Advanced

Interaction Patterns in Human Augmented Mapping

Topp, Elin Anna LU (2017) In Advanced Robotics 31(5). p.258-267
Abstract
Interactive robots can benefit from learning from humans in situated communications, for instance in a guided tour, where the human user explains the environment to a mobile robot, or from being instructed to handle a specific task in an industrial setting. We have to assume that human users do not always give unambiguous or consistent information, which might result in inconsistent information being stored, which can later lead to confusing situations in further interaction with the user. We suggest to consider behavioural features observable from user actions, including commands given to the robot, to detect inconsistencies in the information given by the user with respect to the overall understanding of the situation. We explain our... (More)
Interactive robots can benefit from learning from humans in situated communications, for instance in a guided tour, where the human user explains the environment to a mobile robot, or from being instructed to handle a specific task in an industrial setting. We have to assume that human users do not always give unambiguous or consistent information, which might result in inconsistent information being stored, which can later lead to confusing situations in further interaction with the user. We suggest to consider behavioural features observable from user actions, including commands given to the robot, to detect inconsistencies in the information given by the user with respect to the overall understanding of the situation. We explain our concept of Interaction Patterns and discuss how such patterns can be applied to support hypothesis generation regarding the category of a presented item in a guided tour scenario, which can lead to meaningful request formulation in mixed-initiative interaction. We report on results regarding the identification of Interaction Patterns in data from a user study with 37 subjects. (Less)
Please use this url to cite or link to this publication:
author
organization
publishing date
type
Contribution to journal
publication status
published
subject
in
Advanced Robotics
volume
31
issue
5
pages
258 - 267
publisher
Taylor & Francis
external identifiers
  • scopus:85010987798
  • wos:000396783900004
ISSN
0169-1864
DOI
10.1080/01691864.2017.1281758
language
English
LU publication?
yes
id
3b31960c-4faa-431a-a87a-350adb5d5a55
date added to LUP
2016-12-23 15:24:02
date last changed
2018-01-07 11:42:38
@article{3b31960c-4faa-431a-a87a-350adb5d5a55,
  abstract     = {Interactive robots can benefit from learning from humans in situated communications, for instance in a guided tour, where the human user explains the environment to a mobile robot, or from being instructed to handle a specific task in an industrial setting. We have to assume that human users do not always give unambiguous or consistent information, which might result in inconsistent information being stored, which can later lead to confusing situations in further interaction with the user. We suggest to consider behavioural features observable from user actions, including commands given to the robot, to detect inconsistencies in the information given by the user with respect to the overall understanding of the situation. We explain our concept of Interaction Patterns and discuss how such patterns can be applied to support hypothesis generation regarding the category of a presented item in a guided tour scenario, which can lead to meaningful request formulation in mixed-initiative interaction. We report on results regarding the identification of Interaction Patterns in data from a user study with 37 subjects.},
  author       = {Topp, Elin Anna},
  issn         = {0169-1864},
  language     = {eng},
  number       = {5},
  pages        = {258--267},
  publisher    = {Taylor & Francis},
  series       = {Advanced Robotics},
  title        = {Interaction Patterns in Human Augmented Mapping},
  url          = {http://dx.doi.org/10.1080/01691864.2017.1281758},
  volume       = {31},
  year         = {2017},
}