Project

RealSIM

Period
1 November 2017–31 March 2020
Project manager
Monica Ringvik

Realistic simulation of vehicles for safer, more robust and less expensive development of automated vehicles.

It is very expensive and time consuming to annotate e.g. “real” video data to use it for training of e.g. neural networks. This project aims at developing simulation tools that will make artificial (digitally created) data more realistic (e.g. resembling Swedish roads, cities etc). This shall also include, for example, different weather conditions. Data can then be used for learning intelligent behavior of agents through reinforcement learning and also to generate large amounts of training data for supervised machine learning.

Short facts

Research area
Systems for accident prevention and AD
Financier(s)
VINNOVA/ FFI
Partners
AstaZero
Chalmers University of Technology
Data Intelligence
Wiretronic
Volvo Group
Project type
Project