FloMo: Tractable Motion Prediction with Normalizing Flows
The future motion of traffic participants is inherently uncertain. To plan safely, therefore, an autonomous agent must take into account multiple possible trajectory outcomes and prioritize them. Recently, this problem has been addressed with generative neural networks. However, most generative models either do not learn the true underlying trajectory distribution reliably, or do not allow predictions to be associated with likelihoods. In our work, we model motion prediction directly as a density estimation problem with a normalizing flow between a noise distribution and the future motion distribution. Our model, named FloMo, allows likelihoods to be computed in a single network pass and can be trained directly with maximum likelihood estimation. Furthermore, we propose a method to stabilize training flows on trajectory datasets and a new data augmentation transformation that improves the performance and generalization of our model. Our method achieves state-of-the-art performance on three popular prediction datasets, with a significant gap to most competing models.
FURTHER CURRENT TOPICS
Cognition Factory: Evaluate and visualize camera data
Since the beginning of research on the digital twin, AI specialist Cognition Factory GmbH has focused on processing camera data. In the meantime Dr. Claus Lenz has deployed a large-scale platform
Digital real-time twin of traffic: ready for series production
Expand the test track, deploy new sensors, decentralize software architecture, fuse sensor data for 24/7 operation of a real-time digital twin, and make data packets public: TU Munich has decisively advanced the Providentia++ research project.
Elektrobit: Coining Test Lab to stationary data
Elektrobit lays the foundation for Big Data evaluations of traffic data. Simon Tiedemann on the developments in P++.