Modeling And Simulation Of Microscopic Pedestrian Flow In MTR Hubs

MTR (Mass Transit Rail) is one of the most effective ways to solve traffic problems in large cities. The dominating content of harmonies traffic is the construction of traffic available, functional abundant, resource economic and environment friendly MTR hubs. It is very important to predict the efficiency and safety of pedestrians in MTR hubs in planning and designing such systems. How ever, the design and appraisal of such hubs were based on static and experience method traditionally. It has lots of pitfalls in engineering and theory, such as lack of design criteria, disability in estimating the mustering and evacuation efficiency of the station, little cognition on the relationship between passenger and traffic etc. Answers of above mentioned questions can provide methods to forecast the passenger "capacity" of both planning and existing hubs. The paper tries to answer these questions by discussing the pedestrian flow and its micro-simulation. The study is based on two research projects: one is practical project aims for estimating XUANWUMEN and CHONGWENMEN hub layout in Beijing MTR network; another is a theoretical project names "Traffic Design Theory of Interchange Hub in Passenger Dedicated Railway Line" supported by NSFC (National Scientific Foundation Committee).Firstly, the paper reviews a detailed state-of-the-art of research on the passenger flows in MTR hubs. Based on the deficiency of the existing study, it sets up a frame of passenger mustering and evacuating theory in MTR hubs, this frame shows the key problems and methods of the theory with two lines: "point-line-network" and "individual-group-flow". In the rest chapters, it studies the modeling and simulating of passenger flows in MTR hub based on pedestrian behavior which is the foundation in the frame.Passenger surveys were conducted to research pedestrian behavior at both macroscopic and microscopic level. The macroscopic level behavior was analyzed by references and the microscopic level behavior was extracted from field data in six parts: facility, direction, trips' aim, congestions, periods and evacuation. A semiautomatic data collecting system was developed. It develops a static way to analyze passenger flows in bottle neck, queues and delaysIt then presents a pedestrian mustering and evacuation model based on the behavior studied, a conversely improved A star algorithms was developed, such famous problems as decision synchronize conflict and cross deadlock was solved using new concepts with dynamic games and exchange mechanism. The proposed model combines the advantage of the multi-agent systems, cellular automata, "social force" and steering behavior, and makes it more flexible and powerful.It illustrates that the new model should be verified using four criteria: fundamental diagram, self-organization, emergency evacuation and practical data. It means the model should be effective not only on traffic problems but also on pedestrian self-organization. Also, it should be reproduce some phenomenon in the real world such as mustering and evacuation. The model is demonstrated to be effect on these four criteria. The proposed model is tested to be the only effective model of the existing model. The model also complement the pedestrian self-organization theory, proposes the mathematic relationships among self organization parameters.The paper studies two approaches of simulating pedestrian systems, discrete simulation and continuous simulation. It proposes the transfer condition of the two simulation systems and discusses the key techniques of the system implementation. It then gives the appraisal criteria of the MTR passenger flow. These index using the data from simulation systems as independent variable, and considers the capacity from safety, efficiency and comfortable.It presents the multi-agent simulation model based on the proposed study, simulation software named MTR-PedSIM has been developed using the proposed model. Repetitious pedestrian simulation cases were tested, the condition, mechanism and essence of pedestrian self-organization was found by these tests.At last, an existing hub and a planning hub were simulated using the MTR-PedSIM as case study. In the simulation process, more results were achieved. It is found that the bottleneck spread at temporal and spatial dimensions, pedestrian evacuation state shifts under emergency situation. The study simulation reproduces the characteristics of passenger flow and spatial occupation in the hub. It gives a new dynamic solution to the hub estimation and optimization problems. It shows that the result is credible and promising.