Skip to Content

a. Prototype

Version 13.1 by Deepali Prabhu on 2023/04/10 15:25

The main requirements from our robot to implement our design were:

  • Human like calm voice
  • Ability to show images on the tablet
  • Ability to add gestures such as point to PwD, point to Self, point at object.
  • Ability to show emotion through eye colour, gestures, dialogue.
  • Ability to play music

We picked the Pepper robot for the same as it could cater to our requirements. In addition to this we implemented the following to improve the human-robot interaction.

Dialog

[1] provided communicative strategies to ensure effective interaction with PwD.

  • The speed was reduced to ensure that the PwD understood clearly on what the robot had to say and had sufficient time to reflect and respond.
  • The pitch and volume were adjusted according to the surrounding to ensure no agitation.
  • Cognitive processing for those with dementia is also undermined by the grammatical complexity of the stimuli. Hence, the dialog was kept simple and short. The instructions were broken down into smaller steps to ensure progress.
  • To inculcate theory of mind[], an optimal level of feedback from PwD was triggered to ensure

Gestures

Referred implementation from: https://github.com/frietz58/WoZ4U

However, based on our requirements, we modified the above codebase in this repository: https://github.com/shambhawipal-tud/WoZ4U.git

Our prototype consists of two vertical slices of the functionalities that we wanted to achieve. The first slice consists of helping the PwD draw by using only a pencil. The second slice consists of functionality where the PwD draws and also paints using watercolors. 

We performed user testing in two parts to answer the research question: a Control group and a Test group.

Control group:

  • It consists of a painting session in which the user painted by following instructions from a piece of paper.
  • No interaction with Pepper is performed here.

Test group:

  • It consists of a painting session with PaintPal where Pepper guides the user through the painting process.

We had 18 participants in our user evaluation. Because of the lack of participants, we conducted a within-subject study. However, to reduce the transfer effects, we conducted the experiment in a randomized manner, where 50% of the participants were first in the test group and 50% of the participants were first in the control group.

The video of our prototype only shows one slice of the functionality where the user just used the pencil to draw.

Prototype video: Prototype.mp4

[1] Weirather RR. Communication strategies to assist comprehension in dementia. Hawaii Med J. 2010 Mar;69(3):72-4. PMID: 20397508; PMCID: PMC3104619.