Variable Impedance Control in End-Effector Space: An Action Space for Reinforcement Learning in Contact-Rich Tasks

@article{MartnMartn2019VariableIC,
  title={Variable Impedance Control in End-Effector Space: An Action Space for Reinforcement Learning in Contact-Rich Tasks},
  author={Roberto Mart{\'i}n-Mart{\'i}n and M. Lee and Rachel Gardner and S. Savarese and Jeannette Bohg and Animesh Garg},
  journal={2019 IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS)},
  year={2019},
  pages={1010-1017}
}
  • Roberto Martín-Martín, M. Lee, +3 authors Animesh Garg
  • Published 2019
  • Computer Science
  • 2019 IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS)
  • Reinforcement Learning (RL) of contact-rich manipulation tasks has yielded impressive results in recent years. While many studies in RL focus on varying the observation space or reward model, few efforts focused on the choice of action space (e.g. joint or end-effector space, position, velocity, etc.). However, studies in robot motion control indicate that choosing an action space that conforms to the characteristics of the task can simplify exploration and improve robustness to disturbances… CONTINUE READING
    25 Citations
    Learning Variable Impedance Control for Contact Sensitive Tasks
    • 8
    • PDF
    Learning Task Space Actions for Bipedal Locomotion
    • PDF
    Deep Reinforcement Learning for Contact-Rich Skills Using Compliant Movement Primitives
    • PDF
    Learning Force Control for Contact-Rich Manipulation Tasks With Rigid Position-Controlled Robots
    • 3
    • PDF
    Neural Dynamic Policies for End-to-End Sensorimotor Learning
    • Highly Influenced
    • PDF
    Variable Impedance Control and Learning—A Review
    • PDF
    Stability-Guaranteed Reinforcement Learning for Contact-Rich Manipulation
    • 2
    • Highly Influenced
    • PDF

    References

    SHOWING 1-10 OF 57 REFERENCES
    Learning force control policies for compliant manipulation
    • 132
    • PDF
    Force-based variable impedance learning for robotic manipulation
    • 17
    • PDF
    Learning variable impedance control
    • 232
    • Highly Influential
    • PDF
    Learning Compliant Manipulation through Kinesthetic and Tactile Human-Robot Interaction
    • 101
    • PDF
    Learning-based control strategy for safe human-robot interaction exploiting task and robot redundancies
    • 125
    • PDF
    Optimal distribution of contact forces with inverse-dynamics control
    • 158
    • PDF
    Learning locomotion skills using DeepRL: does the choice of action space matter?
    • 60
    • PDF
    Learning a Structured Neural Network Policy for a Hopping Task
    • 6
    • PDF
    End-to-End Training of Deep Visuomotor Policies
    • 1,950
    • PDF