SocNavShield
Ensuring Safety in Social Robot Navigation via Model Predictive Shielding
News
[2026-03-XX]: The project was kicked off!
About
SocNavShield is dedicated to bridging the gap between sociability and safety in autonomous robot navigation. While modern robots are becoming more adaptive, they often struggle to balance fluid movement with the strict safety and social norms required in human-centered environments. Powered by the research of ENSTA and École Polytechnique within Institut Polytechnique de Paris, we are developing a hybrid navigation architecture integrating Deep Reinforcement Learning with Model Predictive Shielding. Our goal is to provide an open-source, verifiable framework that ensures robots navigate human environments safely, comfortably, and politely.
Team
 |
 |
 |
 |
 |
|---|---|---|---|---|
| Zhi Yan (PI) Associate Professor |
Elena Vanneaux (Co-I) Assistant Professor |
Sergio Mover (Co-I) Full Professor |
Luiza Soares Master’s student |
Naomy Biaggi Master’s student |
This project has received funding from Hi! PARIS under grant agreement No XXX.
