Online movement adaptation based on previous sensor experiences

@article{Pastor2011OnlineMA,
  title={Online movement adaptation based on previous sensor experiences},
  author={Peter Pastor and Ludovic Righetti and Mrinal Kalakrishnan and Stefan Schaal},
  journal={2011 IEEE/RSJ International Conference on Intelligent Robots and Systems},
  year={2011},
  pages={365-371}
}
  • P. Pastor, L. Righetti, S. Schaal
  • Published 5 December 2011
  • Computer Science
  • 2011 IEEE/RSJ International Conference on Intelligent Robots and Systems
Personal robots can only become widespread if they are capable of safely operating among humans. In uncertain and highly dynamic environments such as human households, robots need to be able to instantly adapt their behavior to unforseen events. In this paper, we propose a general framework to achieve very contact-reactive motions for robotic grasping and manipulation. Associating stereotypical movements to particular tasks enables our system to use previous sensor experiences as a predictive… 

Figures from this paper

Probabilistic movement primitives under unknown system dynamics

TLDR
This work presents a reformulation of the ProMPs that allows accurate reproduction of the skill without modeling the system dynamics and derives a variable-stiffness controller in closed form that reproduces the trajectory distribution and the interaction forces present in the demonstrations.

Learning human compliant behavior from demonstration for force-based robot manipulation

TLDR
A hybrid external position/force control is presented to enable robot to produce a human-like compliant behavior, by combining probability modeling approach and dynamic system approach.

Residual Learning from Demonstration: Adapting Dynamic Movement Primitives for Contact-rich Insertion Tasks

TLDR
This work proposes a framework called residual learning from demonstration (rLfD) that combines dynamic movement primitives (DMP) that rely on behavioural cloning with a reinforcement learning (RL) based residual correction policy and shows that rLFD outperforms alternatives and improves the generalisation abilities of DMPs.

Robot Skill Representation, Learning and Control withProbabilistic Movement Primitives

TLDR
A novel movement primitive representation that not only models the shape of the movement but also its uncertainty in time is introduced, creating a mathematically sound framework that is capable of adapting skills to environmental changes as well as adapting the execution speed online.

Sensor-based failure prediction in biped locomotion

  • João AndréL. CostaC. Santos
  • Engineering, Computer Science
    2014 IEEE International Conference on Autonomous Robot Systems and Competitions (ICARSC)
  • 2014
TLDR
A bio-inspired architecture that adapts online the generated biped locomotion using sensory feedback and implements a failure detection system based on Associative Skill Memories (ASM), extending the concept to the specific case of biping locomotion.

Coupling of Trajectories for Human–Robot Cooperative Tasks

TLDR
This chapter presents a method for coupling of robot trajectories to the measured force feedback arising from the interaction with the environment, where the environment can be an object, a robot, or a person.

Skill Memory in Biped Locomotion

TLDR
This paper builds a predictive model of sensor traces that enables early failure detection by means of a skill memory, and extends the concept of Associative Skill Memories to periodic movements by introducing several mechanisms into the training workflow such that representation becomes time invariant and easily parameterizable.

Model-free Probabilistic Movement Primitives for physical interaction

TLDR
The model-free ProMPs are introduced, that are learning jointly the movement and the necessary actions from a few demonstrations, and derive a variable stiffness controller analytically.

Accelerating synchronization of movement primitives: Dual-arm discrete-periodic motion of a humanoid robot

  • A. GamsA. UdeJ. Morimoto
  • Computer Science
    2015 IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS)
  • 2015
TLDR
This paper shows how to accelerate trajectory adaptation based on learning of coupling terms in the framework of dynamic movement primitives (DMPs) and applies ideas from feedback error learning to iterative learning control (ILC).
...

References

SHOWING 1-10 OF 21 REFERENCES

Learning force control policies for compliant manipulation

TLDR
This work presents an approach to acquiring manipulation skills on compliant robots through reinforcement learning, and uses the Policy Improvement with Path Integrals (PI2) algorithm to learn these force/torque profiles by optimizing a cost function that measures task success.

Towards Reliable Grasping and Manipulation in Household Environments

TLDR
This work combines aspects such as scene interpretation from 3D range data, grasp planning, motion planning, and grasp failure identification and recovery using tactile sensors, aiming to address the uncertainty due to sensor and execution errors.

Learning motor primitives for robotics

  • J. KoberJan Peters
  • Computer Science
    2009 IEEE International Conference on Robotics and Automation
  • 2009
TLDR
It is shown that two new motor skills, i.e., Ball-in-a-Cup and Ball-Paddling, can be learned on a real Barrett WAM robot arm at a pace similar to human learning while achieving a significantly more reliable final performance.

On-line periodic movement and force-profile learning for adaptation to new surfaces

TLDR
A new method is proposed that enables the robot to adapt its motion to different surfaces and is applied to the ARMAR-IIIb humanoid robot, where it is used for learning and imitating a periodic task of wiping a kitchen table.

Policy search for motor primitives in robotics

TLDR
A novel EM-inspired algorithm for policy learning that is particularly well-suited for dynamical system motor primitives is introduced and applied in the context of motor learning and can learn a complex Ball-in-a-Cup task on a real Barrett WAM™ robot arm.

Challenges for Robot Manipulation in Human Environments

TLDR
From the Robotics Science and Systems Workshop: Manipulation for Human Environments, a perspective on this exciting area of robotics is presented, as informed by the workshop and the research.

Learning to grasp under uncertainty

TLDR
An approach is presented that enables robots to learn motion primitives that are robust towards state estimation uncertainties and that the robot learns to use fine manipulation strategies to maneuver the object into a pose at which closing the hand to perform the grasp is more likely to succeed.

A Sensitive Approach to Grasping

Experimental results in psychology have shown the important role of manipulation in guiding infant development. This has inspired work in developmental robotics as well. In this case, however, the

Skill learning and task outcome prediction for manipulation

TLDR
This work presents a Reinforcement Learning based approach to acquiring new motor skills from demonstration that allows the robot to learn fine manipulation skills and significantly improve its success rate and skill level starting from a possibly coarse demonstration.

Operational Space Control: A Theoretical and Empirical Comparison

TLDR
An extensive empirical results demonstrate that one of the simplified acceleration-based approaches can be advantageous in terms of task performance, ease of parameter tuning, and general robustness and compliance in the face of inevitable modeling errors.