Scaling Planning for Automated Driving using Simplistic Synthetic Data

TL;DR

This paper demonstrates that simplified synthetic data can effectively train deep learning models for automated driving planning, reducing reliance on realistic datasets and enabling targeted scenario generation for improved real-world performance.

Contribution

It introduces a methodology using simplistic simulated data for planning in automated driving, challenging the need for highly realistic simulations or large real-world datasets.

Findings

Reliable driving achieved with lightweight simulated data

Targeted data augmentation improves sim-to-real transfer

Incremental testing identifies and reduces sim-to-real gaps

Abstract

We challenge the perceived consensus that the application of deep learning to solve the automated driving planning task necessarily requires huge amounts of real-world data or highly realistic simulation. Focusing on a roundabout scenario, we show that this requirement can be relaxed in favour of targeted, simplistic simulated data. A benefit is that such data can be easily generated for critical scenarios that are typically underrepresented in realistic datasets. By applying vanilla behavioural cloning almost exclusively to lightweight simulated data, we achieve reliable and comfortable driving in a real-world test vehicle. We leverage an incremental development approach that includes regular in-vehicle testing to identify sim-to-real gaps, targeted data augmentation, and training scenario variations. In addition to a detailed description of the methodology, we share our lessons learned, touching upon scenario generation, simulation features, and evaluation metrics.