Rolling Ahead Diffusion for Traffic Scene Simulation

TL;DR

This paper introduces a rolling diffusion model for traffic scene simulation that balances reactive behavior and computational efficiency by predicting immediate and partially noised future steps simultaneously.

Contribution

The paper proposes a novel rolling diffusion approach that combines autoregressive and MPC-like methods for efficient, reactive traffic scenario generation.

Findings

Achieves faster simulation than traditional diffusion AR models.

Maintains reactive capabilities comparable to MPC-based methods.

Offers a computationally efficient compromise for traffic scene simulation.

Abstract

Realistic driving simulation requires that NPCs not only mimic natural driving behaviors but also react to the behavior of other simulated agents. Recent developments in diffusion-based scenario generation focus on creating diverse and realistic traffic scenarios by jointly modelling the motion of all the agents in the scene. However, these traffic scenarios do not react when the motion of agents deviates from their modelled trajectories. For example, the ego-agent can be controlled by a stand along motion planner. To produce reactive scenarios with joint scenario models, the model must regenerate the scenario at each timestep based on new observations in a Model Predictive Control (MPC) fashion. Although reactive, this method is time-consuming, as one complete possible future for all NPCs is generated per simulation step. Alternatively, one can utilize an autoregressive model (AR) to predict only the immediate next-step future for all NPCs. Although faster, this method lacks the capability for advanced planning. We present a rolling diffusion based traffic scene generation model which mixes the benefits of both methods by predicting the next step future and simultaneously predicting partially noised further future steps at the same time. We show that such model is efficient compared to diffusion model based AR, achieving a beneficial compromise between reactivity and computational efficiency.