Intelligent Tramways through Sense, Learn and React
Semantic segmentation in the rail context for intelligent rail vehicles with machine learning approaches © AIT
Intelligent rail vehicles can increase safety by avoiding collisions and increase cost efficiency through (partial) automation of rolling stock operation. The aim of INTELLiTRAM is to create technology that enables novel assistance systems and, under certain circumstances, automated tram operation. The research approach is to develop modern image-based deep learning concepts and to extend them to the specifics of the tram and railway environment. Through "intelligence" in the recording and interpretation of dynamic traffic scenes, dangerous situations are to be identified even before a potential collision partner is within the structure gauge. The scenario of a driverless operation on factory premises is to be played through as a demonstration.