Trajectory Prediction for Autonomous Driving Using a Transformer Network

TL;DR

This paper presents a transformer-based multi-modal trajectory prediction framework for autonomous driving, utilizing semantic maps and an auxiliary loss to improve accuracy and feasibility of predicted agent trajectories.

Contribution

Introduces a novel transformer-based model with semantic map inputs and an auxiliary loss for more accurate and feasible trajectory predictions in autonomous driving.

Findings

Achieves state-of-the-art performance on Lyft l5kit dataset.

Significantly improves prediction accuracy.

Enhances trajectory feasibility by penalizing off-road predictions.

Abstract

Predicting the trajectories of surrounding agents is still considered one of the most challenging tasks for autonomous driving. In this paper, we introduce a multi-modal trajectory prediction framework based on the transformer network. The semantic maps of each agent are used as inputs to convolutional networks to automatically derive relevant contextual information. A novel auxiliary loss that penalizes unfeasible off-road predictions is also proposed in this study. Experiments on the Lyft l5kit dataset show that the proposed model achieves state-of-the-art performance, substantially improving the accuracy and feasibility of the prediction outcomes.