STOchastic Traffic NEtworks (STONE)
Understanding efficiency and behavior aspects in partially automated (vehicular) technology in large-scale (traffic) context is an unsolved problem nowadays. Our main goal is to develop learning methods for uncertain traffic networks. We will rely on interdisciplinary approaches between i) mathematical sciences (stochastic Partial Differential Equations sPDE) ii) traffic flow theory (hyperbolic conservation vehicular laws, network efficiency) and iii) traffic safety analysis (driver behavior, collision probability) glued by probabilistic machine leaning concept. We propose to to learn the 1) parameter variation and 2) solution of sPDE network models. Whilst learning parameters will deliver a probabilistic picture of partially automated traffic networks in terms of congestions, driver behavior, network capacity; learning the solutions of such models enables us to short term predict traffic network behavior.
This page is only available in english
- Chalmers AI Research Centre (CHAIR) (Centre, Sweden)
- Chalmers (Site, Sweden)
- Chalmers (Publisher, Sweden)