Improving Autonomous Driving Safety with POP: A Framework for Accurate Partially Observed Trajectory Predictions

TL;DR

This paper introduces POP, a novel framework for accurate trajectory prediction in autonomous driving under partial observations, combining self-supervised learning and feature distillation to improve safety and performance.

Contribution

The paper presents a new trajectory prediction framework that effectively handles partial observations using self-supervised learning and feature distillation, outperforming baseline methods.

Findings

POP achieves comparable results to top methods in open-loop tests.

POP outperforms baseline in closed-loop safety simulations.

Qualitative results show improved safety and reasonableness in predictions.

Abstract

Accurate trajectory prediction is crucial for safe and efficient autonomous driving, but handling partial observations presents significant challenges. To address this, we propose a novel trajectory prediction framework called Partial Observations Prediction (POP) for congested urban road scenarios. The framework consists of two key stages: self-supervised learning (SSL) and feature distillation. POP first employs SLL to help the model learn to reconstruct history representations, and then utilizes feature distillation as the fine-tuning task to transfer knowledge from the teacher model, which has been pre-trained with complete observations, to the student model, which has only few observations. POP achieves comparable results to top-performing methods in open-loop experiments and outperforms the baseline method in closed-loop simulations, including safety metrics. Qualitative results illustrate the superiority of POP in providing reasonable and safe trajectory predictions.