TrajDiffuse: A Conditional Diffusion Model for Environment-Aware Trajectory Prediction

TL;DR

TrajDiffuse is a novel diffusion-based model for environment-aware trajectory prediction that produces diverse, accurate, and collision-free trajectories by integrating map-based guidance into the denoising process.

Contribution

The paper introduces TrajDiffuse, a planning-based conditional diffusion model that effectively incorporates environmental constraints into trajectory prediction.

Findings

Matches or exceeds state-of-the-art accuracy and diversity.

Generates trajectories that adhere closely to environmental constraints.

Demonstrates superior performance on nuScenes and PFSD datasets.

Abstract

Accurate prediction of human or vehicle trajectories with good diversity that captures their stochastic nature is an essential task for many applications. However, many trajectory prediction models produce unreasonable trajectory samples that focus on improving diversity or accuracy while neglecting other key requirements, such as collision avoidance with the surrounding environment. In this work, we propose TrajDiffuse, a planning-based trajectory prediction method using a novel guided conditional diffusion model. We form the trajectory prediction problem as a denoising impaint task and design a map-based guidance term for the diffusion process. TrajDiffuse is able to generate trajectory predictions that match or exceed the accuracy and diversity of the SOTA, while adhering almost perfectly to environmental constraints. We demonstrate the utility of our model through experiments on the nuScenes and PFSD datasets and provide an extensive benchmark analysis against the SOTA methods.