TraViA is a new visualisation/annotation tool that supports the pNEUMA dataset.
Olger Siebinga, a PhD student from Delft University of Technology, developed TraViA in Python 3 to provide solutions for common problems when working with open datasets. As he reports, “TraViA can be used to visualize and annotate data from highD, pNEUMA, and NGSIM and uses generic vehicle objects to store the state of vehicles at a specific time. This makes it possible to validate and test models or controllers on multiple datasets in parallel, without having to cope with the different dataset formats.”
Currently, Olger is using TraViA for model validation of an inverse-reinforcement- learning-based driver model – pending publication. Besides that, TraViA is used for educational purposes, allowing students at TU Delft to explore big naturalistic datasets by providing them with an accessible, GUI-based starting point.
This work is licensed under a CC BY 4.0.
Software repository: https://github.com/tud-hri/travia