EUVIMEDEUVIMED entdecken
Open Access · CC BY 4.0

End-to-End Imitation Learning for Socially Aware Navigation Using Memory-Based Trajectory and Action Planning

Alejandro Olivas, Miguel Ángel Muñoz-Bañón, Fernando Torres

IEEE Access · 2026 · 14 · 81979-81993

DOI: 10.1109/access.2026.3694511 ↗

Kurz gefasst

Fachartikel in der Zeitschrift IEEE Access, 2026. Verfasst von Alejandro Olivas und 2 weiteren Autor:innen. Die Arbeit nennt 67 Literaturangaben. Der Volltext ist frei zugänglich (Open Access, CC BY 4.0). Gefördert von: Ministry of Science and Innovation of the Spanish Government through the Research Project, Formación del Personal Investigador [Research Staff Formation (FPI)], Valencian Community Government and the European Union through the Fellowship.

AbstractBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

The increasing deployment of service robots and autonomous vehicles operating in human-populated environments underscores the importance of planning trajectories to ensure pedestrian comfort while optimizing robot travel times. End-to-end models are promising for shortcutting the complex problem of pedestrian detection, trajectory prediction, and robot movement planning by using a single network. End-to-end Imitation Learning (IL) has been widely adopted in self-driving because it enables implicit understanding of complex environments and expert driving behaviour, whereas socially aware navigation has usually relied on non-end-to-end Reinforcement Learning (RL) approaches that learn from abstracted environment representations, prioritizing information relevant to human interaction. In this research, we aim to address socially aware navigation generalizing the capabilities of the end-to-end IL approaches. For this, we present Memory-based Trajectory and Action Planning (MTAP), which performs intelligent pedestrian avoidance while maintaining a deeper understanding of the environment and the expert driving behaviour. Our method predicts the trajectory and the control action commands directly from the information of the sensors and the global target, employing a recurrent layer to anticipate pedestrian movements based on past frames. The network is trained and evaluated in the CARLA simulator, where we introduce a benchmark for assessing pedestrian avoidance efficacy, presenting also new social navigation metrics to measure both trajectory efficiency and pedestrian comfort during avoidance maneuvers. The code and results are publicly available: https://aurova-projects.github.io/MTAP/

Förderung

  • Ministry of Science and Innovation of the Spanish Government through the Research Project
  • Formación del Personal Investigador [Research Staff Formation (FPI)]
  • Valencian Community Government and the European Union through the Fellowship

Zitieren

APA

Olivas, A., Muñoz-Bañón, M. Á., & Torres, F. (2026). End-to-End Imitation Learning for Socially Aware Navigation Using Memory-Based Trajectory and Action Planning. IEEE Access, 14, 81979-81993. https://doi.org/10.1109/access.2026.3694511

Vancouver

Olivas A, Muñoz-Bañón MÁ, Torres F. End-to-End Imitation Learning for Socially Aware Navigation Using Memory-Based Trajectory and Action Planning. IEEE Access. 2026;14:81979-81993. doi:10.1109/access.2026.3694511

BibTeX
@article{Olivas2026,
  author = {Olivas, Alejandro and Muñoz-Bañón, Miguel Ángel and Torres, Fernando},
  title = {End-to-End Imitation Learning for Socially Aware Navigation Using Memory-Based Trajectory and Action Planning},
  journal = {IEEE Access},
  year = {2026},
  volume = {14},
  pages = {81979--81993},
  doi = {10.1109/access.2026.3694511}
}
Kostenlos starten →