Information Technology Reference
In-Depth Information
Ta b l e 5 .
Response times with different lexicons
small lexicon large lexicon
utt. i
0.07 sec
0.08 sec
utt. ii
0.10 sec
0.14 sec
utt. iii
0.68 sec
0.90 sec
utt. iv
0.76 sec
1.15 sec
utt. v
2.35 sec
2.51 sec
4.3
Discussion
An important point towards successful human-robot interaction with respect to the
user's patience is the system's reaction time. The average human attention span (for
focused attention, i.e. the short-term response to a stimulus) is considered to be ap-
proximately eight seconds [6]. Therefore, the time we require to process the utterance
of a user and react in some way must not exceed 8 seconds. Suitable reactions are the
execution of a request, rejection, or to start a clarification process.
Hence, the question whether computation times are reasonable is in fact the question
whether the computation times exceed eight seconds. Nonetheless, the answer is not as
easy as the question. The optimised system performs well in a realistic test scenario as
shown by the last row of Table 3. In turn, complex test scenarios can lead to serious
problems as Table 4 indicated. However, we saw that ambiguity is a smaller problem
than the length of an utterance
1
. Skills that have more than three parameters are rare in
the field of mobile service robots. In fact, the skills with four or five parameters we used
in the tests of Table 4 needed to be created artificially in lack of realistic examples.
5
Conclusions and Future Work
We presented a system for interpreting commands issued to a domestic service robot
using decision-theoretic planning. The proposed system allows for a flexible matching
of utterances and robot capabilities and is able to handle faulty or incomplete commands
by using clarification. It is also able to provide explanations in case the user's request
cannot be executed and is rejected. The system covers a broader set of possible requests
than existing systems with small and fixed grammars. Also, it performs fast enough to
prevent annoying the user or loosing his or her attention.
Our next step is to deploy the system in a RoboCup@Home competition to test its
applicability in a real setup. A possible extension of the approach could be to include a
list of the
n
most probable interpretations and to verify with the user on which of these
should be executed. Moreover, properly integrating the use of adverbials as qualifiers
for nouns both in the grammar and the interpretation process would further improve the
system's capabilities.
1
By the length of an utterance, we mean the number of spoken objects.
