Obstacle Avoidance Using Hybrid Reinforcement Learning in Dynamic Environments

A tétel rövid nézete
dc.contributor.advisorAlmusawi, Husam Abdulkareem
dc.contributor.authorPham, Tuan Nam
dc.contributor.departmentDE--Műszaki Kar
dc.date.accessioned2026-06-02T08:45:31Z
dc.date.available2026-06-02T08:45:31Z
dc.date.created2026-05-02
dc.description.abstractThis thesis presents a hybrid reinforcement learning approach for obstacle avoidance with the Festo Robotino 4 omnidirectional mobile robot in dynamic environments. The proposed method combines a classical Pure Pursuit controller with a Soft Actor-Critic (SAC) agent that uses a custom 1D CNN (LidarCNN1D) to extract spatial features from stacked LiDAR observations. The system is implemented in ROS 2 Humble and Gazebo, and trained with a three-stage curriculum that scales from a 10×10 m room with static obstacles up to a 6×6 m room with six dynamic pedestrians. Four methods — Hybrid SAC, Optimized SAC, Pure SAC, and Pure PPO — are compared to isolate the contribution of each architectural choice. The Hybrid SAC achieves the best results, reaching a 99% success rate in the hardest training stage and 84% in the most difficult unseen evaluation environment, while maintaining the lowest collision rate (16%). The findings show that CNN-based feature extraction, frame stacking and Pure Pursuit guidance combine to deliver robust real-time navigation suitable for indoor service robots.
dc.description.courseMechatronical Engineeringen
dc.description.degreeMSc/MA
dc.format.extent94
dc.identifier.urihttps://hdl.handle.net/2437/407659
dc.language.isoen
dc.rights.infoHozzáférhető a 2022 decemberi felsőoktatási törvénymódosítás értelmében.
dc.subjectobstacle avoidance, deep reinforcement learning, mobile robot navigation
dc.subject.dspaceEngineering Sciences::Engineering
dc.titleObstacle Avoidance Using Hybrid Reinforcement Learning in Dynamic Environments
Fájlok
Eredeti köteg (ORIGINAL bundle)
Megjelenítve 1 - 2 (Összesen 2)
Nincs kép
Név:
Thesis classification request.pdf
Méret:
94.49 KB
Formátum:
Adobe Portable Document Format
Leírás:
Megjelenítés
Nincs kép
Név:
Obstacles_Avoidance_Hybrid_RL.pdf
Méret:
2.73 MB
Formátum:
Adobe Portable Document Format
Leírás:
Megjelenítés
Engedélyek köteg
Megjelenítve 1 - 1 (Összesen 1)
Nincs kép
Név:
license.txt
Méret:
1.69 KB
Formátum:
Item-specific license agreed upon to submission
Leírás:
Letöltés
Gyűjtemények
Theses (Department of Electrical Engineering and Mechatronics)