Improving Out-of-Distribution Generalization of Trajectory Prediction for Autonomous Driving via Polynomial Representations

TL;DR

This paper introduces a polynomial-based trajectory prediction method that enhances out-of-distribution robustness in autonomous driving, demonstrating improved generalization and efficiency over existing models.

Contribution

The paper proposes a novel polynomial representation approach for trajectory prediction that improves OoD robustness and reduces model complexity and training effort.

Findings

Achieves near state-of-the-art ID performance with smaller models.

Significantly improves OoD robustness compared to existing methods.

Highlights the importance of OoD testing in model evaluation.

Abstract

Robustness against Out-of-Distribution (OoD) samples is a key performance indicator of a trajectory prediction model. However, the development and ranking of state-of-the-art (SotA) models are driven by their In-Distribution (ID) performance on individual competition datasets. We present an OoD testing protocol that homogenizes datasets and prediction tasks across two large-scale motion datasets. We introduce a novel prediction algorithm based on polynomial representations for agent trajectory and road geometry on both the input and output sides of the model. With a much smaller model size, training effort, and inference time, we reach near SotA performance for ID testing and significantly improve robustness in OoD testing. Within our OoD testing protocol, we further study two augmentation strategies of SotA models and their effects on model generalization. Highlighting the contrast between ID and OoD performance, we suggest adding OoD testing to the evaluation criteria of trajectory prediction models.