A Diffusion-Model of Joint Interactive Navigation

TL;DR

This paper introduces DJINN, a diffusion-based model for generating realistic and diverse traffic scenarios in autonomous vehicle simulation, enabling flexible, conditioned sampling of joint agent trajectories.

Contribution

The paper presents DJINN, a novel diffusion model that jointly diffuses multiple agent trajectories conditioned on various observations, achieving state-of-the-art performance and flexible scenario generation.

Findings

State-of-the-art performance on joint trajectory metrics

Flexible conditional sampling including goal-based and behavior-class scenarios

Effective generation of diverse, realistic traffic scenarios

Abstract

Simulation of autonomous vehicle systems requires that simulated traffic participants exhibit diverse and realistic behaviors. The use of prerecorded real-world traffic scenarios in simulation ensures realism but the rarity of safety critical events makes large scale collection of driving scenarios expensive. In this paper, we present DJINN - a diffusion based method of generating traffic scenarios. Our approach jointly diffuses the trajectories of all agents, conditioned on a flexible set of state observations from the past, present, or future. On popular trajectory forecasting datasets, we report state of the art performance on joint trajectory metrics. In addition, we demonstrate how DJINN flexibly enables direct test-time sampling from a variety of valuable conditional distributions including goal-based sampling, behavior-class sampling, and scenario editing.