Data-driven simulation of pedestrian collision avoidance with a nonparametric neural network

NEUROCOMPUTING

卷:379

Article

Martin, Rafael F.¹  Parisi, Daniel R.² 

[1] Inst Tecnol Buenos Aires, Lavarden 315,C1437FBG, Ca De Buenos Aires, Argentina

[2] Inst Tecnol Buenos Aires, CONICET, Lavarden 315,C1437FBG, Ca De Buenos Aires, Argentina

关键词: Pedestrian dynamics;    Data-driven simulation;    Navigation;    Steering;    Generalized regression neural network;    Artificial intelligence;   

DOI  :  10.1016/j.neucom.2019.10.062

来源: Elsevier

【 摘 要 】

Data-driven simulation of pedestrian dynamics is an incipient and promising approach for building reliable microscopic pedestrian models. We propose a methodology based on generalized regression neural networks, which does not have to deal with a huge number of free parameters as in the case of multi-layer neural networks. Although the method is general, we focus on the one pedestrian - one obstacle problem. Experimental data were collected in a motion capture laboratory providing high-precision trajectories. The proposed model allows us to simulate the trajectory of a pedestrian avoiding an obstacle from any direction. Together with the methodology specifications, we provide the data set needed for performing the simulations of this kind of pedestrian dynamic system. (C) 2019 Elsevier B.V. All rights reserved.

【 授权许可】

Free