Transfer Learning Study of Motion Transformer-based Trajectory Predictions

Abstract

Trajectory planning in autonomous driving is highly dependent on predicting the emergent behavior of other road users. Learning-based methods are currently showing impressive results in simulation-based challenges, with transformer-based architectures technologically leading the way. Ultimately, however, predictions are needed in the real world. In addition to the shifts from simulation to the real world, many vehicle- and country-specific shifts, i.e. differences in sensor systems, fusion and perception algorithms as well as traffic rules and laws, are on the agenda. Since models that can cover all system setups and design domains at once are not yet foreseeable, model adaptation plays a central role. Therefore, a simulation-based study on transfer learning techniques is conducted on basis of a transformer-based model. Furthermore, the study aims to provide insights into possible trade-offs between computational time and performance to support effective transfers into the real world.

Published in 2024 IEEE Intelligent Vehicles Symposium (IV), Jeju Shinhwa World, Jeju Island, Korea, June 2-5, 2024

References in corpus (9)

• Adam: A Method for Stochastic Optimization
• Deep Visual Domain Adaptation: A Survey
• Domain Generalization: A Survey
• One weird trick for parallelizing convolutional neural networks
• Transfusion: Understanding Transfer Learning for Medical Imaging
• Multi-task learning for natural language processing in the 2020s: where are we going?
• Motion Transformer with Global Intention Localization and Local Movement Refinement
• Shifts 2.0: Extending The Dataset of Real Distributional Shifts
• MGTR: Multi-Granular Transformer for Motion Prediction with LiDAR