DESTINE: Dynamic Goal Queries with Temporal Transductive Alignment for Trajectory Prediction

TL;DR

DESTINE is a novel trajectory prediction method that dynamically estimates goals and aligns trajectories temporally, achieving state-of-the-art results by integrating coarse-to-fine predictions and attention mechanisms.

Contribution

The paper introduces DESTINE, a new approach that dynamically predicts goals without relying on road structures and aligns trajectories using a novel temporal attention module.

Findings

Achieves state-of-the-art performance on Argoverse benchmark.

Effectively predicts map-compliant trajectories with dynamic goals.

Demonstrates the importance of temporal alignment in trajectory prediction.

Abstract

Predicting temporally consistent road users' trajectories in a multi-agent setting is a challenging task due to unknown characteristics of agents and their varying intentions. Besides using semantic map information and modeling interactions, it is important to build an effective mechanism capable of reasoning about behaviors at different levels of granularity. To this end, we propose Dynamic goal quErieS with temporal Transductive alIgNmEnt (DESTINE) method. Unlike past arts, our approach 1) dynamically predicts agents' goals irrespective of particular road structures, such as lanes, allowing the method to produce a more accurate estimation of destinations; 2) achieves map compliant predictions by generating future trajectories in a coarse-to-fine fashion, where the coarser predictions at a lower frame rate serve as intermediate goals; and 3) uses an attention module designed to temporally align predicted trajectories via masked attention. Using the common Argoverse benchmark dataset, we show that our method achieves state-of-the-art performance on various metrics, and further investigate the contributions of proposed modules via comprehensive ablation studies.