Multiple Futures Prediction

TL;DR

This paper presents a probabilistic, data-driven framework for multi-agent motion prediction that models multiple potential futures using latent variables, enabling improved planning and state-of-the-art trajectory prediction accuracy.

Contribution

It introduces a novel probabilistic framework with semantic latent variables and a dynamic attention encoder for scalable, multimodal multi-agent trajectory prediction without explicit labels.

Findings

Achieves state-of-the-art results on vehicle trajectory datasets.

Effectively models multimodal futures with latent variables.

Scales efficiently to any number of agents.

Abstract

Temporal prediction is critical for making intelligent and robust decisions in complex dynamic environments. Motion prediction needs to model the inherently uncertain future which often contains multiple potential outcomes, due to multi-agent interactions and the latent goals of others. Towards these goals, we introduce a probabilistic framework that efficiently learns latent variables to jointly model the multi-step future motions of agents in a scene. Our framework is data-driven and learns semantically meaningful latent variables to represent the multimodal future, without requiring explicit labels. Using a dynamic attention-based state encoder, we learn to encode the past as well as the future interactions among agents, efficiently scaling to any number of agents. Finally, our model can be used for planning via computing a conditional probability density over the trajectories of other agents given a hypothetical rollout of the 'self' agent. We demonstrate our algorithms by predicting vehicle trajectories of both simulated and real data, demonstrating the state-of-the-art results on several vehicle trajectory datasets.