A Dynamic Mission Planning Framework Combining Improved A∗ and Neural-OFMPC
12th International Conference on Control, Decision and Information Technologies, CoDIT 2026, Bari, İtalya, 13 - 16 Temmuz 2026, ss.86-91, (Tam Metin Bildiri)
- Yayın Türü: Bildiri / Tam Metin Bildiri
- Doi Numarası: 10.1109/codit70676.2026.11631126
- Basıldığı Şehir: Bari
- Basıldığı Ülke: İtalya
- Sayfa Sayıları: ss.86-91
- Anahtar Kelimeler: Dynamic Environments, Hierarchical mission planning, Improved A, Motion Planning, MPC, Neural Model Predictive Control, Path Planning
- Hacettepe Üniversitesi Adresli: Evet
Özet
This paper presents a hierarchical mission planning framework in dynamic environments. The framework combines an improved A∗ global path planner with a Neural Optimization-Free Model Predictive Control (Neural-OFMPC) local controller. The Improved A∗ algorithm generates safer and smoother reference paths by incorporating safety distance weighting and key-turning node extraction, reducing unnecessary heading changes and simplifying local tracking. At the control level, Neural-OFMPC augments an optimization-free MPC with a lightweight neural residual policy that adaptively refines conservative control actions. The neural controller enables proactive avoidance maneuvers, such as steering and acceleration, while preserving safety through predictive collision checking and a minimal safety shield. Since no online optimization is required, the proposed approach maintains low and consistent computational cost.Simulation results demonstrate that the proposed framework improves success rate, execution time, and navigation efficiency compared to conventional OFMPC-based approaches, particularly in complex scenarios with dynamic obstacles. These results indicate that the proposed method is well-suited for real-time mission planning on resource-constrained robotic platforms.