Bus-stop Operating State Recognition And Capacity Study Based On Bus GPS Data

In recent years,domestic public transportation has developed rapidly,but as the scale of public transportation has expanded,the demand for passenger travel has also increased.The contradiction between supply and demand has become increasingly apparent.As an indispensable part of the daily operation of public transportation,analyzing the operating status of the bus stops and improving its capacity has certain research significance for improving the overall service level of the bus and maintaining the good operation of the urban public transportation system.In order to improve the efficiency and capacity of bus stops,and avoid large-scale queuing congestion in the stop,this paper analyzes the influencing factors of the bus stop,and takes the bus stop in the queue state as the key research object,and constructs the calculation model of the capacity of the bus stop and average berth blocking time under the variable overtaking rule to study the reasonable overtaking behavior under different conditions to maximize the capacity of the bus stop.Based on the GPS data,combined with the location data of the bus stops,this paper uses the method of extracting the data around the stops to extract the characteristic data of the bus at the stops,and combines the clustering algorithm to divide the stops into 2 categories.Compare the distribution fitting characteristics of the operating data of different types of stops.On this basis,the importance of different influencing variables to different types of stops is studied.Through the Logit model and the random forest model,the relationship between the operating characteristic data of the bus stop,the attribute data of the stop itself and the geographical variable data around the stop is deeply explored.The two types of models established were compared using traditional statistical analysis results and machine learning analysis results to analyze the importance of different influence factors for the classification of different types of stops.Secondly,through the recognition of different operating states of bus stops,based on the time-space trajectory diagram and state transition diagram,combined with Monte Carlo algorithm.A calculation model of the capacity of bus stops with different numbers of berths under the queuing state is proposed.The cellular automata model is used to construct a kind of operation simulation model of the bus stop in the queue state,and the simulation value is compared with the theoretical value obtained from the theoretical model to verify the reliability of the theoretical model.Finally,it analyzes the capacity model and the average berth blocking time model of the bus stops in the queue state under different overtaking rules.Through comparative analysis of the results,reasonable overtaking rules under different conditions are obtained.When the environment near the bus stop permits,when the variation coefficient of the bus service time is between 0.2 and 0.9,the free-overtaking rule is recommended.If only one overtaking behavior is allowed,it is recommended to allow overtaking to leave the stop when the variation coefficient of the bus service time is greater than 0.43.