Study On Cellular Automata Model Of Continuous Flow In Urban Narrow Motor Vehicle Lane Environment

The concepts of reducing urban traffic pressure and improving the pedestrian environment have introduced new demands for the allocation of road space.As a form of road slimming,the narrowing of motor vehicle lane width has been proposed by scholars from the actual measurement and simulated driving to optimize the lane width standard in the specification.However,the application and promotion of narrow motor vehicle lanes still need to improve the research on the macroscopic traffic flow characteristics and microscopic vehicle behavior of the road after lane narrowing,as well as the research on traffic flow simulation model.This paper takes the continuous flow section of urban narrow motorway as the research object.Based on the analysis and modeling of macro and micro traffic flow characteristics,a cellular automata model is constructed to simulate the traffic flow phenomenon and capacity of narrow motorway.Firstly,through a comparative study of macroscopic traffic flow between narrow lanes and conventional wide lanes,it is demonstrated the necessity of studying the macroscopic characteristics of traffic flow in narrow lanes,as well as investigating micro-level vehicle behaviors and simulating models to explain their macroscopic differences.Field investigations are conducted to examine the macroscopic parameters of continuous flow under different service levels of narrow lanes and conventional lanes.Statistical methods are employed to analyze the differences,as well as the relationship between traffic volume,density,and speed.The results indicate that there are lateral influences among vehicles in narrow lanes,leading to lower and more stable speeds compared to conventional lanes.The capacity of a two-lane road with 2.8 meters lane width in the same direction is approximately 2500 vehicles per hour,and the optimal density ranges from 60 to80 vehicles per kilometer,both of which are lower than those of roads with 3.25 meters lane width.Secondly,the microcharacteristics of traffic flow are analyzed using vehicle trajectory data from narrow lanes,providing a foundation for simulation models.By analyzing the differences in vehicle behaviors between narrow lanes and conventional lanes,the study describes how vehicle behavior in narrow lanes is influenced by multi-lane vehicles and vehicle types,resulting in special phenomena such as "staggered following." Vehicle trajectory data is extracted from unmanned aerial vehicle(UAV)videos of continuous flow in narrow lanes.Taking into account different traffic conditions,the narrow lanes are divided into 12 different scenarios based on vehicle behavior.The characteristics and relationships of vehicle speed,lateral spacing from other vehicles,and longitudinal spacing are quantitatively analyzed under stable conditions in each scenario.Then,a selection model for vehicle acceleration,deceleration,and uniform speed behavior in narrow lanes is constructed based on CART classification trees.This model quantifies the microcharacteristics of traffic flow in narrow lanes and provides updated conditions for simulation models.For the six main scenarios divided based on vehicle behavior in narrow lanes,specific feature variables and target variables are selected.Machine learning modeling analysis is then conducted to obtain the probabilities of vehicle acceleration,deceleration,and uniform speed choices in each scenario,along with the corresponding feature variable partitioning conditions for behavior selection.The model results align with the statistical analysis of vehicle trajectory data,effectively describing the behavioral patterns of vehicles in various scenarios of narrow lanes.Finally,a cellular automaton model is constructed based on the behavioral patterns of vehicles in multiple scenarios of narrow lanes.This model is validated using macroscopic fundamental diagrams of measured traffic flows and used to simulate the capacity of narrow lanes.The results of the microscopic traffic flow behavior modeling are simplified,and the updating rules for narrow lane cellular automata are established based on the longitudinal spacing between vehicles.Separate updating rules are defined for small vehicles and large vehicles.Three sets of simulation control experiments are conducted,varying the random deceleration probability,lane interaction,and longitudinal spacing partitioning conditions,and the random deceleration probability parameters are calibrated based on the fundamental diagrams of traffic flow.The model is then used to simulate the capacity of a 2.8m wide one-way,two-lane road with four different proportions of large vehicles,obtaining reduction factors for the capacity of narrow lanes corresponding to each proportion of large vehicles.