Abstract Visual Programming of Social Robots for Novice Users

Brown, Onis (2021) Abstract Visual Programming of Social Robots for Novice Users. MRes thesis, University of Lincoln.

Abstract Visual Programming of Social Robots for Novice Users
MRes Thesis
Brown, Onis -Computer Science - March 2021.pdf - Whole Document

Item Type:Thesis (MRes)
Item Status:Live Archive


Human Robot Interaction is a rapidly growing field that blends multiple disciplines
with highly varied research methods. To better leverage experts in non-engineering
(software/hardware) fields and increase engagement with the wider public, development of HRI components such as robot behaviours need to be made more accessible.
This Project aims to develop and evaluate a system with a suitably high level of abstraction so that novices can create high level behaviours or tasks without having to
deal with complexities such as low-level hardware control or the myriad of individual
behaviours that combine to form seemingly straightforward actions. The system was
developed for use on simulated and physical forms of the Scitos G5, currently operating at the collection museum. Over the course of the project, two evaluations
have been performed. A study on the behaviours exhibited by the robot and a linear
version of the visual programming interface, and then workload and usability associated with the system interface after making more high-level blocks and programming
constructs available. The study at the collection museum was used to inform the
other study taking place using simulated sessions with the robot. The second study
was an evaluation of a more complex version of the system. Across both studies,
the system was found to be favourable according to SUS responses, with TLX responses showing that the mental demand for the system was low in relation to other
dimensions. The latter study showed that that some people found that elements of
the visual programming were slightly confusing, however the functionality of blocks
and their relation to performing certain actions were easily learned. Results suggest
further study may be useful with future development of this or similar systems in
wider contexts.

Keywords:Human-computer interaction, human-robot interaction
Subjects:G Mathematical and Computer Sciences > G440 Human-computer Interaction
Divisions:College of Science > School of Computer Science
ID Code:48584
Deposited On:16 Mar 2022 12:31

Repository Staff Only: item control page