Vehicle Behavior Prediction Using Deep Learning
Title: Vehicle Behavior Prediction Using Deep Learning
DNr: Berzelius-2024-206
Project Type: LiU Berzelius
Principal Investigator: Erik Frisk <>
Affiliation: Linköpings universitet
Duration: 2024-05-15 – 2024-09-01
Classification: 20201


A pivotal aspect of advancing self-driving vehicles is equipping them with the ability to accurately forecast the behavior (motion and intention) of surround- ing road users. However, merely making accurate predictions is not sufficient. To enable safe and robust decision-making processes, it is crucial that these methods also effectively handle the multimodality and uncertainty in forecasted behaviors. As a natural consequence, probabilistic modeling has become a cor- nerstone in devising robust trajectory prediction methods. In this project, we seek to explore the potential of generative modeling for trajectory prediction applications.