Autonomous Navigation of a Mobile Robot in a Known Map with Dynamic Obstacles
12th International Conference on Control, Decision and Information Technologies, CoDIT 2026, Bari, İtalya, 13 - 16 Temmuz 2026, ss.3186-3191, (Tam Metin Bildiri)
- Yayın Türü: Bildiri / Tam Metin Bildiri
- Doi Numarası: 10.1109/codit70676.2026.11630742
- Basıldığı Şehir: Bari
- Basıldığı Ülke: İtalya
- Sayfa Sayıları: ss.3186-3191
- Anahtar Kelimeler: autonomous navigation, dynamic obstacle avoidance, mobile robots, MPPI, Nav2, ROS 2, skid-steer
- Hacettepe Üniversitesi Adresli: Evet
Özet
The expansion of autonomous mobile robots from industrial facilities into dynamic environments shared with humans has made it imperative for navigation algorithms to operate reliably under new uncertainties. While traditional navigation methods succeed in static maps, they often cause the "freezing robot"problem in chaotic scenarios involving suddenly moving obstacles. This study presents the mechanical, electronic, and software design of an integrated autonomous navigation system capable of safely and smoothly avoiding static and dynamic obstacles on a known map. Local planning is performed using the Model Predictive Path Integral (MPPI) controller on a custom-designed skid-steer mobile robot platform, featuring a slip-aware EKF configuration and an asymmetric acceleration envelope, with full parameter settings reported for reproducibility. We present this as a systems-integration and reproducibility study rather than a new algorithmic contribution. In a representative single-run real-world evaluation, the tuned stack cleared 23 unmapped static obstacles and 17 dynamic encounters (pedestrians and a quadruped robot) with zero collisions and no observed freeze events under our success definitions; we do not claim a statistical success rate, and multi-trial validation is the primary follow-up. A detailed event-level analysis demonstrates the interplay between the MPPI planner's backward yielding capabilities and Behavior-Tree (BT) recovery on human-shared corridors.