Traffic evacuation simulation based on multi-level driving decision model.
Traffic evacuation is a critical task in disaster management. Planning its evacuation in advance requires taking many factors into consideration such as the destination shelter locations and numbers, the number of vehicles to clear, the traffic congestions as well as traffic road configurations. A traffic evacuation simulation tool can provide the emergency managers with the flexibility of exploring various scenarios for identifying more accurate model to plan their evacuation. This paper presents a traffic evacuation simulation system based on integrated multi-level driving-decision models which generate agents behavior in a unified framework. In this framework, each agent undergoes a Strategic, Cognitive, Tactical and Operational (SCTO) decision process, in order to make a driving decision. An agent actions are determined by a combination, on each process level, of various existing behavior models widely used in different driving simulation models. A wide spectrum of variability in each agent decision and driving behaviors, such as in pre-evacuation activities, in choice of route, and in the following or overtaking the car ahead, are represented in the SCTO decision process models to simulate various scenarios. We present the formal model for the agent and the multi-level decision models. A prototype simulation system that reflects the multi-level driving-decision process modeling is developed and implemented. Our SCTO framework is validated by comparing with MATSim tool, and the experimental results of evacuation simulation models are compared with the existing evacuation plan for densely populated Beijing, China in terms of various performance metrics. Our simulation system shows promising results to support emergency managers in designing and evaluating more realistic traffic evacuation plans with multi-level agent decision models that reflect different levels of individual variability of handling stress situations. The flexible combination of existing behavior and decision models can help generating the best evacuation plan to manage each crisis with unique characteristics, rather than resorting to a fixed evacuation plan. [ABSTRACT FROM AUTHOR] Copyright of Transportation Research: Part C is the property of Pergamon Press - An Imprint of Elsevier Science and its content may not be copied or emailed to multiple sites or posted to a listserv without the copyright holder's express written permission. However, users may print, download, or email articles for individual use. This abstract may be abridged. No warranty is given about the accuracy of the copy. Users should refer to the original published version of the material for the full abstract. (Copyright applies to all Abstracts.)
Statistical analysis of the data collected by running the simulation for the framework validation and experimental evaluation. Furthermore, the data was compared with MATSim.
Designing and implementing a traffic simulation system in order to test it under various emergency scenarios. Afterwards, the framework was validated by comparing the results of their system with a popular traffic simulation tool MATSim that has been effectively applied for evacuation modeling. Then, an experimental evaluation was done.
Clearance time, 2) total evacuation rates and mean evacuation rates per shelter, 3) panic population, 4) number of jammed intersections
The plan was to present a formal model of agent-based driving behavior using multi-level decision process that reuses and flexibly combines existing behavior models, including evacuee’s decisions, route choices, and traffic flow models. Driving variability of each agent is modeled through different decisions making perspectives from each layer. No deviations reported.
Data was collected by running the simulation for the framework validation and experimental evaluation. GIS Data was acquired from OSM.
The authors have designed and implemented a traffic simulation system based on the mentioned framework and have tested it under various emergency scenarios in order to identify traffic jams locations and to help develop strategies for selecting shelters.
The authors plan to integrate additional factors to better simulate the outcomes of evacuation-related decisions. Such as varying types of evacuation agents, including vehicles, bicycles and pedestrians.The authors will also study how psychological characteristics affect drivers’ behaviors, especially how such affects might be integrated into an agent’s decision model in which evacuation rates can be estimated with varying proportions of agents failing to follow an evacuation plan.
The simulation system shows promising results to support emergency managers in designing and evaluating more realistic traffic evacuation plans with mutli-level agent’s decision models that reflect different levels of individual variability of handling stress situations.When multiple models and behavior patterns were turned off, the results are comparable to those of the MATSimThe results also show that using the additional route choice model, i.e., potential network model, reduces the agent’s travel time from that of the shortest path modelThe results also showed a significant impact of the nervousness value on the clearance time. When evacuees are extremely panicked, the clearance time increased rapidly.The experimental results show that it is beneficial to integrate more variability of driving behaviors to a model frameworkThe results also hos that while the buildings of more shelters may increase overall evacuation rates. It may diminish efficient utilization of sheltersThe approach would provide emergency officials with a clearer view, enabling them to visualize outcomes of the possible alternative evacuation plans under various possible scenarios, and to more accurately predict and plan clearance time so that they can choose the optimal plan.
This paper presents a traffic evacuation simulation system based on integrated multi-level driving-decision models which generate agents’ behavior in a united framework. The proposed integrated simulation system includes four key components: geospatial manager, agent manager, behavior manager, and mission manager. Together they provide the capability to simulate complex scenarios.
 |
Die Portfolio of Solutions Website wurde ursprünglich im Rahmen des DRIVER+ Projekts entwickelt. Heute wird das Service von der AIT Austrian Institute of Technology GmbH. zum Nutzen des europäischen Krisenmanagements betrieben. PoS ist vom Disaster Competence Network Austria (DCNA) sowie von den H2020-Projekten STAMINA und TeamAware befürwortet und unterstützt. |