trajectory data of taxi,bus and JIATU(internet of vehicles),this paper takes the single-point signalized intersection of mountainous city as the research object,and proposes new characterization parameters and extraction methods based on vehicle operating characteristics and laws;It has defined a two-level evaluation index system for signalized intersections,which realizes analysising and evaluating of the operation status of intersections by the multi-level and angles,from the turn,import,and intersection.The main research contents are as follows:(1)Reviewing and summary of exsiting research on indexes and methods of signalized intersection operation evaluation.The estimation methods of two core indicators of delay and queue length are comprehensively combed,which is difficult to promote and apply as those estimation models need to manually collecte and and input multiple traffic parameters,also the assumptions;besides,it has clarified and explained the shortcomings of the existing intersection evaluation criteria and evaluation methods,facing the actual needs of conduction,selection and optimization of signal control schemes at present,and the challenges also be proposed in the construction,extraction and application of evaluation indicators.(2)Evaluation index is constructed Considering the multi-source trajectory data feature based on the vehicle's operating characteristics and laws at the signalized intersection.The traffic characteristics such as random arrival,speed fluctuation,queuing accumulation and dissipation of the vehicle at the signal intersection are qualitatively analyzed.The characteristics of the space-time traffic parameters and the influence relationship under the guidance of traffic demand,intersection channelization and signal timing are analyzed.The analysis results show that the remaining queues in the cycle have a significant impact on the traffic efficiency of the intersection;Based on which,and considering continuous time,position,speed and other characteristics of the trajectory data,from the perspectives of vehicle travel time benefit,lane or approach space utilization,lane group or approach balance of supply and demand,The second level evaluation indexs are proposed,including the average travel time intersection,the 95~thh percent of queue length and the two parking rate.(3)Traffic parameter extraction and sample classification method construction of vehicle operation at intersection.Based on the multi-source trajectory data,the ArcGIS platform and Python scripting language,a complete vehicle initial queue length and intersection travel time batch extraction method are constructed.The vehicle parking number identification and sample classification method based on the continuous velocity variation characteristics of trajectory data are proposed.The results show that the method can accurately distinguish the import,steering,initial queue length,intersection travel time,and number of stops of individual vehicles,which lays a foundation for the evaluation of intersection operation.(4)Construction of comprehensive evaluation index model for signalized intersection operation.Based on the secondary evaluation index,the calculation model of comprehensive evaluation index for intersection operation is introduced.Considering the different value orientations of various social groups,it has comprehensively investigated the sensitivity and attention by traffic management departments,environmental protection departments and ordinary social travellers to indicators,and convinced index weight coefficients of the average travel time,the95~thh percent of queue length and the two parking.The Gaussian hybrid clustering model is used to classify the comprehensive evaluation index and service level,and finally establishes Five levels of evaluation criteria.(5)Example verification and analysis.The typical signalized intersections in Chongqing were selected and compared with the actual traffic operation state by simulation and this method,which have verified the technical feasibility of the method.