Adaptive Proxemic Path Planning for Mobile Robots in Dynamic Human Environments with Predicted Motion

De Sousa R.M.; Diaz-Amado J.; Barros P.V.A.; De Souza Silva L.H.; De Oliveira J.P.G. · Proceedings - 2026 Brazilian Conference on Robotics, CROS 2026 (2026) · DOI: 10.1109/cros69211.2026.11565544

Resumen

This paper presents an advanced path planning framework for mobile robots operating in environments shared with humans, focusing on enhancing human-robot interaction through adaptive proxemic behavior and dynamic human motion prediction. We extend the A∗ algorithm to incorporate a novel cost function that considers both static proxemic zones around humans and their estimated future trajectories. The proposed approach, termed "A∗ Dynamic Prediction,"aims to generate robot paths that not only avoid immediate human personal spaces but also anticipate and circumvent potential future conflicts with human movement. We compare this new method against traditional non-proxemic, rigid proxemic, and adaptive proxemic planning strategies. Our simulations demonstrate that by integrating dynamic human motion prediction, the robot can achieve smoother, safer, and more socially acceptable trajectories, significantly reducing proxemic invasions and potential collisions while maintaining efficient path lengths. The results highlight the importance of predictive capabilities in fostering more natural and harmonious human-robot coexistence in complex, dynamic settings.

Instituciones: Universidad Católica San Pablo