Dorine C. Duives

Exit Choice Decisions during Pedestrian Evacuations of Buildings

This paper assesses how exit choice behavior influences the total egress behavior of a crowd during building evacuation. A discrete choice model is proposed to represent pedestrian exit choice decisions during evacuation. The proposed model is based on responses to an Internet questionnaire conducted in the Netherlands and the United States. The results of the Internet questionnaire, consisting of 20 choice experiments by 100 respondents making trade-offs between distance, angular deviation, and group following behavior, are used to estimate a multinomial logit model. Accordingly, a new pedestrian simulation approach that uses a microscopic cellular automata model (EvacPed) is presented. Three levels of decision making are incorporated: global exit choice, regional adaptive route choice, and local operative behavior. Simulations estimate the difference in evacuation time between four exit choice strategies: least distance path, least travel time path, discrete choice incorporating hive knowledge, and discrete choice incorporating vision field. The results suggest that group following behavior has a large impact on the effectiveness of building evacuation.

Trajectory Analysis of Pedestrian Crowd Movements at a Dutch Music Festival

Major pedestrian crowd movements have proven to be volatile in the past. The contribution of this paper is to assess a new recording and a new analysis technique to improve the currently limited theoretical knowledge on the movement of individual pedestrians within a crowd. Using a camera attached to a UAV, footage of pedestrian crowd movements has been recorded at the major music festival Lowlands in the Netherlands. This paper describes a detection and tracking tool, resulting in trajectory information for all pedestrians visible within the footage. Footage recorded at Lowlands is used to assess the new techniques’ contribution, resulting in trajectory analyses of individual pedestrian movements in a crowd.

How do people cycle in Amsterdam, Netherlands? : Estimating cyclists' route choice determinants with GPS data from an urban area

Nowadays, the bicycle is seen as a sustainable and healthy substitute for the car in urban environments. The Netherlands is the leading country in bicycle use, especially in urban environments. Yet route choice models featuring inner-city travel that includes cyclists are lacking. This study estimated a cyclists’ route choice model for the inner city of Amsterdam, Netherlands, on the basis of 3,045 trips collected with GPS data. The main contribution of this study was the construction of the choice set with an empirical approach, which used only the observed trips in the data set to compose the choice alternatives. The findings suggested that cyclists were insensitive to separate cycle paths in Amsterdam, a city characterized by a dense cycle path network in which cycling was the most prominent mode of travel. In addition, cyclists were found to minimize travel distance and the number of intersections per kilometer. The impact of distance on route choice increased during the morning peak when schedule constraints were more prevalent. Furthermore, overlapping routes were more likely to be chosen by cyclists, everything else being the same.

Quantitative Estimation of Self-Organization in Bi-directional and Crossing Flows During Crowd Movements

Understanding emerging phenomena in crowd movements is necessary to understand how pedestrians behave during these movements under different circumstances and over time. Measures able to identify self-organization patterns are currently scarce. In the present study the way in which three measures (the cluster-method (Moussaid, et al. PLoS Comput Biol 8(3):e1002442, 2012), Efficiency (Helbing (1997) Verkehrsdynamik – Neue physikalische Modellierungskonzepte, 1st edn. Springer, Berlin/Heidelberg, p. 46), and Polarization (Hemelrijk and Hildenbrandt, PLoS ONE 6(8):e22479, 2011)) identify the presence of self-organization within crowd movements. Trajectory data sets resulting from a laboratory experiment and several simulations are used as a basis for the assessment. It was found for all three methods that the extent to which self-organization can be accurately predicted depends on the flow situation. Furthermore, two out of three methods were able to detect the presence of self-organization in pedestrian flows at all.

Pedestrian Evacuation Optimization Dynamic Programming in Continuous Space and Time

This paper deals with the optimal allocation of routes, destination, and departure times to members of a crowd, for instance in case of an evacuation or another hazardous situation in which the people need to leave the area as quickly as possible. The generic approach minimizes the evacuation times, considering the demand dependent waiting times at bottlenecks within the considered infrastructure. We present the mathematical optimization problem for both the optimal instructions, and the continuum model describing the pedestrian flow dynamics. The key contribution of the approach is that it solves the evacuation problem considering the entire solution space in a continuous manner (i.e. both the time dimension and the routing), implying that for each location and for each time instant the optimal path towards the most favorable exit is calculated, taking into consideration the traffic flow operations along the routes. The approach is generic in the sense that different network loading models can be used, and that a variety of components can be added to the optimization objective without loss of generality. Next to presenting the framework and the mathematical model, we propose an iterative numerical solver to compute the optimal instructions. We demonstrate the abilities and opportunities of this optimization framework with two case studies.

Monitoring the Number of Pedestrians in an Area: The Applicability of Counting Systems for Density State Estimation

Crowd monitoring systems are more and more used to support crowd management organizations. Currently, counting systems are often used to provide quantitative insights into the pedestrian traffic state, since they are fairly easy to install and the accuracy is reasonably good under normal conditions. However, there are no sensor systems that are 100% accurate. Detection errors might have severe consequences for the density state estimation at large squares. The consequences of these errors for pedestrian state estimation have not yet been determined. This paper studies the impact of one specific type of detection error on the functionality of counting camera systems for density state estimation, namely, a randomly occurring “false negative” detection error. The impact is determined via two tracks, a theoretical track and a simulation track. The latter track studies the distribution of the cumulative number of pedestrians after 24 hours for three stylized cases by means of Monte Carlo simulations. This paper finds that counting camera systems, which have a detection error that is not correlated with the flow rate, provide a reasonably good estimation of the density within an area. At the same time, if the detection error is correlated with the flow rate, counting camera systems should only be used in the situation where symmetric demand patterns are expected.

Operational Walking Dynamics of Crowds Modeled with Linear Regression

In this study the influence of interaction characteristics on operational walking dynamics within a crowd—specifically the influence of the distance headway, time headway, angle of sight, angle of interaction, walking speed, and number of pedestrians located nearby on a pedestrian’s change in speed and direction—is investigated. To the authors’ best knowledge, this is the first time that the combined effect of the characteristics of interactions between pedestrians on the operational walking dynamics of pedestrians has been quantified. The walking speed and the number of pedestrians in close proximity were found to influence the adaptation of speed and direction. The other characteristics of the interaction affect either the change in speed (i.e., distance headway and interaction angle) or the change in direction (i.e., time headway and angle of sight). The results of this study strongly indicate that the density experienced by pedestrians is not the only characteristic of the crowd that affects pedestrians’ operational walking dynamics. Consequently, to model crowd movements correctly, the other characteristics of the interaction must also be taken into account in pedestrian flow theory and simulation models.

Comparison of three algorithms for real-time pedestrian state estimation - supporting a monitoring dashboard for large-scale events

Technical advancements allow for the development of crowd monitoring and management support systems to ensure the safety of pedestrians during large-scale events. This way, pedestrian behaviour can be monitored during events, and potential dangerous situations can be identified in a timely manner. A real-time prototype Crowd Monitoring Dashboard has been developed for a large nautical event in Amsterdam in 2015. Three main functions of such a system have been identified: the real-time data collection, the traffic engineering functions and the visualization. This paper focuses on three state estimation algorithms in the second component - the traffic engineering functions. These algorithms are presented and cross compared in terms of performance, data usage and practical applications, based on the empirical data collected from this large-scale event. The outcome can shed useful light on the feasibility of the data sources and estimation methods for future real-time applications on crowd monitoring and management for large-scale events.

Sensitivity Analysis of the Local Route Choice Parameters of the Continuum Model Regarding Pedestrian Movement Phenomena

Numerous pedestrian simulation models have been proposed in the last decade, many of which simulate the movement behaviour of pedestrians microscopically. However, the numerous degrees of freedom of microscopic models complicate the calibration process severely. Moreover, the computation speed of many microscopic simulation models leaves much to be desired. Especially in cases where computation speed is essential and no microscopic data is available to calibrate the model, macroscopic models outperform microscopic models. This study provides a detailed assessment of the impact of the combination of delay and density within the formulation of the local route choice behaviour on the predicted movement dynamics of the crowd in a continuum model (Physica A 416:684–694, 2014 [9]; Transp Res Part C Emerg Technol 59:183–197, 2015 [10]). This study aims to understand how the parameters of the respective versions of the continuum model influence the development crowd movement phenomena. The impact of the parameter sets of the continuum model is assessed for a uni-directional bottleneck, uni-directional corner rounding, bidirectional straight walking and intersecting movements.