Kashif Zia

LifeBelt: Silent Directional Guidance for Crowd Evacuation

The effectiveness of an evacuation process in an emergency situation is heavily dependent on the understanding of the dynamics of crowds. Evidence has been delivered supporting the hypothesis that crowd behavior is self referential in the sense that movement patterns of individuals constitutes crowd movement, which in turn impacts the behavior of individuals. In this paper we build on this evidence and propose a belt like wearable device for vibro tactile directional guidance, LifeBelt, to notify individuals in panic about exits. To assess the potential improvement in evacuation efficiency when using the LifeBelt instead of when not, we empirically analyze the behavior of individuals in panic on the microscopic level, and parameterize large scale simulations (of up to 2000 individuals) with this evidence on the macroscopic level. Simulations show that LifeBelt based guidance can damp panic growth and increase the number of successful evacuations per unit time.

LifeBelt: Crowd Evacuation Based on Vibro-Tactile Guidance

To increase crowd evacuation efficiency in emergency situations, the authors developed a wearable computer, LifeBelt, to guide individuals from emergency situations to escape. The LifeBelt collects position and orientation information about individuals from embedded sensors, shares this information with a background emergency coordination system, receives an individualized optimized escape strategy, and guides the individual via vibro-tactile stimulation, indicating direction and distance to the recommended exit. Simulation models indicate that the number of successful evacuations per unit time improve significantly using the LifeBelt technology.

On the Efficiency of LifeBelt Based Crowd Evacuation

To support the evacuation process of crowds from emergency situations, we have developed a wearable device, LifeBelt, for vibro tactile guidance of individuals in panic towards exits. Since visual and auditory perception is overwhelmed in situations of panic, exit sign and loudspeaker based evacuation often appears ineffective. LifeBelt as a coordinated navigation device builds on a subtle directionalvibration stimulation to navigate and guide individuals to escape.This paper compares the effectiveness of a LifeBelt supported vs. non-supported evacuation from spaces with multiple exits. Based on an extended Moore’s model of neighborhood for next step behavior, and a predicted shortest time to escape metric, we investigate on the efficiency effect of (i) individuals following the nearest exit (without LifeBelt), and (ii) individuals adheringLifeBelt recommendations towards the earliest escape exit. Large scale simulations show that even in situations of growing panic, LifeBelt improves evacuation efficiency by more than 34,5 %.

Grouping behaviour in AmI-enabled crowd evacuation

Grouping behaviour occurs often in crowd evacuation. On the one hand, groups are needed for efficient evacuation. On the other hand, large uncontrolled groups (herds) may cause clogging and increase panic. The mechanisms of emergence of leaders and groups in complex socio-technical systems with intelligent technical components are not well understood. This paper presents the first attempt to unveil the role of AmI technology in formation of spontaneous groups in crowd evacuation. To this end several hypotheses were formulated, which were tested by simulation experiments based on a cognitive agent model. The checking of the hypotheses was done in the context of a train station evacuation scenario. The general outcome is that in a system with scarce and uncertain information, AmI technology can be used to stimulate emergence of leaders and groups to increase the efficiency of evacuation. Furthermore, a large penetration rate of ambient devices may be unnecessary and even not appropriate for fluent evacuation.

An Agent-Based Parallel Geo-Simulation of Urban Mobility during City-scale Evacuation

The simulation of urban mobility is a modeling challenge due to the complexity and scale. The complexity in modeling a social agent is due to three reasons: (i) the agent is behaviorally complex itself due to several interrelated/overlapping modeling aspects; (ii) the setting in which a social agent operates usually demands a multi-resolution approach; and (iii) the consideration of real spatial and population data is the underpinning that has to be realized. In this paper, we propose an agent-based parallel geo-simulation framework of urban mobility based on necessary modeling aspects. The aspect-oriented modeling paradigm relates the models vertically as well as horizontally and highlights the situations requiring multi-resolution interfacing. The framework takes into consideration the importance of technological footprints embedded with social behavior along with essential space and mobility features keeping focus on the importance of the city-scale scenario. We have used a real, high-quality raster map of a medium-sized city in central Europe converting it into a cellular automata (CA). The fine-grained CA readily supports pedestrian mobility and can easily be extended to support other mobility modes. The urban mobility simulation is performed on a real parallel and distributed hardware platform using a CA compatible software platform. Considering city-wide mobility in an emergency scenario, an analysis of the simulation efficiency and agent behavioral response is presented.

Potential of Social Modelling in Socio-Technical Systems

Abstract While modelling (and simulating) a socio-technical system, the primary focus has been on the technological side almost ignoring the social dimension of it. However, with recent advancements in Ambient Intelligence (AmI) and emotions sensing technologies, it is now possible to fill this gap. To this end, we have integrated cognitive decision making model abstracted from psychological, neurological and social theories of human behaviour in evacuation situations into a simulated environment, assisted by AmI technology. Initial simulation findings are reported in this paper, keeping focus on a scenario in which a small population of agents is technologically assisted. The results are: (i) the technologically assisted agents emerge as leaders during evacuation changing the intentions of most of the agents, and (ii) even a small population of such leaders is sufficient to guarantee an efficient evacuation due to leader following behaviour emerging as a social phenomenon.

Collective Attention through Public Displays

The dynamics of collective attention emerging out of individual viewing experiences from public displays appear to be among the most demanding challenges in understanding the mechanisms of self-adaptation of public opinion. In this paper we approach a model of collective attention from observations of the attention of individuals estimated from their efforts expressing interest. Extending on SEEV, an established individual attention model from cognitive science, attention estimates from spontaneous passer-bys in front of public displays are used to describe a collective attention model at the scale of society. The model is validated via a large scale simulation experiment reflecting the demographics and the morphology of a whole city, together with population densities, mobility patterns and individual decision making on a 2048 node shared memory multiprocessor (SGI Altix Ultra Violet 1000, Repast HPC). Simulations how collective attention emerges from local spots of attention towards city scale opinion building and consensus finding.

Emotional Decision Making in Large Crowds

Currently it is widely recognised that emotions of people influence their decisions. In this paper the role of emotions in social decision making in large technically assisted crowds is investigated. For this a formal, computational model is proposed, which integrates existing neurological and cognitive theories of affective decision making. Based on this model several variants of a large scale crowd evacuation scenario with technically assisted agents were simulated. By analysis of the simulation results it was established that spread of emotions in a crowd increases resistance of agent groups to opinion changes and supports continuity of decision making in a group with technically assisted agents.

Scenario Based Modeling for Very Large Scale Simulations

In order to develop complexity science based modeling, prediction and simulation methods for large scale socio-technical systems in an Ambient Intelligence (AmI) based smart environment, we propose a scenario based modeling approach. With a case study on AmI technology to support the evacuation from emergency scenarios, i.e. the Life Belt, a wearable computing systems for vibro-tactile directional guidance, we introduce the concept of model scaling from a micro to a macro level. Aligned with the scenario, we present how crowd simulation strategies encoded into a small scale simulation setup can be extended to a mixed-level simulation based on combining model aspects also coming from the large scale model. The experimental results of a real evacuation trail at a local railway station are incorporated to compare the evacuation efficiency for three strategies: (i) Potential Map, (ii) Evacuees familiarity of the exits and (iii) Exits usage optimization. A comparison with the earlier results from small scale simulation suggest that a real large scale simulation results may not be similar to that of small scale simulation due to dynamics of crowd built up and complexity of building structure.

Modular Simulation-Based Physical and Emotional Assessment of Ambient Intelligence in Traffic

In this letter, we detail a modular approach for measuring the secondary physical and emotional effects of ambient intelligence (AmI) technology in traffic. Using the case of merges on to a highway, we assess the results of a system that advises the driver to change early to a lane on the left to create space for merging cars downstream (tested using a cellular automata simulation). The indirect impact of the system downstream, namely how the remaining lane changes from the merge lane to the innermost lane proceed, is then evaluated using a time-discrete, space-continuous microscopic traffic simulation tool. This yields detailed results concerning driver interactions that can also be used to derive an estimate of driver anger in the situation. We have used real geographic, traffic and psychological data to test the system, and different models are used to accomplish various tasks. The approach yields (surprisingly) negative results concerning the indirect emotional impact of this AmI intervention which may be due to the nature of the lane changing model used and the chosen parameters. We argue that such an approach is also applicable to similar types of systems, where different data and model types are suited to different scenario elements.