Biomedical Engineering Reference
In-Depth Information
6.3
User Satisfaction
Overall, it shows a very good user satisfaction for touch and speech modalities, with
averagely 4.7 and 4.3 out of 5. Specifically, the speech and textual outputs are consi-
dered appropriately constructed; the interface is intuitive and easy to understand; the
process to perform the task is feasible; and the system is considered user-friendly.
However, the scores of system behavior and user perspective for the speech modal-
ity were a bit lower than the others. This is mainly due to the problem of the automat-
ic speech recognizer, which could trigger unexpected system responses, and therefore
make the future use from the user perspective less attractive. This impression is also
reflected to every other aspect for speech modality, with overall lower score com-
pared to touch input.
Table 4.
The assessment of subjective user satisfaction
Touch
Speech
Mean
Std.
Mean
Std.
System behavior
4.8
0.3
3.6
0.7
Speech output
4.7
0.5
4.6
0.6
Textual output
4.9
0.3
4.5
0.5
Interface presentation
4.8
0.3
4.7
0.4
Task performing
4.5
0.3
4.3
0.5
User-friendliness
4.7
0.4
4.4
0.6
User perspective
4.3
0.9
3.9
0.8
Overall
4.7
0.2
4.3
0.4
7
Conclusions and Future Work
This paper summarized our work on multimodal interaction for elderly persons, cen-
tering the following two essential aspects:
•
The design and implementation of a multimodal interactive system according to a
number of elderly-friendly guidelines concerning with the basic design principles of
conventional interactive interfaces and ageing centered characteristics;
•
The modelling and development of multimodal interaction using a tool-supported,
formally tractable and extensible unified dialogue modelling approach.
In order to evaluate the minutely designed and developed multimodal interactive sys-
tem, an experimental study was conducted with 31 elderly persons and concerned
with the touch and speech input modality respectively. The evaluation showed high
effectiveness, sufficient efficiency and a high satisfaction of the participants with our
system for both modalities. Due to the problem caused by the automatic speech rec-
ognition, touch modality displays a better performance with respect to effectiveness
and is preferred by the elderly. But the speech modality helped the participants in a
more efficient way. Thus, the combination of both modalities is motivated.
The presented work served as a continuing step towards building an effective, effi-
cient, adaptive and robust multimodal interactive framework extensively for elderly
persons. The result of a further study focusing on the touch and speech combined
modality is being analyzed. Corpus-based supervised and reinforcement learning
Search WWH ::
Custom Search