Simulation Of Pedestrian Dynamics Research Articles

Learning from experimental data to simulate pedestrian dynamics

Accuracy is a critical concern for those engaged in the simulation of pedestrian dynamics. Although a range of pedestrian models has been developed, the performance of these models is heavily limited by simplifications inherent in the usual knowledge-based approach. In this study, we propose data-driven models of pedestrian dynamics based on the machine learning approaches. The models are constructed in an iterative manner; thus, the long-term future trajectories are predicted using the current output as the input at the next timestep. The performance of our models is evaluated by comparison with an well-researched knowledge-based pedestrian model. The evaluation included two metrics: the ADE/FDE evaluation metric and another metric that measures whether the simulation corresponded to ground truths from the collective dynamics perspective. Experiments on two experimental datasets show that our models outperform the knowledge-based model. Our findings highlight the considerable potential of the data-driven approach for handling pedestrian simulation tasks.

Simulation of pedestrian dynamics based withemantic trajectory segmentation

The article analyzes the existing methods of information processing necessary for the functioning of the system of intelligent control over unregulated pedestrian crossings based on aggregation and data processing by means of IOT. The state space model of the switching Kalman filter is considered, the development of mathematical software for the analysis and processing of information based on the results of intelligent control over unregulated pedestrian crossings, in particular with semantic segmentation of trajectories using agent-based models, is carried out. An MDA (Markov Decision Process) state space model is presented, a Hidden Markov Model (HMM) which has discrete hidden variables. The developments for the development of the following subsystems are presented: activity detector subsystem. Receives video frames as input, supports the static object model (background model) and returns the hotspot mask for the current frame; subsystems for detecting and tracking objects (pedestrians and cars). Based on the video frame and hotspot mask, it detects and accompanies objects of a given class, returning their coordinates; trajectory analysis subsystem. Analyzing the history of movement of pedestrians and cars, returns the facts of traffic violations.

Agent-based simulation of pedestrian dynamics for exposure time estimation in epidemic risk assessment

PurposeWith the coronavirus disease 2019 (COVID-19) pandemic spreading across the world, protective measures for containing the virus are essential, especially as long as no vaccine or effective treatment is available. One important measure is the so-called physical distancing or social distancing.MethodsIn this paper, we propose an agent-based numerical simulation of pedestrian dynamics in order to assess the behavior of pedestrians in public places in the context of contact transmission of infectious diseases like COVID-19, and to gather insights about exposure times and the overall effectiveness of distancing measures.ResultsTo abide by the minimum distance of 1.5 m stipulated by the German government at an infection rate of 2%, our simulation results suggest that a density of one person per 16m2 or below is sufficient.ConclusionsThe results of this study give insight into how physical distancing as a protective measure can be carried out more efficiently to help reduce the spread of COVID-19.

A generic and density-sensitive method for multi-scale pedestrian dynamics

The simulation of pedestrian dynamics is required for a wide range of applications, including crowd management, evacuation scenarios, and the design of buildings. Microscopic simulators are based on the simulation of individual agents and their mutual interactions. In comparison to macroscopic simulators that are based on network-flow models, microscopic simulators are capable of resolving movement patterns and potentially risky situations more precisely thanks to the modeling of individual behaviour, however, with the cost of much higher computational effort. Regarding the spatial resolution, microscopic simulators are either based on continuous (SpaceCont) or discrete (SpaceDisc) approaches.Whereas in the former, a pedestrian can take any position in the continuous space, in the latter, space is discretized into cells, thus restricting the possible positions of a person. SpaceDisc approaches are less precise, but computationally more efficient and thus typically applied in scenarios with a large number of persons or expanded simulation areas, respectively. For many applications, the spatial resolution of a discrete model is sufficient for observing the phenomena in question. However, in situations with high densities or small-scale obstacles, a continuous model is necessary to obtain realistic results.To combine the advantages of both approaches, we propose to integrate SpaceCont and SpaceDisc into a hybrid simulation model. Such a hybrid approach allows simulating critical regions with a continuous spatial resolution and uncritical ones with discrete spatial resolution while enabling consistent information exchange between the two simulation models. We introduce a generic approach that provides consistent solutions for the challenges resulting from coupling diverging time steps and spatial resolutions. Furthermore, we introduce a dynamic and density-sensitive approach to detect dense areas during the simulation run. If a critical region is detected, the simulation model used in this area is dynamically switched to a space-continuous one. The correctness of the hybrid model is evaluated by comparison with the simulation results of other, well-established simulation models.

Data-driven simulation of pedestrian collision avoidance with a nonparametric neural network

Data-driven simulation of pedestrian dynamics is an incipient and promising approach for building reliable microscopic pedestrian models. We propose a methodology based on generalized regression neural networks, which does not have to deal with a huge number of free parameters as in the case of multilayer neural networks. Although the method is general, we focus on the one pedestrian - one obstacle problem. Experimental data were collected in a motion capture laboratory providing high-precision trajectories. The proposed model allows us to simulate the trajectory of a pedestrian avoiding an obstacle from any direction. Together with the methodology specifications, we provide the data set needed for performing the simulations of this kind of pedestrian dynamic system.

Unpredictability in pedestrian flow: The impact of stochasticity and anxiety in the event of an emergency

Random interactions among pedestrians, usually produced by psychological effects (e.g. anxiety), can produce collective nontrivial phenomena in their flow dynamics. Here, we improve our understanding on these phenomena by implementing cellular automaton (CA) simulations in a simplified one-dimensional corridor model. We consider different types of randomness that could be used to parametrize social, cultural or psychological behaviors; their incidence on the average evacuation time and standard deviation are studied. The randomness in the pedestrians’ motion allows separating between different phases of movement, which in the end, drives the dynamics. Our study allows to distinguish among emergent collective phenomena, unpredictability, and unexpected Brownian type behavior in certain regimes, that occur in one-dimensional hallways, such as exists in airplanes, cinemas, massive sports events, etc. These findings may be relevant to evacuation procedures in the event of an emergency, where anxiety may compromise people’s performance and the effectiveness of the evacuation plan. Therefore, it could be highly useful to introduce these aspects in the simulation of pedestrian dynamics.

A mesoscopic model for large-scale simulation of pedestrian dynamics

A mesoscopic pedestrian model is proposed, considering pedestrians as individuals and describing their movement by means of aggregate density-flow relationships. The model builds on a stochastic process, describing transition rates among adjacent sites on a lattice. Each lattice can contain several pedestrians. The approach is minimal and fast to simulate, and, by construction, capable of capturing population heterogeneity as well as variability in walking behaviour and en-route path choice. The model is more efficient than microscopic models, and potentially more accurate than macroscopic ones. We calibrate and validate the model using real data and carry out several numerical experiments to present its key properties and possible applications for simulation of large-scale scenarios.

Simulation of Pedestrian Dynamics with Macroscopic and Microscopic Mathematical Models

Simulation of Pedestrian Dynamics with Macroscopic and Microscopic Mathematical Models

Special issue on Simulation with Cellular Automata

Cellular Automata (CA) are discrete dynamical systems. The definition of relatively simple, local interactions results in complex global behaviors of the whole system. This makes it possible to create and study models of many complex systems in different fields such as physics, engineering, environment science, social science, and life sciences. Such an approach is very promising, and during the past few decades CA-based modeling has attracted a growing number of researchers and practitioners interested in studying models of complex phenomena. In this special issue entitled ‘‘Simulation with Cellular Automata,’’ we have proceeded with an open-call announcement targeting novel simulation CA-based studies as well as revised and extended papers presented during the 11 International Conference on Cellular Automata for Research and Industry, held at AGH University of Science and Technology in September 2014. After a thorough and meticulous review process, we finally came up with seven (7) accepted papers dealing with various and, in most cases, interdisciplinary topics under the umbrella of CAbased simulation. More specifically, in the first paper, ‘‘Modelling and Simulation of Complex Cellular Automata Using CellDEVS,’’ Gabriel Wainer and Joaquin Fernandez present a discrete event formalism Cell-DEVS, which enables timing delay constructions. Cell-DEVS is introduced as a formalism supporting a discrete event approach to define cellular models with timing delay constructions and using a simple definition of complex timing. The application of the Cell-DEVS paradigm enables more efficient cellshaped modeling by also supporting the inherent complexity of the proposed models. The second paper is ‘‘Modeling Pedestrian Behavior with Cell-DEVS: Theory and Applications’’ co-authored by Ala’a Al-Habashna and Gabriel Wainer. On the basis of the Cell-DEVS approach, an entity-based modeling and simulation of crowd dynamics, applied in an analysis of evacuation and occupancy levels of multiple-floor buildings, is analytically described. Specifically, the authors present a practical application of Cell-DEVS formalism in the modeling and simulation of pedestrian dynamics and their behavior in different situations. The third paper, entitled ‘‘The Spatial Reproduction Number in a Cellular Automaton Model for Vector-Borne Diseases Applied to the Transmission of Chagas Disease,’’ is co-authored by Baki Cisse, Samira El Yacoubi, and Sebastien Gourbiere. The proposed approach is based on CA and applied to Chagas disease for homogeneous as well as heterogeneous landscapes. It refers to modeling of spatial changes in the basic reproduction number for vector-borne diseases. The main outcome from the performed simulations is an effect of local vector dispersal in a heterogeneous landscape for Chagas disease transmission. The fourth paper, entitled ‘‘Graph Cellular Automata Approach to the Maximum Lifetime Coverage Problem in Wireless Sensor Networks’’ by Antonina Tretyakova, Franciszek Seredynski, and Pascal Bouvry, proposes the generalization of a CA concept called Graph Cellular Automata to solve the maximum lifetime coverage problem in wireless sensor networks (WSN). A new localized algorithm is presented, which uses only some knowledge about the neighboring sensors of WSNs, still able to selfreorganize in such a way to preserve a required coverage ratio and prolong lifetime of the WSN. The fifth paper, ‘‘Cellular Automata Approach for Parallel Machine Scheduling Problem’’ authored by Edyta Kucharska, Katarzyna Grobler-Debska, Krzysztof Rączka, and Lidia Dutkiewicz, deals with scheduling problems with parallel machines using the CA paradigm. An asynchronous non-homogenous cellular automaton is

A hybrid multi-scale approach for simulation of pedestrian dynamics

One of the most important aspects for a realistic prediction of pedestrian flows is the modelling of human navigation in normal situations such as early design phases of buildings or transportation systems and hubs as well as in evacuation studies to enhance safety in existing infrastructures. To overcome the limitations of current navigation models, this paper proposes a new hybrid multi-scale model, which closely links information between the small-scale and large-scale navigation layer to improve the navigational behaviour. In the presented hybrid navigation model, graph-based methods using visibility graphs are used to model large-scale wayfinding decisions. The pedestrians’ movements between two nodes of the navigation graph are modelled by means of a dynamic navigation field. The navigation field is updated dynamically during the runtime of the simulation, explicitly considering other pedestrians for determining the fastest path.The proposed hybrid approach provides a realistic modelling of human navigational behaviour and thus a realistic prediction of flows since it reflects the human cognitive processes triggered by wayfinding tasks. This includes taking into account other pedestrians for routing decisions who are visible from the current position of the considered pedestrian. The paper discusses the concept and the technical details of the proposed hybrid multi-scale approach in detail and presents an extensive case study demonstrating its advantages.

Crowd and environmental management during mass gatherings

Crowds are a feature of large cities, occurring not only at mass gatherings but also at routine events such as the journey to work. To address extreme crowding, various computer models for crowd movement have been developed in the past decade, and we review these and show how they can be used to identify health and safety issues. State-of-the-art models that simulate the spread of epidemics operate on a population level, but the collection of fine-scale data might enable the development of models for epidemics that operate on a microscopic scale, similar to models for crowd movement. We provide an example of such simulations, showing how an individual-based crowd model can mirror aggregate susceptible-infected-recovered models that have been the main models for epidemics so far.

From Ant Trails to Pedestrian Dynamics

This paper presents a model for the simulation of pedestrian dynamics inspired by the behaviour of ants in ant trails. Ants communicate by producing a pheromone that can be smelled by other ants. In this model, pedestrians produce a virtual pheromone that influences the motion of others. In this way all interactions are strictly local, and so even large crowds can be simulated very efficiently. Nevertheless, the model is able to reproduce the collective effects observed empirically, eg the formation of lanes in counterflow. As an application, we reproduce a surprising result found in experiments of evacuation from an aircraft.

From Ant Trails to Pedestrian Dynamics

This paper presents a model for the simulation of pedestrian dynamics inspired by the behaviour of ants in ant trails. Ants communicate by producing a pheromone that can be smelled by other ants. In this model, pedestrians produce a virtual pheromone that influences the motion of others. In this way all interactions are strictly local, and so even large crowds can be simulated very efficiently. Nevertheless, the model is able to reproduce the collective effects observed empirically, eg the formation of lanes in counterflow. As an application, we reproduce a surprising result found in experiments of evacuation from an aircraft.

Simulation of pedestrian dynamics using a two-dimensional cellular automaton

We propose a two-dimensional cellular automaton model to simulate pedestrian traffic. It is a v max=1 model with exclusion statistics and parallel dynamics. Long-range interactions between the pedestrians are mediated by a so-called floor field which modifies the transition rates to neighbouring cells. This field, which can be discrete or continuous, is subject to diffusion and decay. Furthermore it can be modified by the motion of the pedestrians. Therefore, the model uses an idea similar to chemotaxis, but with pedestrians following a virtual rather than a chemical trace. Our main goal is to show that the introduction of such a floor field is sufficient to model collective effects and self-organization encountered in pedestrian dynamics, e.g. lane formation in counterflow through a large corridor. As an application we also present simulations of the evacuation of a large room with reduced visibility, e.g. due to failure of lights or smoke.