CellEVAC: An adaptive guidance system for crowd evacuation through behavioral optimization

Abstract

A critical aspect of crowds' evacuation processes is the dynamism of individual decision making. Identifying optimal strategies at an individual level may improve both evacuation time and safety, which is essential for developing efficient evacuation systems. Here, we investigate how to favor a coordinated group dynamic through optimal exit-choice instructions using behavioral strategy optimization. We propose and evaluate an adaptive guidance system (Cell-based Crowd Evacuation, CellEVAC) that dynamically allocates colors to cells in a cell-based pedestrian positioning infrastructure, to provide efficient exit-choice indications. The operational module of CellEVAC implements an optimized discrete-choice model that integrates the influential factors that would make evacuees adapt their exit choice. To optimize the model, we used a simulation-optimization modeling framework that integrates microscopic pedestrian simulation based on the classical Social Force Model. In the majority of studies, the objective has been to optimize evacuation time. In contrast, we paid particular attention to safety by using Pedestrian Fundamental Diagrams that model the dynamics of the exit gates. CellEVAC has been tested in a simulated real scenario (Madrid Arena) under different external pedestrian flow patterns that simulate complex pedestrian interactions. Results showed that CellEVAC outperforms evacuation processes in which the system is not used, with an exponential improvement as interactions become complex. We compared our system with an existing approach based on Cartesian Genetic Programming. Our system exhibited a better overall performance in terms of safety, evacuation time, and the number of revisions of exit-choice decisions. Further analyses also revealed that Cartesian Genetic Programming generates less natural pedestrian reactions and movements than CellEVAC. The fact that the decision logic module is built upon a behavioral model seems to favor a more natural and effective response. We also found that our proposal has a positive influence on evacuations even for a low compliance rate (40%).

LED wristbands for Cell-based Crowd Evacuation: an Adaptive Exit-choice Guidance System Architecture

“…Thus, there are research efforts in the area of real-time guidance for crowd evacuations that have focused on studying mechanisms for providing pedestrians with optimal exit-choice information. A promising line of research in this area is that of cell-based evacuation systems [1,22,23]. These systems rely on a cell-based pedestrian positioning infrastructure such that pedestrians in a cell are assumed to receive the same exit gate instructions.…”

“…To solve this problem, the authors adopt the Cartesian Genetic Programming (CGP) [25]. We developed in [23] an adaptive cell-based crowd evacuation system (CellEVAC) based on behavioral optimization that searches for the optimal evacuation plan through meta-heuristic optimization methodology. As in [22], we obtain adaptive evacuation plans capable of responding to changing environmental conditions.…”

“…With the purposes mentioned above, this paper proposes a system architecture for our adaptive cell-based evacuation system CellEVAC [23]. The proposed system architecture consists of three subsystems: (i) monitoring and control logic subsystem, (ii) active RFID cell-node network, and (iii) radio-controlled LED wristband subsystem.…”

“…In [23], we assumed an error-free positioning infrastructure, where pedestrians were supposed to be equipped with a generic device (smart-phone or wearable device) with ideal location-aware and color display capabilities. In this paper, the proposed positioning system's RFID communication channels are modeled using a log-normal propagation model.…”

“…To perform simulation and simulation-optimization experiments, we have used the simulation-optimization modeling framework that we developed in [23]. This framework integrates a microscopic pedestrian simulation based on the classical Social Force Model (SFM) [26].…”