A Data-Efficient Progressive Learning Framework for Robot Scooping Task

Robot scooping is a challenging and important task in robotic tool manipulation research due to the complex relationship between the robot, the tool, and target objects/environment. Taking into account different tools, different target objects and varying environments, the required scooping manipulation strategy usually varies greatly. Even considering a specific type of spoon, the question of how to obtain a policy model that requires less demonstration data but shows better generalization capabilities deserves further exploration. In this paper, we propose a progressive learning framework for general robot scooping tasks, which requires a limited number of demonstrations but shows promising generalization capability. We first learn a scooping policy via human demonstrations with a specific setup. We then use this as a pre-train model for reinforcement learning in a curriculum manner to achieve a scooping strategy that is generalizable to different task setups. Finally, we evaluate the capabilities of the policy with a series of experiments both in simulation and on a real robot.

Authors

• Shuai Wang 0007
• Entang Wang
• Bidan Huang
• Chong Zhang
• Wei Wang
• Yu Zheng 0001

Keywords

• Shape
• Multimodal sensors
• Imitation learning
• Reinforcement learning
• Containers
• Stability analysis
• Data models
• Robots
• Learning Framework
• Progressive Learning
• Scooping Task
• Use Of Tools
• Important Task
• Target Object
• Generalization Capability
• Policy Model
• Reinforcement Learning Model
• Real Robot
• Number Of Demonstrations
• Model Performance
• Upper Limit
• Sequence Of Actions
• Difficulty Level
• Control Network
• Simulation Environment
• Object Size
• Depth Images
• End-effector
• Curriculum Learning
• Visual Encoding
• Catastrophic Forgetting
• Lowest Average Score
• Random Training
• Sense Of Hearing
• Task Environment
• Pose Changes
• Training Policy