A Diffusion-Model of Joint Interactive Navigation
This addresses the need for scalable and realistic traffic simulation in autonomous driving, though it appears incremental as an application of diffusion models to a known bottleneck in scenario generation.
The paper tackles the problem of generating diverse and realistic traffic scenarios for autonomous vehicle simulation by introducing DJINN, a diffusion-based method that jointly diffuses trajectories of all agents with flexible conditioning, achieving state-of-the-art performance on trajectory forecasting datasets.
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.