![](bg4.png)
algorithm and the attributes mentioned before can be
used to correctly identify movements made by
people whose movements are not listed in the case
base. The proposed algorithm can distinguish the
movement’s orientation and can distinguish squares
from circles and in which direction they were done.
After this analysis, it was noted that most
incorrect identifications where due to bad input data.
It was mentioned that a linear movement over an
axis resembles a one-period sine wave and that
information was used in the development of the
CBR identifier, but if the data received isn’t
minimally similar to that sine wave, if it resembles a
slope for instance, the movement’s attributes will
make the system answer incorrectly. This can be
avoided if the user starts and ends the movement
with the Wiimote stopped. Another cause of bad
data is the speed at which the movement is done.
Low velocities will result in lower accelerations that,
when combined with the Wiimote’s accelerometer
low signal-to-noise ratio (SNR), will result in bad
attributes that will make the system answer
incorrectly.
…
Figure 7: Correct identification percentage of linear
movements by the CBR algorithm.
5 CONCLUSIONS
An implementation of a local human-robot interface
based on the Nintedo Wii’s remote controller was
presented. Due to the Wiimote features, it is easy to
use and inexpensive.
Three working modes were developed:
• The steering wheel allowed the user to set
references for v and
ω
in order to remote control
the robot in an intuitive way
• The “Follow the Leader” mode that uses
distance and angle to the “leader” and is
interesting to lead the robot to distant locations
quickly teaming up with the leader
• The “Movement Identifier” is an interface
usable in future for general commands.
The first two methods use a fuzzy controller. The
movement identifier uses Case Based Reasoning.
User tests hint that system usability is easy.
In the future, new working modes might be
added to the interface as well as new movements.
Another possibility is to allow for multiple
Wiimotes. It will also be very interesting to use the
Wiimote’s speaker to feedback perceived commands
to the user.
ACKNOWLEDGEMENTS
This work was partially supported by FCT Project
PTDC/EIA/70695/2006 "ACORD - Adaptative
Coordination of Robotic Teams".
REFERENCES
O. Mayora-Ibarra 2003. From HCI to HRI - Usability
Inspection in “Multimodal Human-Robotics
Interactions”, Proceedings of the 2003 IEEE
International Workshop on Robot and Human
Interactive Communication.
Perzanowski, D.; Schultz, A.C.; Adams, W.; Marsh, E.;
Bugajska, M. 2001. Building a Multimodal Human-
Robot Interface, in Intelligent Systems,IEEE, vol. 16,
Issue 1, Jan-Feb 2001 pp. 16 – 21.
Sousa, A. 2006 - CleanRob project official website.
Available at http://paginas.fe.up.pt/~cleanrob.
D. Driankov, 1996 An Introduction to Fuzzy Control –
Second Edition, Berlin, Springer, 1996, chapters 2,3
and 4.
Jenson, J., 2005. Design of Fuzzy Controllers in European
Network for Fuzzy Logic and Uncertainty Modeling in
Information Technology,
http://www.iau.dtu.dk/~jj/pubs/design.pdf
Aamodt, A. and Plaza, E, 1994. Case-Based Reasoning:
Foundational Issues, Methodological Variations, and
System Approaches in Artificial Intelligence
Communications 7, no. 1, 1994, pp. 39-52
ICAART 2009 - International Conference on Agents and Artificial Intelligence
348