Search

Abstract: The evacuation process in a teaching building with two neighboring exits is investigated by means of experiment and modeling. The basic parameters such as flow, density and velocity of pedestrians in the exit area are measured. The exit-selecting phenomenon in the experiment is analyzed, and it is found that pedestrians prefer selecting the closer exit even though the other exit is only a little far. In order to understand the phenomenon, we reproduce the experiment process with a modified biased random walk model, in which the preference of closer exit is achieved using the drift direction and the drift force. Our simulation results afford a calibrated value of the drift force, especially when it is 0.56, there is good agreement between the simulation results and the experimental results on the number of pedestrians selecting the closer exit, the average velocity through the exits, the cumulative distribution of the instantaneous velocity and the fundamental diagram of the flow through exits. According to the further simulation results, it is found that pedestrians tend to select the exit with shorter distance to them, especially when the people density is small or medium. But if the density is large enough, the flow rates of the two exits will become comparable because of the detour behaviors. It reflects the fact that a crowd of people may not be rational to optimize the usage of multi-exits, especially in an emergency. [Copyright &y& Elsevier]

Efficient passenger flow assignment is essential for ensuring the safe and stable operation of the urban rail transit (URT). This paper proposes a new passenger flow assignment approach to address this challenge. The approach generates two agents: connected vehicle and unconnected passenger, to simulate vehicle operation and passenger travel for the URT network. It derives the passenger trajectory by the passenger route evolution mechanism to achieve the passenger flow assignment for the whole network. Furthermore, it provides statistics on platform passenger flow when obtaining the overall station passenger flow. Meanwhile, the approach employed Automatic Fare Collection (AFC) data as input to analyze the passenger flow features of the Shanghai Metro. The passenger flow assignment results can reveal the start and end of morning and evening peak hours at a time granularity of seconds. That is, 7:16:20 and 9:33:00 for the morning peak and 16:49:00 and 19:22:50 for the evening peak, respectively. The results also reveal the route choice pattern of passengers. Passengers have preferences for each type of route, in descending order: the shortest route, the minimum number of stations route, and the minimum number of transfer stations route, when they have a choice of these three routes. Moreover, the results can accurately identify large passenger flow stations and provide data support for station passenger flow management strategies. • This paper proposes a new passenger flow assignment approach. • The approach simulates connected vehicles and unconnected passengers. • Count platform passenger flow when obtaining the overall station flow. • Analyzing passenger route choice preferences and large passenger flow stations. [ABSTRACT FROM AUTHOR]

The stairwell is the main evacuation channel of super high-rise buildings, so revealing the evacuation characteristics on stairs in such buildings based on a full-scale experiment is critical for safer evacuation. However, currently there are few studies focusing on the evacuation performance of crowd in such buildings, especially for the collective movement scenarios. Here, an evacuation experiment is carried out in Shanghai Tower with a vertical height of 583 m, which is the second tallest building in the world. The evacuation speeds for 69 participants in 58 stair sections are extracted and the effects of travel distance, gender, age and density on descent speeds are analyzed and discussed. According to the result that the group of participants with a 9.63% higher travel height spends a 16.39% longer evacuation time, yet within each group the speed does not decrease with the increase of the moving distance, we propose an interesting hypothesis that the travel distance may affect descent speed, but it needs to be further verified. The speeds of the female are significantly lower than those of the male and the female are more sensitive to the increase of travel distance. At the same time, younger participants show better evacuation ability than the older ones, including that they evacuate faster and are less affected by the confluence. Furthermore, a new method is proposed to calculate density and the fundamental diagram These results are useful for the validation and calibration of related evacuation models, and can provide basic data for the design of emergency evacuation facilities in super high-rise buildings. • An evacuation drill is conducted in a super high-rise building up to 126 stories. • Pedestrian evacuation speeds in different stair sections are extracted. • The effects of travel distance, gender, age and congestion on descent speed are analyzed and discussed. • A new measurement method for fundamental diagram is proposed. [ABSTRACT FROM AUTHOR]

Calculating the evacuation time during staircase evacuation is essential to evaluate the security conditions of high-rise buildings under emergency scenarios. A three-stage hierarchical model for calculating emergency evacuation time is constructed by combining discrete state update rules and continuous forms in computation to investigate the evacuation process of pedestrians in multi-floor buildings in emergency circumstances. Combined with simulation analysis, the model considers the capacity constraint of stairwell and congestion effects and calculates the evacuation time hierarchically. The model's validity is verified by the existing evacuation simulator for a hypothetical building case study, coupled with a comprehensive computational comparison. Calculations on entire and separate evacuation times are explored considering different stairway capacity restrictions and geometric features. Computational results demonstrated that this model is effective and may be regarded as a valid tool for staircase evacuation time calculation in the context of high-rise buildings. • A macro-mathematical model combining discrete state update procedures and continuous calculation formulas. • Three stages are subdivided for quick calculation of evacuation time hierarchically. • Evacuation time calculation and computational configuration under dual supporting of theoretical model and software simulation. • The alterations in staircase capacity, stair width and landing area are addressed. [ABSTRACT FROM AUTHOR]

Vehicle travel time is widely regarded as an important performance of the road traffic as well as a significant decision factors for travelers. While the travel time is variable because of the stochastic factors such as traffic flow, signals, and some other emergency conditions. Considering the dynamic characteristics of road traffic flow and the importance of shortening travel time for emergency transport task, in this paper, a basic vehicle travel time model based on empirical velocity–density function is established, in which the travel time is divided into road running time and signal waiting time. Simulations based on the basic model are conducted to investigate the comprehensive effect of the multiple factors on the travel time, such as traffic density, number of signals, circle time, green ratio and offset time between adjacent signals. A green-wave traffic objective-oriented method is also explored to achieve the shortest travel time. The results demonstrate that plateaus and sloped plateaus would appear under the comprehensive influence of both traffic density and traffic light number or signal parameters. It is found there is a minimum traffic time with a certain offset time under different traffic density when the traffic lights on a route do not change synchronously. The fitting formula and theoretical formula for the optimal offset time are also proposed, which can be used to achieve green-wave traffic in a practical case. • A practical method to estimate the route travel time is proposed by establishing a travel time model. • The model could calculate both road running time and waiting time before signals. • The comprehensive influence of multiple factors on the travel time is investigated. • There is a minimum traffic time with a certain offset time under different traffic density. • The fitting formula and theoretical formula are proposed to achieve a green-wave traffic. [ABSTRACT FROM AUTHOR]

Humans plan their movements mainly based on visual information. However, agents in few existing evacuation models perceive the environment by using visual information. To obtain the visual features of an individual in a discrete field, a Dynamic Scanning Algorithm (DSA) is proposed. DSA introduces the ray-scanning of LIDAR, a "laser" is released by an agent to detect the nearest object that intersects it. By pre-storing the grids crossed by the rays, the efficiency of DSA is significantly improved. Using the scan results as inputs, an evacuation model has been developed based on the Deep Reinforcement Learning with Double Q-learning (DDQN). The parameters of DSA are calibrated at first, and a group of parameters with a good balance between efficiency and accuracy are recommended. Furthermore, the fundamental diagram is reproduced to calibrate the reward values in DDQN. At last, trajectories and behaviors in the model are studied by using the calibrated parameters. Results show that the movement trajectories are affected by visible distances and reward values, and some effectiveness are consistent with that in experiments. Besides, the exit selection behavior and the lane formation behavior are observed in simulation without introducing any special designed rules. DSA provides a new method to obtain the first-person environmental information in discrete field, and the DSA&DDQN based model defines a new ray-scan-based evacuation modeling scheme. • Propose a Dynamic Scanning Algorithm to obtain the environmental information in discrete features field. • By pre-storing the grids crossed by the rays, the efficiency of DSA is significantly improved. • An evacuation model has been developed based on a reinforcement learning algorithm by using the scan results as inputs. • The fundamental diagram, trajectories and behaviors of specific evacuation scenarios are reproduced by the model. • The DSA&DDQN based model defines a new ray-scan-based evacuation modeling scheme. [ABSTRACT FROM AUTHOR]

In this paper the difference of fundamental diagrams in pedestrian flow is investigated. A new measurement method in which the moving back is considered as negative contribution to pedestrian flow is proposed based on previous study. Firstly the comparison of different measurement methods is made and small discrepancy in the fundamental diagrams of the same experiment is observed. In order to remove the effect of different measurement methods on final results, we compare the fundamental diagrams of Chinese and German experiments by using the same measurement method and surprisingly large difference is found. From the analysis of experimental video, it is observed the motivation and competitiveness of participants in two experiments are quite different, which plays a dominant role on the large difference between fundamental diagrams. To make it more tenable, we further analyze two German experiments (Hermes experiment and BaSiGo experiment) in which the participants have the same average age. The free velocity is adopted to measure pedestrian’s motivation in the experiment and the analysis results confirm that the discrepancy of motivation and competitiveness of participants in different experiments leads to the large difference in the fundamental diagrams. The study may be helpful to understand the effect of pedestrian behavior on global flow and microscopic dynamics. [ABSTRACT FROM AUTHOR]

Pedestrian inflow process frequently occurs in various pedestrian facilities in our daily life. Great importance should be attached to the study of this process. In this paper, we explore the pedestrian inflow process in a room with a separate entrance and exit. Two kinds of experiments are conducted: experiment 1 has no inactive persons and primarily focuses on analyzing the features of the normal pedestrian inflow process and analyzing the representative spatial parameters in the steady state, while experiment 2 involves the influence of the inactive persons. In order to quantitatively discuss the distribution of pedestrians in the steady state, we adopt several analytical methods, such as the Voronoi diagram method, proxemics, and point pattern analysis. Some features of the inflow process are captured. The distribution of pedestrians in the steady state is not uniform. The proxemics and attraction to the exit are both considered to affect pedestrians’ distribution in the inflow process. The presence of inactive persons may have an impact on both the inflow and outflow processes. Practical suggestions are provided for the managers of pedestrian facilities. [ABSTRACT FROM AUTHOR]

The merging flow happens when pedestrians coming from the stair (referred to as stair crowd) encounter those coming from the floor (referred to as floor crowd) on the stair landing, which is usually thought to significantly reduce movement efficiency during building evacuation. However, currently investigations on it are still rare especially the experimental ones with trajectories. In this paper, characteristics of stair landing merging flow are analyzed, based on precise trajectories extracted from the heads through a controlled experiment. Due to the interference of floor crowd, the average path of stair crowd in merging flow experiment is less smooth, compared with that in experiment with no merging flow. Pedestrian total travel time and total travel distance on the stair landing exhibit a linear relation, but the fitting functions are varied depending on the movement directions (ascent and descent), merging conditions (with and without merging flow), and crowd types (floor crowd and stair crowd). The sum of flow rates for stair crowd and floor crowd before merging is larger than the flow rate of overall crowd after merging, due to the occurrence of congestion after merging. Microscopic fundamental diagrams based on Voronoi diagram for different movements are derived and compared. For movements without merging flow, the speed on the stair landing is higher during descent movement. While for movements with merging flow, the speed on the stair landing is higher during ascent movement, which results from the higher speed for ascent floor crowd because ascent floor crowd is not affected by the transition from stair to landing when just stepping onto the stair landing and is less affected by the merging phenomenon. • A controlled experiment on pedestrian movements on the stair landing with and without merging flow is conducted. • The average walking path of stair crowd becomes less smooth with the interference of floor crowd. • Pedestrian total travel time and total travel distance on the stair landing exhibit a linear relation. • The average flow rate after merging is smaller than that before merging due to congestion after merging. • In our experimental setup, ascent floor crowd can move the fastest, leading to higher speed in ascent merging flow. [ABSTRACT FROM AUTHOR]

The evacuation process from a room without visibility is investigated by both experiment and modeling. Some typical characteristics of blind evacuation, including the preference of choosing left-hand side direction and following behavior, are found from the experiment. Meanwhile, different strategies of conflict resolution are observed in the experiment. Based on the experimental observation, a multi-grid model for evacuation without visibility is built in this paper. Simulation results of the model agree well with the experiments. Furthermore, the effect of exit width, number of exits and initial density on evacuation are studied, and results show that exit width has little impact on evacuation time and increasing number of exits is an effective way to decrease evacuation time. Finally, simulations of evacuation under normal and no visibility are compared, and the differences for two conditions are predicted. The comparison results also demonstrate that the blind evacuation is much slower than evacuation under normal visibility, which is match with the practical experience. A similar point is that the distributions of time interval in both situations satisfy power-law relation approximately. The study may be useful for understanding the egress behaviors and developing efficient evacuation strategy and plan to guide pedestrian evacuation without visibility. [ABSTRACT FROM AUTHOR]

In this paper, we conduct a single-file experiment on stairs under two motivations, namely normal walking condition and fast walking condition at various densities. Totally four types of movements are analyzed and compared, namely normal walking-ascending movement, normal walking-descending movement, fast walking-ascending movement and fast walking-descending movement. Compared with the normal walking condition, in fast walking condition pedestrians move with higher velocity, which can be clearly observed in color-coded trajectories and time–space diagrams. We further make pairwise comparisons of density–velocity and headway–velocity relations for four types of movement, and critical parameters such as minimum headway, adaptation time, critical headway and free-flow velocity are extracted. It is found that the adaptation time is negatively correlated with free-flow velocity, indicating that pedestrians under higher movement motivation have a stronger ability to adapt their velocities. Besides, the values of time to collision are smaller for ascending movement than descending movement, apart from those at high density. In addition, we also analyze the following behavior through the relation between velocities of two adjacent pedestrians. The correlation analysis indicates that velocities are most correlated in fast walking-descending movement, and least correlated in normal walking-ascending movement. It is hoped that our study can be useful for the modeling of stair motion and the design of building stairway. [ABSTRACT FROM AUTHOR]

In this paper, we propose a microscopic method to evaluate continuous pedestrian dynamic models at the trajectory level. By comparing the experimental and the simulated trajectory in four directions, the evaluation of the model can be described by a radar chart, with which both qualitative and quantitative conclusions can be obtained. In order to demonstrate our method, we evaluate a social force model by 1,936 trajectories with graded densities in three different scenarios. Three qualitative conclusions are obtained by observing radar charts of the simulation of the unidirectional experiments. All of them are verified by the comparison of the macroscopic parameters. Besides, we find that a model with smaller error in our method always has a better performance at the macroscopic level. At last, the possible quantitative descriptions of the method are discussed. Compared to the evaluations by comparing collective features like fundamental diagram, our method is general, comprehensive and quantitative. The method provides a new possibility to evaluate any continuous pedestrian dynamic model in any scenario with a standard process. • A trajectory level method is proposed to evaluate continues pedestrian models. • A standard process of evaluating a continuous model is included in the method. • The evaluation results of the method can be visualized by a radar chart. • A social force model is evaluated with 1,936 trajectories in three scenarios. • The method is verified by macroscopic level comparisons. [ABSTRACT FROM AUTHOR]