Ana Čukarska and Felix Dietrich collaborated with colleagues from the TUM School of Engineering and Design Patrick Berggold, Stavros Nousias and André Borrmann, in the scope of the FORWARD Project funded by TUM GNI and produced a systematic literature review paper on machine learning methods in pedestrian and evacuation dynamics. We highlight the importance of this new area of research and identify the current challenges and open questions that remain in the field. Two modeling trends became apparent: modeling individual human trajectories, and modelling collective behavior; most research so far has predominantly focused on individual trajectories, and used methods from the families of deep learning (LSTM, Transformers) and reinforcement learning.
Please have a look at round paper: Machine learning in pedestrian and evacuation dynamics for the built environment: A systematic literature review
DOI: https://doi.org/10.1016/j.ssci.2026.107143