One Thousand and One Hours: Self-driving Motion Prediction Dataset

TL;DR

This paper introduces the largest self-driving motion prediction dataset, with over 1,000 hours of data, enabling significant advancements in machine learning tasks for autonomous vehicles.

Contribution

It provides the largest and most detailed dataset for self-driving motion prediction, including extensive scene data, semantic maps, and aerial views, to facilitate research and development.

Findings

Dataset significantly improves performance in self-driving motion prediction tasks.

Large-scale data enhances the accuracy of motion forecasting models.

The dataset supports development of motion planning and simulation tools.

Abstract

Motivated by the impact of large-scale datasets on ML systems we present the largest self-driving dataset for motion prediction to date, containing over 1,000 hours of data. This was collected by a fleet of 20 autonomous vehicles along a fixed route in Palo Alto, California, over a four-month period. It consists of 170,000 scenes, where each scene is 25 seconds long and captures the perception output of the self-driving system, which encodes the precise positions and motions of nearby vehicles, cyclists, and pedestrians over time. On top of this, the dataset contains a high-definition semantic map with 15,242 labelled elements and a high-definition aerial view over the area. We show that using a dataset of this size dramatically improves performance for key self-driving problems. Combined with the provided software kit, this collection forms the largest and most detailed dataset to date for the development of self-driving machine learning tasks, such as motion forecasting, motion planning and simulation. The full dataset is available at http://level5.lyft.com/.