Relation-based Motion Prediction using Traffic Scene Graphs

TL;DR

This paper introduces a novel traffic scene graph representation using semantic relations and employs Graph Neural Networks to improve motion prediction accuracy for autonomous driving, notably enhancing acceleration prediction by up to 12%.

Contribution

It presents a new approach to model traffic scenes with spatial semantic graphs and demonstrates that explicit relation modeling improves prediction performance.

Findings

Acceleration prediction improved by up to 12%.

Including previous scene data yields 73% performance boost.

Explicit relation modeling enhances traffic participant predictions.

Abstract

Representing relevant information of a traffic scene and understanding its environment is crucial for the success of autonomous driving. Modeling the surrounding of an autonomous car using semantic relations, i.e., how different traffic participants relate in the context of traffic rule based behaviors, is hardly been considered in previous work. This stems from the fact that these relations are hard to extract from real-world traffic scenes. In this work, we model traffic scenes in a form of spatial semantic scene graphs for various different predictions about the traffic participants, e.g., acceleration and deceleration. Our learning and inference approach uses Graph Neural Networks (GNNs) and shows that incorporating explicit information about the spatial semantic relations between traffic participants improves the predicdtion results. Specifically, the acceleration prediction of traffic participants is improved by up to 12% compared to the baselines, which do not exploit this explicit information. Furthermore, by including additional information about previous scenes, we achieve 73% improvements.