AGMA: Adaptive Gaussian Mixture Anchors for Prior-Guided Multimodal Human Trajectory Forecasting

TL;DR

AGMA introduces a scene-adaptive prior modeling approach using Gaussian mixture anchors, significantly improving multimodal human trajectory prediction accuracy and diversity across multiple datasets.

Contribution

The paper presents AGMA, a novel method that constructs expressive, scene-adaptive priors for better human trajectory forecasting, addressing prior misalignment issues.

Findings

AGMA achieves state-of-the-art results on ETH-UCY, Stanford Drone, and JRDB datasets.

High-quality, scene-adaptive priors improve prediction accuracy and diversity.

Theoretical analysis links prior quality to lower bounds of prediction error.

Abstract

Human trajectory forecasting requires capturing the multimodal nature of pedestrian behavior. However, existing approaches suffer from prior misalignment. Their learned or fixed priors often fail to capture the full distribution of plausible futures, limiting both prediction accuracy and diversity. We theoretically establish that prediction error is lower-bounded by prior quality, making prior modeling a key performance bottleneck. Guided by this insight, we propose AGMA (Adaptive Gaussian Mixture Anchors), which constructs expressive priors through two stages: extracting diverse behavioral patterns from training data and distilling them into a scene-adaptive global prior for inference. Extensive experiments on ETH-UCY, Stanford Drone, and JRDB datasets demonstrate that AGMA achieves state-of-the-art performance, confirming the critical role of high-quality priors in trajectory forecasting.