Sim-to-Real Reinforcement Learning for Deformable Object Manipulation

Abstract

We have seen much recent progress in rigid object manipulation, but interaction with deformable objects has notably lagged behind. Due to the large configuration space of deformable objects, solutions using traditional modelling approaches require significant engineering work. Perhaps then, bypassing the need for explicit modelling and instead learning the control in an end-to-end manner serves as a better approach? Despite the growing interest in the use of end-to-end robot learning approaches, only a small amount of work has focused on their applicability to deformable object manipulation. Moreover, due to the large amount of data needed to learn these end-to-end solutions, an emerging trend is to learn control policies in simulation and then transfer them over to the real world. To-date, no work has explored whether it is possible to learn and transfer deformable object policies. We believe that if sim-to-real methods are to be employed further, then it should be possible to learn to interact with a wide variety of objects, and not only rigid objects. In this work, we use a combination of state-of-the-art deep reinforcement learning algorithms to solve the problem of manipulating deformable objects (specifically cloth). We evaluate our approach on three tasks --- folding a towel up to a mark, folding a face towel diagonally, and draping a piece of cloth over a hanger. Our agents are fully trained in simulation with domain randomisation, and then successfully deployed in the real world without having seen any real deformable objects.

arXiv commentPublished at the Conference on Robot Learning (CoRL) 2018

References in corpus (9)

• Continuous control with deep reinforcement learning
• Prioritized Experience Replay
• Leveraging Demonstrations for Deep Reinforcement Learning on Robotics Problems with Sparse Rewards
• Domain Randomization for Transferring Deep Neural Networks from Simulation to the Real World
• Transferring End-to-End Visuomotor Control from Simulation to Real World for a Multi-Stage Task
• 3D Simulation for Robot Arm Control with Deep Q-Learning
• Deep Reinforcement Learning for Robotic Manipulation with Asynchronous Off-Policy Updates
• Learning Robust Bed Making using Deep Imitation Learning with DART
• Folding Deformable Objects using Predictive Simulation and Trajectory Optimization

Cited by in corpus (25)

• Open-Sourced Reinforcement Learning Environments for Surgical Robotics
• Automated shapeshifting for function recovery in damaged robots
• Scaling data-driven robotics with reward sketching and batch reinforcement learning
• When Deep Learning Meets Data Alignment: A Review on Deep Registration Networks (DRNs)
• DisCoRL: Continual Reinforcement Learning via Policy Distillation
• Deep Reinforcement Learning in Computer Vision: A Comprehensive Survey
• Learning to Manipulate Deformable Objects without Demonstrations
• PlasticineLab: A Soft-Body Manipulation Benchmark with Differentiable Physics
• Learning Arbitrary-Goal Fabric Folding with One Hour of Real Robot Experience
• Hierarchical Foresight: Self-Supervised Learning of Long-Horizon Tasks via Visual Subgoal Generation
• Q-Learning for Continuous Actions with Cross-Entropy Guided Policies
• Time Reversal as Self-Supervision
• ROBEL: Robotics Benchmarks for Learning with Low-Cost Robots
• Diverse Auto-Curriculum is Critical for Successful Real-World Multiagent Learning Systems
• RetinaGAN: An Object-aware Approach to Sim-to-Real Transfer
• Untangling Dense Knots by Learning Task-Relevant Keypoints
• The Importance and the Limitations of Sim2Real for Robotic Manipulation in Precision Agriculture
• Auto-Tuned Sim-to-Real Transfer
• MetaGrasp: Data Efficient Grasping by Affordance Interpreter Network
• Continual Learning: Tackling Catastrophic Forgetting in Deep Neural Networks with Replay Processes
• Scalable sim-to-real transfer of soft robot designs
• Towards Closing the Sim-to-Real Gap in Collaborative Multi-Robot Deep Reinforcement Learning
• Human Correspondence Consensus for 3D Object Semantic Understanding
• End-to-End Egospheric Spatial Memory
• Ubiquitous Distributed Deep Reinforcement Learning at the Edge: Analyzing Byzantine Agents in Discrete Action Spaces