Modular Neural Network Policies for Learning In-Flight Object Catching with a Robot Hand-Arm Systems
Published in IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS), 2023
Recommended citation: Wenbin Hu, Fernando Acero, Eleftherios Triantafyllidis, Zhaocheng Liu and Zhibin Li. (2023). "Modular Neural Network Policies for Learning In-flight Object Catching with a Robot Hand-Arm System." in IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS). https://ieeexplore.ieee.org/abstract/document/10341463
The paper introduces a modular framework for training robot hand-arm systems to catch flying objects, combining supervised learning for trajectory prediction and pose ranking with deep reinforcement learning for reaching and grasping actions. Extensive simulations show high success rates with the system generalising beyond its training to catch various household objects.