Trustworthy Pedestrian Trajectory Prediction via Pattern-Aware Interaction Modeling

TL;DR

This paper introduces InSyn, a Transformer-based model for pedestrian trajectory prediction that explicitly models interaction patterns for more trustworthy and accurate predictions, especially in dense scenarios, and proposes a training strategy to improve initial prediction accuracy.

Contribution

The paper presents InSyn, a novel interaction-aware Transformer model with a new training strategy, enhancing transparency and accuracy in pedestrian trajectory prediction.

Findings

Outperforms recent black-box models in accuracy, especially in high-density scenarios.

Effectively models diverse interaction patterns among pedestrians.

Reduces initial-step prediction error by approximately 6.58% with SSOS strategy.

Abstract

Accurate and reliable pedestrian trajectory prediction is critical for the application of intelligent applications, yet achieving trustworthy prediction remains highly challenging due to the complexity of interactions among pedestrians. Previous methods often adopt black-box modeling of pedestrian interactions. Despite their strong performance, such opaque modeling limits the reliability of predictions in real-world deployments. To address this issue, we propose InSyn (Interaction-Synchronization Network), a novel Transformer-based model that explicitly captures diverse interaction patterns (e.g., walking in sync or conflicting) while effectively modeling direction-sensitive social behaviors. Additionally, we introduce a training strategy, termed Seq-Start of Seq (SSOS), designed to alleviate the common issue of initial-step divergence in numerical time-series prediction. Experiments on the ETH and UCY datasets demonstrate that our model not only outperforms recent black-box baselines in prediction accuracy, especially under high-density scenarios, but also provides transparent interaction modeling, as shown in the case study. Furthermore, the SSOS strategy proves to be effective in improving sequential prediction performance, reducing the initial-step prediction error by approximately 6.58%. Code is avaliable at https://github.com/rickzky1001/InSyn