What-If Motion Prediction for Autonomous Driving

TL;DR

This paper introduces a recurrent graph-based model for long-term motion prediction in autonomous driving, capable of incorporating hypothetical scenarios and social interactions to improve planning and safety.

Contribution

It presents a novel interpretable attention-based approach that supports counterfactual reasoning with geometric and social context integration.

Findings

Supports counterfactual geometric goals and social interactions

Produces diverse, scenario-conditioned predictions

Enhances planning by reasoning about unlikely futures

Abstract

Forecasting the long-term future motion of road actors is a core challenge to the deployment of safe autonomous vehicles (AVs). Viable solutions must account for both the static geometric context, such as road lanes, and dynamic social interactions arising from multiple actors. While recent deep architectures have achieved state-of-the-art performance on distance-based forecasting metrics, these approaches produce forecasts that are predicted without regard to the AV's intended motion plan. In contrast, we propose a recurrent graph-based attentional approach with interpretable geometric (actor-lane) and social (actor-actor) relationships that supports the injection of counterfactual geometric goals and social contexts. Our model can produce diverse predictions conditioned on hypothetical or "what-if" road lanes and multi-actor interactions. We show that such an approach could be used in the planning loop to reason about unobserved causes or unlikely futures that are directly relevant to the AV's intended route.