Deep Context Maps: Agent Trajectory Prediction using Location-specific Latent Maps

TL;DR

This paper introduces location-specific latent context maps that enhance agent trajectory prediction accuracy in cluttered environments by capturing environment-specific information beyond visual cues.

Contribution

The novel concept of learning and integrating context maps as latent representations into trajectory prediction models is introduced, improving prediction accuracy.

Findings

Learned context maps significantly improve predictor accuracy.

Providing partial map semantics further boosts performance.

Maps capture environment-specific trajectory patterns.

Abstract

In this paper, we propose a novel approach for agent motion prediction in cluttered environments. One of the main challenges in predicting agent motion is accounting for location and context-specific information. Our main contribution is the concept of learning context maps to improve the prediction task. Context maps are a set of location-specific latent maps that are trained alongside the predictor. Thus, the proposed maps are capable of capturing location context beyond visual context cues (e.g. usual average speeds and typical trajectories) or predefined map primitives (such as lanes and stop lines). We pose context map learning as a multi-task training problem and describe our map model and its incorporation into a state-of-the-art trajectory predictor. In extensive experiments, it is shown that use of learned maps can significantly improve predictor accuracy. Furthermore, the performance can be additionally boosted by providing partial knowledge of map semantics.