Ensemble of pre-trained models for long-tailed trajectory prediction

Thuremella, Divya; Yang, Yi; Wanna, Simon; Kunze, Lars; De Martini, Daniele

Abstract

This work explores the application of ensemble modeling to the multidimensional regression problem of tra-jectory prediction for vehicles in urban environments. We show how, perhaps surprisingly, combining state-of-the-art deep learning models out-of-the-box (without retraining or fine-tuning) with a simple weighted average method can enhance the overall prediction. Indeed, while in general combining trajectory prediction models is not straightforward, this simple approach enhances performance by 10% over the best prediction model, especially in the long-tailed metrics. We show that this performance improvement holds on both the NuScenes and Argoverse datasets, and that these improvements are made across the dataset distribution.

• Presentation Conference Type: Conference Paper (unpublished)
• Conference Name: 2025 IEEE 28th International Conference on Intelligent Transportation Systems (ITSC 2025)
• Start Date: Nov 18, 2025
• End Date: Nov 21, 2025
• Acceptance Date: Jul 1, 2025
• Deposit Date: Aug 22, 2025
• Peer Reviewed: Peer Reviewed
• Public URL: https://uwe-repository.worktribe.com/output/14832567
• Other Repo URL: https://ora.ox.ac.uk/objects/uuid:db33f2b6-cf1a-40ad-9da3-7a89b37d4dc9