FORWARD
Pedestrian dynamics prediction for safe and flow-efficient building design
Principal Investigators
Prof. Dr.-Ing. André Borrmann
Chair of Computational Modeling and Simulation, TUM School of Engineering and Design
Prof. Dr. rer. nat. Felix Dietrich
Professor for Physics-enhanced Machine Learning, TUM School of Computation, Information and Technology
Project summary
Project FORWARD aims to automate the prediction of pedestrian flow in indoor and outdoor spaces allowing architects and engineers to consider and modify bottlenecks when designing buildings and infrastructure facilities, such as train stations or airports. While pedestrian simulation has been investigated for a long time and commercial implementations exist, they are compute-intensive and thus require long execution time. To allow building designers to interactively explore different design options and assess their suitability in terms of pedestrian flow, the FORWARD project will employ Machine Learning (ML) methods to perform predictions that are instantly available and provide suitable accuracy. The project will investigate a variety of different ML models to identify the best performing ones. For training ML models with a broad spectrum of building designs, parametric building modelling will be exploited extensively. The cross-fertilization of parametric model design and machine learning methods will allow the research consortium to efficiently predict the pedestrian flow in building models, to identify bottlenecks and evaluate their capacity to handle pedestrian traffic, estimate maximum acceptable flow, and evacuation times.