Synapse
⌘+K
Synapse
PulseExploreClubsResearchersJournals
Instagram
HomeClubsExplore
June 19, 2026Scientia Sinica Informationis

PHGDiff: physics-enhanced hypergraph diffusion model for crowd trajectory prediction

View Full Paper
Ask AI
Bookmark
Share

Authors

QNQingying NIUQJQiankun JIN

Discussion

Loading...

Member takes

Overview

Randomized trial demonstrates enhanced trajectory prediction in crowds, suggesting improvements for autonomous systems.

Key Points

  • The study aims to improve pedestrian trajectory prediction by addressing uncertainties in human behavior and social interactions.
  • Developed a physics-enhanced hypergraph diffusion model (PHGDiff) for trajectory prediction.
  • Incorporated a dynamic hypergraph mechanism to represent social interactions through hyperedges.
  • Utilized a noise-aware physical constraint scheduler to balance diversity and physical consistency.
  • PHGDiff achieved superior trajectory prediction accuracy compared to existing models.
  • Demonstrated improved diversity in predicted trajectories while maintaining physical plausibility.
  • Extensive testing on ETH and UCY datasets confirmed robust performance.

Cite This Study

NIU et al. (2026) studied this question.

synapsesocial.com/papers/6a34df7565a5b0777af2e808https://doi.org/10.1360/ssi-2025-0299
View Full Paper
Ask AI
Bookmark
Share

Also Consider

Synapse has enriched 5 closely related papers on similar clinical questions. Consider them for comparative context:

  1. 1Social Physics Informed Diffusion Model for Crowd Simulation2024 · 16 citations
  2. 2DifTraj: Diffusion Inspired by Intrinsic Intention and Extrinsic Interaction for Multi-Modal Trajectory Prediction2024 · 1 citations
  3. 3Dynamic Population Distribution Aware Human Trajectory Generation with Diffusion Model2026 · 2 citations
  4. 4Diff-Pre: A Diffusion Framework for Trajectory Prediction2025 · 6 citations
  5. 5Stochastic human trajectory prediction <i>via</i> interaction-aware diffusion model2026