48 Just ordinAry robots
2.2.1 Experiences of Early Adaptors: Roombarization
Research by Sung, Grinter, Christensen, and Guo (2008) shows that
people who buy a vacuum-cleaning robot belong to the group of early
adopters, namely curious and well-educated young people, and that
the general public has not been reached. From the study by Sung
etal., it is clear how these early adopters experienced the vacuum-
cleaning robot and how they used it. In general, they are positive
about the robot. Users with young children or pets were signicantly
positive. is is a dierent use of the vacuum-cleaning robot for
them: the robot invites babies or very young children to learn to crawl
or walk, because they want to follow the robot and the robot chases
the pets, or the pets (especially birds) are driven around the room on
top of the robot.
A striking fact that emerges from research is that people who own
vacuum-cleaning robots tidy their home more often. A path must
be cleared in order to make room for the vacuum-cleaning robot.
In addition, small items, for example, small parts of toys, must be
removed from the oor; otherwise, the robot sucks them up. As a
result, for most people this leads to promoting household order, as
they will go through the house with a cleaning cloth. In addition, a
large majority still continues to use the regular vacuum cleaner for
several reasons. e robot version does not always manage to get the
oor completely dust-free, and pet hair in particular is sometimes left
behind. Sometimes speedy cleaning is desired, and since the vacuum
robot is unable to walk up stairs, reach into corners or other di-
cult places, such as vacuuming a round cables or removing cobwebs,
people will still use the regular vacuum cleaner.
Here we see a tendency comparable to what occurred with the
introduction of the regular vacuum cleaner. It was expected then that
housecleaning could be done signicantly more swiftly, so housewives
would gain time for other activities. However, it turned out that hygiene
standards were soon raised considerably, which ultimately required more
time for cleaning than before the era of the vacuum cleaner (Cowan,
1983). us, domestic labor does not necessarily get lighter because of
the introduction of the vacuum-cleaning robot, as the advertisement
promises us: “ink of your most tedious housecleaning task. Now
think about never having to do it again. Indoors and out, our robots
49home is where the robot is
are engineered for cleaning performance and convenience, bringing
the latest robotic technology to real-world homes.”*
e study by Sung, Grinter etal. (2008) showed that almost all
users of a robotic vacuum cleaner made changes to the organization
of their home and their home furniture. e more tidy and less fur-
nished the household is, the easier it is to make use of that vacuum-
cleaning robot. is process of rationalizing the environment so that
the vacuum-cleaning robot can do its job better is known as room-
barization (Sung, Guo, Grinter, & Christensen, 2007), referring to
the Roomba. Typical modications are moving or hiding cables and
cords, removing deep pile carpet, removing lightweight objects from
the oor, and moving furniture. When purchasing new furniture
or oors, one should take the capabilities of the vacuum-cleaning
robot into account. An inhibiting factor for the rise of the commer-
cial vacuum-cleaning robot probably lies in this need for a structured
environment. e history of technology research shows that the inter-
est in new devices quickly decreases when existing practices require
too many changes (Oldenziel, 2001). at this probably also holds
for vacuum-cleaning robots follows from the study by Vaussard etal.
(2014). In this study, most households stopped using the robot after
a while, because they became disappointed as they actually assessed
the robot’s relevance within their own ecosystem; in other words,
they assessed how well the robot integrates inside the user’s space and
perception and considered the fact that the robot does not actually
decrease the amount of work for the user.
Presently, the huge potential for selling vacuum-cleaning robots
has not yet been fullled. In a Dutch magazine for computer technol-
ogy, 24 dierent vacuum-cleaning robots were tested, including the
most popular, such as the Roomba, and its conclusion is telling in that,
economically, there is currently no convincing argument to buy any of
the tested robots.
†
Most of the reviews indicate that vacuum-cleaning
robots cannot entirely replace the traditional human-driven vacuum
cleaners.
‡
Acceptance of useful household appliances is dicult if
*
www.irobot.com/.
†
C’t Magazine voor Computertechniek 2011, 11, 106. See also http://robot-vacuum-
review.toptenreviews.com/.
‡
http://www.cnet.com/news/robot-vacuum-roundup/.
50 Just ordinAry robots
those devices only work partially, and the regular household appli-
ance remains an indispensable device in the household.
A historical technological argument to the detriment of the
vacuum-cleaning robot is that devices that can save time and labor
lose out to devices used for relaxation. In the 1930s, many house-
holds bought a radio and then often had no money left for a wash-
ing machine (Bowden & Oer, 1996). We see a similar tendency in
relation to the vacuum-cleaning robot. Despite the economic crisis,
people are ocking to buy expensive game systems for entertainment,
such as the Nintendo Wii and the Microsoft Xbox 360. More than
100 million Wiis have been sold since its introduction in 2006.*
e experiences and problems outlined above with the vacuum-
cleaning robot also occur with the lawnmower robot and the mop
robot. With those, a form of roombarization will also take place. For
the lawnmower robot, the owner must have a power socket outside for
charging and a at lawn without water features. ere should be no
trees in the yard from which fruit or branches might drop, because the
robot will have problems with those objects, and loose items should
not be strewn on the lawn as the robot mower tends to mow much
more frequently than a regular human-operated mower. Additionally,
for the robotic lawnmower, it is recommended that children are not
playing in the yard when the robot is working, because of safety issues
and also because a robot can be too attractive as a toy (see Section 2.4.2
regarding liability). Although the lawnmower robot operates on very
low electric power, it can sever ngers or toes when these come into
contact with the rotating blades. An integrated safety circuit therefore
switches the blades o automatically when the machine is lifted.
2.2.2 Reducing the Complexity of Household Tasks
It appears that household tasks require very complex decisions of a
robot, because they are comparable to solving the so-called frame
problem. In such a problem, a clear criterion is missing for checking
whether a proposed solution is acceptable. In a domestic task, often
unconsciously, we search, organize, and select all kinds of relevant
information and we will take a decision on the basis of that information.
*
www.nintendo.co.jp/ir/library/historical_data/pdf/consolidated_sales_e1406.pdf.
51home is where the robot is
e environment of the household tasks is not static, but changes con-
stantly. For example, the contents of the laundry basket, clothes that
need folding, change each time we use it. Dierent clothes, some even
inside out, will be in the basket. e problem of forcing a robot to
adapt to these changes, which humans often consider to be common
sense, is the basis of the frame problem in articial intelligence, and it
is very hard to represent (see also Floridi, 2014).
e degree of diculty is shown by research from the University
of California at Berkeley, which aims to develop a robot able to fold
laundry.* e main result that emerged is how dicult it really is to
fold textiles mechanically, which is a relatively simple task for humans.
Eventually, a robot was developed that took nearly 25minutes to
fold one towel.
†
Also, for the rst prototypes of cooking robots, it
shows that simply breaking an egg into a pan forms an insurmount-
able obstacle. e prospect that we will not have to do any cooking in
the near future is still very far away. Nevertheless, there are already
cooking robots in highly structured environments where cooking is
reduced to a few basic sequences of actions, where the vegetables are
precut and washed for the robot and the ingredients have already
been put in scales at xed locations. For a particular dish, the robot
will take the necessary ingredients for the recipe from the dishes by
following preprogrammed instructions, put them into a wok or pan
and stir them together. e cooking robot may not be suitable for a
household, but it might well be useful in school cafeterias and mili-
tary camps, for example, in terms of mass production and where a
prefabricated kitchen is tted.
‡
To deal with the complexity of cooking, the environment is
therefore rationalized: a closed microenvironment is created within
an unstructured environment. A good example is one of the most
successful household devices—the dishwasher. By redening the
process of washing the dishes by spraying hot water at the dishes,
instead of manual dishwashing, which relies largely on physical
scrubbing to remove soiling, a closed microenvironment is created to
deal with the complex task of washing the dishes in a dynamic and
*
newscenter.berkeley.edu/2010/04/02/robot/.
†
www.youtube.com/watch?v=pjk69h9P8.
‡
www.twanetwerk.nl/default.ashx?DocumentID=11839.
52 Just ordinAry robots
unstructuredenvironment. Another recent household device, Tubie
the ironing robot, has become available for U.S. $1100.* With Tubie,
the process of ironing is also redened. Tubie is in fact a doll over
which the consumer pulls the garment. en a motor inates the
doll, and hot air is blown all through the garment. It thus takes about
5minutes per garment before the ironing program is over. In contrast
with the dishwasher, although we do not have to iron, we must be
constantly present during the process to pull the clothing on and o
the doll and guide the deation process. erefore, Tubie does not
fully take over the job, and consequently ironing in this way takes
even longer than if we had used the old-fashioned steam iron.
e process of rationalizing the environment also applies to the
vacuum-cleaning robot, as we have already seen, which we call room-
barization. Before using the robot, the environment needs to be pre-
pared so that the robot can do the task properly without, for example,
sucking up a sales receipt. When cleaning up the room, we nd all
kinds of stu that we may or may not throw out or move—a loose
Lego brick is thrown back into the Lego box, a sales receipt on the
oor is perhaps not just vacuumed up, but we rst check whether we
still need to keep it for guarantee purposes. ese are all decisions that
are impossible (especially for the time being) for a vacuum- cleaning
robot to make. To enable them to make these decisions, engineers
are trying to make robots smarter by connecting them to a cognitive
control system. e international robotics project “Web-Enabled and
Experience-Based Cognitive Robots at Learn Complex Everyday
Manipulation,” from 2011,
†
funded by the European Commission, has
had this aim. Such cognitive controls allow robots to pick up infor-
mation available on the Internet and transfer it into proper behavior.
ink of web instructions, online dictionaries, and online encyclo-
pedias. In this way, a household robot—as is expected—could even
learn to cook simple dishes based on recipes found on the web. In
addition, numerous robotic technology–based domotics solutions (see
Chapter 3) and ambient intelligence (Aarts & Encarnação, 2006)
‡
*
www.ironing-machines-tubie.com/.
†
ias.cs.tum.edu/research/robohow.
‡
In ambient intelligence, the emphasis is more on increasing comfort for humans (cre-
ating smart homes, etc.) than on a robot’s specic actions. Further equipping robots
with intelligent functionality we can perhaps le under the label ambient intelligence.
..................Content has been hidden....................
You can't read the all page of ebook, please click here login for view all page.