The growth of the personal robotics market is showing signs of mirroring the early growth of personal computers. While this market is nowhere near as large as that of the personal computer, it is as large as that of traditional industrial robotics, and it is growing quickly. According to studies by the Japan Robotics Association, the United Nations Economic Commission, and the International Federation of Robotics, the personal and service robotics market is already equal to that of industrial robotics at about 5,400,000,000 U.S. dollars. By 2025 it is projected to be four times the size of the industrial robotics market, or about 51,700,000,000 U.S. dollars, and this is excluding military robotics and entertainment robotics which would greatly increase this dollar amount.69
This explosive growth is garnering the same kind of investor excitement as the dotcom boom of the 1990s and a few large trade shows have been organized to help hype the technology and funnel investment dollars into this industry.70 Behind the hype and over exuberance occasioned by the introduction of personal robotics technology, there is an interesting and significant reality. Slowly but surely, more or less autonomous machines are making their way into our lives, from expensive robotic toys like the Sony Aibo robotic dog, to robotic vacuum cleaners and lawn-mowers, all the way to the new crop of robotic weapons platforms currently deployed in the Middle East (Aproberts, 2004).
One of the most socially interesting developments in robotics technology has been the creation of robotic companions built to suit the emotional needs of children, the elderly, and even love sick young adults. These robots are primarily designed by Korean and Japanese companies and research centers that are keenly interested in building machines that are more than simply appliances: they are interested in making our future friends.
We can see two distinct design paradigms forming in the burgeoning personal robotics industry. For the sake of discussion I will call them the ‘effective’ and the ‘affective’ design paradigms. For example, American and European robotics companies have largely focused on very utilitarian, or effective, implementations of robotics technologies by building robotic vacuum cleaners, lawnmowers, and weapons platforms. Japanese and Korean companies have pursued the more playful or affective aspect of robotics, building ingenious robotic pets, dolls, and humanoid companions. Sony, Honda, and Hitachi have all built extremely expensive humanoid robotic mascots that dance and wow the crowds at tradeshows and in advertising.
Effective design here refers to the interpretation of robots as tools or appliances meant to automate some formerly human activity. Effective design in robotics is the design strategy that seeks to remove some task from the human lifeworld and delegate it to robotics technology that can deal with the problem with little or no human direction. The robot effectively takes over some task that is too mundane, dirty, dangerous, or otherwise distasteful to leave to humans. An example of an effective robotic design that is already in place might be a vacuum cleaning robot that is programmed to come out of its charging station at night so it can vacuum a room and have it ready before its owners awake.
Affective design seeks to imbed the robot deeply into the lifeworld of the humans with which it interacts. These machines are built to elicit, and even ‘experience’ emotion, in order to bond more fully with their human users. This is an intriguing notion, and it is by far the more radical of the two design paradigms found in robotics today. It is this design strategy that we will focus on in this chapter. In sections four and five we will look at a few examples of this technology and explore some of the motivations of the engineers working on these machines.
It would be too simplistic to suggest that the differences between effective and affective robotics design are entirely accounted for by diversity in culture since we will see that there are important researchers in the West that are making many breakthroughs in the affective design paradigm and the Japanese have lead the world in building factory robots that are firmly in the effective robotics design paradigm. However, it is true that one finds a more ready acceptance amongst consumers of friendly and good-humored robotic designs in the East, especially in Japan.
Before we look at some of the interesting affective robots that have already been built, we need to review some of the insights that have influenced the robotics movement towards affective robotics design.
The roboticist Takayuki Kanda and other researchers from the Advanced Telecommunications Research Institute Intelligent Robotics and Communications Labs in Kyoto (ATR), in conjunction with a number of Japanese Universities, have studied the psychological and sociological factors that can be observed during human robot interactions. They state that, “[f]or realizing a robot working in human society, interaction with humans is the key issue” (Kanda et al., 2001). They add that to achieve a robot that can elicit positive emotional responses from its human users, the robot needs to have some understanding of human psychology and group dynamics so that it can more fully interact with those around it.
Takayuki Kanda’s ATR lab built a robot named “ROBOVIE,” and studied its interactions with human test subjects. ROBOVIE has a vaguely human shape with a head, arms, torso, and a wheeled undercarriage. It is also equipped with an antenna that tracks radio frequency identification (RFID) badges worn by the humans interacting with it. This allows the robot to easily identify the different people it comes into contact with. The ATR researchers believe that a robot is only seen as intelligent by its operators if it both performs actions and expresses its ability to function in a natural and human like way (Kanda et al., 2001). For instance, just having people interact with a robotic head or some other restricted design is not going to draw out affective interactions with the machine, but a robot with a complete body that can interact with users autonomously, “...lets observers easily attribute various intentions to the robot based on its gaze-related movement” (Kanda et al., 2001). The researchers at ATR had the robot interrelate with fifty nine subjects and then asked each of them to fill out a questionnaire. The respondents rated the robot on a seven point scale between twenty eight pairs of opposite traits, such as friendly-unfriendly, exciting-dull, intelligent-unintelligent, etc. They found that close contact with an expressive robot that could accomplish various tasks brought about the most favorable impressions in the subjects (Kanda et al., 2001).
In another set of experiments, the ATR Intelligent Robotics and Communications Lab took ROBOVIE to elementary schools for extended periods of interaction with students in the classroom (Kanda et al., 2004; Kanda and Ishiguro, 2005). The robot was able to interact with students in a modest way engaging with them in about seventy behaviors, including simple games, telling them secrets, giving hugs and kisses to them, and making other friendly gestures. Takayuki Kanda and Hiroshi Ishiguro have been able to design the robot to engage in simple conversations, it can speak some three hundred sentences and understand about fifty words (Kanda and Ishiguro, 2005). This design has proven to be engaging enough to interest some children in interacting with the robot for extended periods of time. In one experiment the robot was programmed gradually to give out more “secret’ information about itself depending on the amount of time the student spent with the robot and this, along with the robots ability to call out student’s names, proved to be a very popular set of behaviors with the students (Kanda et al., 2004). The students wore nametags that had an RFID transmitter in them so that the robot was able to know with whom it was it was interacting. This feature allowed ROBOVIE to track the number and length of interactions it had with various students and also to attempt to deduce the friendship relationships that existed between the students in the classroom, in which it achieved to some moderate success (Kanda et al., 2004). The ATR Labs’ goal is to eventually create a robot that can interact with students in a friendly manner and help teach children in the classroom while building relationships with the students and to, “.help maintain safety in the classroom such as by moderating bullying problems, stopping fights among children, and protecting them from intruders” (Kanda and Ishiguro, 2005).
70
Robonexus is a consumer trade show (http://www.robonexus.com/index.html) and Robobusiness is for industry members (http://www.roboevent.com/).