Seq2Seq Imitation Learning for Tactile Feedback-based Manipulation

Robot control for tactile feedback based manip-ulation can be difficult due to modeling of physical contacts, partial observability of the environment, and noise in perception and control. This work focuses on solving partial observability of contact-rich manipulation tasks as a Sequence-to-Sequence (Seq2Seq) Imitation Learning (IL) problem. The proposed Seq2Seq model first produces a robot-environment interaction sequence to estimate the partially observable environment state variables, and then, the observed interaction sequence is transformed to a control sequence for the task itself. The proposed Seq2Seq IL for tactile feedback based manipulation is experimentally validated on a door-open task in a simulated environment and a snap-on insertion task with a real robot. The model is able to learn both tasks from only 50 expert demonstrations while state-of-the-art reinforcement learning and imitation learning methods fail.

Authors

• Wenyan Yang
• Alexandre Angleraud
• Roel Pieters
• Joni Pajarinen
• Joni-Kristian Kämäräinen

Keywords

• Learning systems
• Robot control
• Tactile sensors
• Reinforcement learning
• Transformers
• Human in the loop
• Trajectory
• Imitation Learning
• State Variables
• Observable Variables
• Manipulation Tasks
• Reinforcement Learning Methods
• Real Robot
• Expert Demonstrations
• Seq2seq Model
• Mixture Model
• Target Object
• Stiffness Matrix
• Hidden State
• Sampling Efficiency
• Tactile Sensor
• Markov Decision Process
• Robot Manipulator
• Rotated Component
• Real Task
• Partially Observable Markov Decision Process
• Simulated Task
• Inverse Reinforcement Learning
• End-effector Pose
• Object Pose
• LSTM Network
• Encoder-decoder Structure
• Executive Skills
• Exploration Stage
• Number Of Demonstrations