MG-GAN: A Multi-Generator Model Preventing Out-of-Distribution Samples in Pedestrian Trajectory Prediction

TL;DR

This paper introduces MG-GAN, a multi-generator model for pedestrian trajectory prediction that effectively captures multiple modes of future trajectories and reduces out-of-distribution samples.

Contribution

The paper proposes a novel multi-generator GAN architecture with a categorical selector to better model multimodal pedestrian trajectories.

Findings

Reduces out-of-distribution samples compared to single-generator models

Effectively captures multiple trajectory modes in complex scenes

Improves prediction accuracy in pedestrian trajectory datasets

Abstract

Pedestrian trajectory prediction is challenging due to its uncertain and multimodal nature. While generative adversarial networks can learn a distribution over future trajectories, they tend to predict out-of-distribution samples when the distribution of future trajectories is a mixture of multiple, possibly disconnected modes. To address this issue, we propose a multi-generator model for pedestrian trajectory prediction. Each generator specializes in learning a distribution over trajectories routing towards one of the primary modes in the scene, while a second network learns a categorical distribution over these generators, conditioned on the dynamics and scene input. This architecture allows us to effectively sample from specialized generators and to significantly reduce the out-of-distribution samples compared to single generator methods.