Tracing the Influence of Predecessors on Trajectory Prediction

TL;DR

This paper introduces the Predecessor-and-Successor (PnS) method, which models the influence of neighboring agents' trajectories to improve prediction accuracy in traffic scenarios, achieving state-of-the-art results without relying on map data.

Contribution

The paper presents a novel PnS approach that incorporates predecessor influence modeling into trajectory prediction, enhancing performance on multiple datasets and reducing dependence on map information.

Findings

Achieved state-of-the-art pedestrian trajectory prediction on ETH/UCY datasets.

Maintained good vehicle trajectory prediction performance without map data on nuScenes.

Effectively aligned motion encodings with multiple potential predecessors.

Abstract

In real-world traffic scenarios, agents such as pedestrians and car drivers often observe neighboring agents who exhibit similar behavior as examples and then mimic their actions to some extent in their own behavior. This information can serve as prior knowledge for trajectory prediction, which is unfortunately largely overlooked in current trajectory prediction models. This paper introduces a novel Predecessor-and-Successor (PnS) method that incorporates a predecessor tracing module to model the influence of predecessors (identified from concurrent neighboring agents) on the successor (target agent) within the same scene. The method utilizes the moving patterns of these predecessors to guide the predictor in trajectory prediction. PnS effectively aligns the motion encodings of the successor with multiple potential predecessors in a probabilistic manner, facilitating the decoding process. We demonstrate the effectiveness of PnS by integrating it into a graph-based predictor for pedestrian trajectory prediction on the ETH/UCY datasets, resulting in a new state-of-the-art performance. Furthermore, we replace the HD map-based scene-context module with our PnS method in a transformer-based predictor for vehicle trajectory prediction on the nuScenes dataset, showing that the predictor maintains good prediction performance even without relying on any map information.