Empirical Investigation On Pedestrian Complex Movement Characteristics And Interaction Behaviors

The movement of pedestrian crowd is complex.There is a rise in the hierarchy of motion complexity from uni-directional flow to multi-directional flow,from free flow to congested flow.Pedestrians are interacting with their walking environment including pedestrian facilities,as well as other road users such as other pedestrians,vehicles and non-motorists such as bicycles and scooters.These interactions can be categorized into four types i.e.Pedestrian to Infrastructure(P2I),Pedestrian to Pedestrian(P2P),Pedestrian to Vehicle(P2V)and Pedestrian to Non-motorists(P2N).As a result of such a variety of interactions,the collective movement characteristics of pedestrian crowd are changed,thereby promoting the constant updates of simulation models.With the increasing computing power of simulation tools and the advancements in microscopic modeling techniques,a large quantity of models begin to consider the microscopic interactions of pedestrian crowd.Therefore,it is imperative to glean deeper insights into the movement characteristics and interaction behaviors of pedestrian crowd under various complex scenarios.However,the verification and validation of pedestrian simulation models require comprehensive,authentic and reliable empirical data.Past studies on crowd motions have been conducted predominantly through theoretical approaches,empirical data are not sufficient.Moreover,previous empirical studies are usually carried out through the analysis of a specific type of movement behavior or phenomenon,seldom of the them have established a theoretical framework for pedestrian crowd studies based on the hierarchy of complex movement patterns and interaction behaviors.Therefore,to this end,this paper first presents an exhaustive review of the empirical studies on pedestrian complex movements and attempts to establish a theoretical framework.Then,through two series of case studies on empirical data collection of complex movement scenarios,further efforts are devoted to explore novel research methodologies and approaches in this field.The detailed contributions of this work are summarized into five bullets as follows:(1)The establishment of theoretical framework for pedestrian complex movement.A state-of-the art review on empirical data collection of pedestrian complex movements for the past two decades is presented.Based on the governing factors of movement characteristics,complex movement patterns and behaviors are categorized into external governed movement and internal driven movement.Apart from the literature review of each type of movement pattern and behavior,an assessment of movement complexity is performed considering the locomotion and behavior aspects of a specific movements.(2)Empirical data collection approaches of pedestrian crowd movement.Based on the state-of-thepractice methods of empirical data collection,three key elements of empirical data collection are summarized i.e.techniques(traditional methods vs.emerging technologies),controllability(controlled experiments vs.field observation)and research subjects(humans vs.animals).Moreover,the strengths,weaknesses,opportunities and threats for different approaches are assessed through SWOT analysis.Further,three data collection approaches including controlled experiments,video processing and smartphone sensing are introduced in details.(3)Macroscopic and microscopic analysis on crowd motion characteristics.A series of trajectory based calculation methods for measuring the macroscopic and microscopic properties of pedestrian crowd e.g.density,speed,flow,time headway and conflicts are developed.Then,the effect of different measurement methods on fundamental diagrams are examined.Meanwhile,time headway distributions are utilized to fit a power law model that can be used to evaluate the flow and clog of bottleneck areas.At last,the above analytical methods are coded in MATLAB and packaged as a software named pedestrian crowd trajectory analysis software.(4)Controlled experiments on pedestrian crowd merging movement at three-leg intersections.A total of 14 sets of experiments are designed to investigate the impact of merging angles,flow directions,desired speeds and vision conditions on crowd dynamics.A sum of 92 human participants are involved in this two series of experiments.Results show that merging movements tend to occur within a fixed region(Pedestrian Merging Zone,PMZ)in which pedestrians' speed reductions and conflicts can be observed.As the rise of merging angle,the safety and mobility situations of PMZ are worsen.In addition,opposed merging has better performance compared with inverse merging in T-junctions.Moreover,desired speed has a significant impact on the fluctuations of speed and travel time before and after merging.At last,it is observed that open vision condition is better than blocked vision condition in terms of the benefits to the coordinated and stabilized operation of merging flows in PMZ.(5)Controlled experiments on pedestrian crowd egress movement at bottlenecks.To examine the adjustments of architectural configurations on crowd egress dynamics,a total of 14 types of exit layouts(considering different exit positions,obstacle sizes and positions)are designed for a square room with a narrow bottleneck door exit.A sum of 50 participants are enrolled to egress through the exit under low and medium competitiveness.Escape order and outflow at exit segment are automatically extracted from the videos.Results show that careful adjustments of obstacle size and position can increase the outflow of evacuees up to 33% compared with standard design at corner exit under high competitive situation.Meanwhile,improper use of obstacle can impede the outflow up to 7%.Moreover,the flow and clog states of exit segment are measured by the time headway distribution fitting with a power law model.The fitting results are consistent with several previous studies on animal egress and granular bottleneck flow.