Generalizing Behavior Trees and Motion-Generator (BTMG) Policy Representation for Robotic Tasks Over Scenario Parameters

Research output: Contribution to conferencePaper, not in proceedingpeer-review

Abstract

We propose a generalisation of a behaviour tree and motiongenerator based robot arm policy representation for learning and solving tasks such as contact-rich tasks like peg insertion or pushing an object. We use planning to generate skill sequences needed to execute these tasks and rely on reinforcement learning to obtain parameters of the policy. We assume gaussian processes as a suitable method for this generalisation and present preliminary, promising results from initial experiments.
Original languageEnglish
Number of pages2
Publication statusPublished - 2022

Subject classification (UKÄ)

  • Robotics

Free keywords

  • Generalization
  • Industrial Robots
  • Reinforcement learning

Fingerprint

Dive into the research topics of 'Generalizing Behavior Trees and Motion-Generator (BTMG) Policy Representation for Robotic Tasks Over Scenario Parameters'. Together they form a unique fingerprint.

Cite this