Moving Horizon Estimation of Pedestrian Interactions Based on Multiple Velocity Fields
This paper describes a model for the interaction between pedestrians in a scene. We consider a recently proposed model for isolated pedestrians in the scene and extend it by adding an interaction term that accounts for attractive/repulsive behaviors among pedestrians. The proposed model combines multiple velocity fields that represent typical motion regimes in the scene and a time-varying interaction term. The estimation of the active velocity field and interaction parameters is achieved by assuming that they remain constant within every instance of an analysis window that slides in time. This strategy is known as the Moving Horizon Estimation (MHE) method. The proposed algorithm is assessed both by using synthetic data and pedestrian trajectories extracted from video streams.
Unable to display preview. Download preview PDF.
- 5.Nascimento, J., Marques, J., Figueiredo, M.A.T.: Classification of complex pedestrian activities from trajectories. In: IEEE Int. Conf. Image Processing, pp. 3481–3484 (September 2010)Google Scholar
- 6.Helbing, D.: A mathematical model for the behavior of pedestrians. Behavioral Science (36), 298–310 (1991)Google Scholar
- 7.Helbing, D., Molnar, P.: Social force model for pedestrian dynamics. Physical Review (E51), 4282–4286 (1995)Google Scholar
- 8.Mehran, R., Oyama, A., Shah, M.: Abnormal crowd behavior detection using social force model. In: IEEE Int. Conf. on Computer Vision and Pattern Recognition, pp. 935–942 (June 2009)Google Scholar
- 9.Zhang, Y., Qin, L., Yao, H., Huang, Q.: Abnormal crowd behavior detection based on social attribute-aware force model. In: IEEE Int. Conf. Image Processing, pp. 2689–2692 (September 2012)Google Scholar