Multiple Future Prediction Leveraging Synthetic Trajectories

TL;DR

This paper introduces a data-driven approach using Markov Chains to generate synthetic trajectories for training multi-future prediction models in autonomous driving, improving accuracy and capturing diverse outcomes.

Contribution

It proposes a novel synthetic data generation method with Markov Chains and a multimodal loss, enhancing trajectory prediction performance.

Findings

Synthetic data improves prediction accuracy.

Combining real and synthetic data yields state-of-the-art results.

The approach captures multiple plausible future trajectories.

Abstract

Trajectory prediction is an important task, especially in autonomous driving. The ability to forecast the position of other moving agents can yield to an effective planning, ensuring safety for the autonomous vehicle as well for the observed entities. In this work we propose a data driven approach based on Markov Chains to generate synthetic trajectories, which are useful for training a multiple future trajectory predictor. The advantages are twofold: on the one hand synthetic samples can be used to augment existing datasets and train more effective predictors; on the other hand, it allows to generate samples with multiple ground truths, corresponding to diverse equally likely outcomes of the observed trajectory. We define a trajectory prediction model and a loss that explicitly address the multimodality of the problem and we show that combining synthetic and real data leads to prediction improvements, obtaining state of the art results.