Research On Calculation Model Of Crowd Behavior Based On Emotional Contagion

Crowd behavior calculation refers to the modeling process including the construction of scenes and crowds corresponding to the real world in virtual space,modeling crowd complex behaviors,and realizing visual simulation.Regardless of the limitation of time and space,the micro behaviors of individuals and macro tendencies of groups are deduced at any scale and perspective.Crowd behavior calculation is an interdisciplinary research field of biology,psychology,computer graphics and other disciplines.It is widely used in security control,building layout,military training,computer games,film effect and so on.More and more scholars pay attention to it.In practical application,people have much higher requirements for the reliability and authenticity of crowd behavior calculation results and calculation performance.It is a great challenge to model the complex factors that affect crowd movements,explore the relationship between these factors and movements,and generate more real crowd movements.Researches have shown that emotion has a significant impact on crowd behaviors.The emotion modeling method in psychology is applied to the calculation method of crowd behavior to improve the similarity between simulation results and real crowd movements.In addition,we also combine the research of crowd behaviors with current social hot events in order to provide support for the decision-making of safety control of the government.Based on the above research background,this dissertation mainly focuses on the crowd behavior calculation model which integrates emotional contagion.In emergency scenarios,affected by the hazard,people are prone to be panicked and panic spreads among crowd.Crowd movements are greatly affected by emotion.Firstly,we study the movements of panicked crowds in this kind of typical scene.We describe the psychological factor(panic emotion)and the physiological factor(physical strength consumption)that have a great influence on crowd movements,and discuss the relationship between them.There are similar situations in crowd riot scenes.The emotions of activists have a great influence on their antagonistic behaviors and the emotion spread among different roles of individuals.In this paper,the emotional contagion mechanism is integrated into violent scenes.In these scenes,cops and activists are two groups with antagonism and game relations.We model the antagonistic emotions among different roles of individuals,integrate the antagonistic emotional contagion into the evolutionary game theory mechanism,and study the influence of antagonistic emotions on antagonistic behaviors.The research on emotion and movement in emergency and antagonistic scenarios involves the calculation of crowd trajectories.However,these works do not fully consider the influence of personality factors on movement trajectories.Besides,in normal scenarios(not emergency scenarios),personality has a very important impact on human behavior.Therefore,we continue to be engaged in the related work of trajectory prediction and personality.We predict the trajectories of pedestrians based on a large amount of real crowd movement data and consider the differences in personality.Model driven and data driven methods are used in the above trajectory calculation models.These methods have their own advantages and disadvantages.Finally,we combine these two kinds of methods and make full use of their advantages.Besides,we describe the emotional state of a group based on group movements and predict future movement trajectories of crowds.This paper focuses on the modeling of crowd movements.We integrate emotion into crowd behavior calculation model and it can simulate more real crowd behaviors.The main contributions of this dissertation are as follows:(1)We propose an emotion-based crowd simulation model integrating physical strength consumption in emergency scenarios.There are many factors that affect crowd movements and it is difficult to describe the relationship between these factors and crowd movements.In order to solve these issues,crowd movements in our model is determined by modeling the influence of physical strength consumption and panic emotion.In particular,human physical strength consumption is computed using a physics-based numerical method.Inspired by the James-Lange theory,panic levels are estimated by means of an enhanced emotional contagion model that leverages the inherent relationship between physical strength consumption and panic.(2)We present an antagonistic crowd simulation model which is integrated with antagonistic emotional contagion and evolutionary game theories.At present,most of the researches focus on modeling the same type of individuals' emotions and movements,and ignore the antagonistic emotions and behaviors among different types of individuals.In order to solve these problems,we combine enhanced susceptibleinfectious-susceptible(SIS)and game approaches,and model the antagonistic emotions between individuals in different roles using two components: mental emotion and external emotion.Our evolutionary game theoretic approach incorporates antagonistic emotional contagion through deterrent force,which is modelled by a mixture of emotional forces and physical forces defeating the opponents.(3)We present a novel trajectory prediction algorithm for pedestrians based on a personality-aware probabilistic feature map.At present,only a few factors are considered in trajectory prediction models and almost no framework can integrate too many factors.In addition,most of the models are based on short-term historical trajectories to predict the future trajectories,and often ignore the impact of personality on pedestrian trajectory.To address these problems,we present a probabilistic feature map.By adjusting its weights,our model contains many factors(especially personality)that affect the trajectories of pedestrians.Furthermore,we estimate the personality features of each pedestrian based on long-term historical trajectories and quantify the relationship between personality features and pedestrian trajectories.(4)We propose a hybrid driven trajectory prediction method based on group emotion.At present,most methods focus on specific scenes to model human behaviors and there is almost no general method.In addition,most of the current researches focus on individual emotion and there are few methods that consider the emotional state of a group.To solve these problems,we take group as the research object and pay attention to the emotion of groups.Besides,we propose a group division method based on emotional contagion.It takes into account the influence of interpersonal relationships on emotion and grouping.We make full use of the advantages of data-driven and model-driven methods and combine them together.