Predicting Future Spatiotemporal Occupancy Grids with Semantics for Autonomous Driving

TL;DR

This paper introduces a semantic-aware environment prediction framework for autonomous vehicles that improves the accuracy and temporal consistency of future occupancy predictions, especially for moving objects, using real-world data.

Contribution

It integrates environment semantics into occupancy prediction, enhancing long-term prediction accuracy and object persistence compared to baseline methods.

Findings

Higher prediction accuracy on Waymo dataset

Better maintenance of moving object appearances over time

Improved long-term occupancy prediction performance

Abstract

For autonomous vehicles to proactively plan safe trajectories and make informed decisions, they must be able to predict the future occupancy states of the local environment. However, common issues with occupancy prediction include predictions where moving objects vanish or become blurred, particularly at longer time horizons. We propose an environment prediction framework that incorporates environment semantics for future occupancy prediction. Our method first semantically segments the environment and uses this information along with the occupancy information to predict the spatiotemporal evolution of the environment. We validate our approach on the real-world Waymo Open Dataset. Compared to baseline methods, our model has higher prediction accuracy and is capable of maintaining moving object appearances in the predictions for longer prediction time horizons.