We gratefully acknowledge support from
the Simons Foundation and member institutions.
Full-text links:


Current browse context:


Change to browse by:


References & Citations

DBLP - CS Bibliography


(what is this?)
CiteULike logo BibSonomy logo Mendeley logo del.icio.us logo Digg logo Reddit logo ScienceWISE logo

Computer Science > Robotics

Title: Human-Robot Kinaesthetic Interaction Based on Free Energy Principle

Abstract: The current study investigated possible human-robot kinaesthetic interaction using a variational recurrent neural network model, called PV-RNN, which is based on the free energy principle. Our prior robotic studies using PV-RNN showed that the nature of interactions between top-down expectation and bottom-up inference is strongly affected by a parameter, called the meta-prior, which regulates the complexity term in free energy.The study also compares the counter force generated when trained transitions are induced by a human experimenter and when untrained transitions are induced. Our experimental results indicated that (1) the human experimenter needs more/less force to induce trained transitions when $w$ is set with larger/smaller values, (2) the human experimenter needs more force to act on the robot when he attempts to induce untrained as opposed to trained movement pattern transitions. Our analysis of time development of essential variables and values in PV-RNN during bodily interaction clarified the mechanism by which gaps in actional intentions between the human experimenter and the robot can be manifested as reaction forces between them.
Comments: 12 pages, 8 figures, journal paper
Subjects: Robotics (cs.RO)
Cite as: arXiv:2303.15213 [cs.RO]
  (or arXiv:2303.15213v1 [cs.RO] for this version)

Submission history

From: Hiroki Sawada [view email]
[v1] Mon, 27 Mar 2023 13:53:07 GMT (9006kb,D)

Link back to: arXiv, form interface, contact.