Research On Simulation Of Crowd Movement Behaviors In Smart Venue

Crowd monitoring and smart venue management become increasingly intelligent with the spread of indoor positioning technology.Analyzing the movement trajectories of crowds in smart venues facilitates the recognition of crowd behavior patterns in the indoor environment,which is useful for the smart venue management.However,crowd movement trajectories in real-world concern personal privacy,so it is important to use simulation data for related studies.But many difficulties in simulating crowd movement behavior result in few effective crowd movement behavior simulation methods and high-quality simulation data in smart venues.First,a realistic simulation scenario needs to be designed in detail;second,the movement behavior of the crowd needs to be reasonable and consistent with the simulation scenario;third,the simulation results need to be validated from multi-perspectives.In this paper,we propose a smart venue crowd movement behavior simulation method and publish a high-quality simulation data.The research work mainly includes:(1)Large-scale academic conference simulation scenario design.The simulation scenario in this paper is a large-scale international academic conference on cyber security.First,through on-site research and interviews,we define seven types of people and design a two-story smart venue physical space that can accommodate more than 5000 people.Then,we define diverse conference events and some complex abnormal events.Finally,we create entity models of the personnel,venue,and event with the rich and realistic multi-dimensional attributes.(2)A "constraint-motivation-control" simulation method for crowd movement behavior in smart venues.In terms of constraint,we design a behavior-constraint model including time constraint,permission constraint,energy constraint,and capacity constraint,which can restrict the movement time,movement area,movement duration,and movement speed of personnel respectively.In terms of motivation,we design a behavior-driven model that drives personnel movement from both personal interest and the attractiveness of the event.In terms of control,we design the individual control strategies for manipulating changes in entity attributes,as well as overall behavior control strategies for simulating the dynamic behavioral judgment and decision processes of personnel.The simulation method can ultimately generate simulation data with reasonable movement behavior and various movement patterns.(3)Two-stage simulation results validation method.In this paper,simulation results are validated using internal and external validation methods.We design a visual analytic system for internal validation to modify the scenario design and simulation details by interactively analyzing the simulation data.In the external validation,we use simulation data as the China Vis Data Challenge 2019 contest,inviting contestants to analyze conference schedules,conclude personnel movement patterns,and identify abnormal events.Based on evaluation results from 75 contest entries and feedback from 359 contestants,we perform a comprehensive validation of the usability,completeness,and efficiency of the simulation data.The results show that the simulation data can effectively reflect the pre-defined scenarios and perform well in completeness and usability.Also,some abnormal events are challenging with complicated associations that can effectively identify the performance of methods,technologies,and systems for simulation data analysis.