Robots Fashion and Physical Programming
Questions To be answered:
Can a robot sleep when it is wearing its pajama?
Can Roomba feel shy and hide?
Can I have my own pet Pleo that looks and acts differently from my sister’s?
Can Pleo dress up according to its mood?
Personalization of Technology
Dressing up toys was the most popular activity which many of us enjoyed doing when we were kids; both girls with their dolls and boys with their boy dolls. Unsurprisingly, we never stopped personalizing our own belonging. We accessorize our mobile phones and electronic gadgets; we place stickers on laptops that say ‘this is mine’. The laptop I’m using to write this article is from work and its surface is filled with stickers reminding me that it is not my personal laptop (check the photo below).
People tend to personalize their digital devices. Likewise, and as previously discussed in Theater HRI experiment, users prefer to be able to change the physical appearance of robots. Generally, appearance is connected to behavior. For example when a little girl dresses her doll in pajamas, she will put her afterwards to sleep imagining that the doll was sleepy. For that reason, connections between physical appearance and behavior should be carefully considered in designing studies of technology products. Changing the behavior of a robot sounds like a very sophisticated task and requires a lot of knowledge in programming and engineering. Though it is true, normal and unspecialized people can program robots behaviors through physical language programming.
Physical programming is the technique of creating new behavior of a program with only manipulating physical codes and cues. The increase need for physical programming for technology that will give a novice more power to control appearance and behavior of robots is the inspirations and motivation for the actDresses[1] case study in LIREC. Next, we will review the study and other related studies in the areas of tangible and visual programming.
Robot Marionettes
How about robot marionettes? Attaching strings to a robot to control its behavior is a possible option one could try. In contrast to marionettes, robots are of rather heavier substance, which requires very strong strings to achieve the desired movement. The unfeasibility of this approach doesn’t negate that fact that it is possible to manipulate robots with strings. The Tangible Media Group at MIT[2] has (like turning with certain degrees, and moving to certain distances), and one for sensors of the robot (like temperature, light), and a set of string. The physical programming in this method depends on linking actions to sensors via strings. When the sensors are triggered, the action(s) they are linked to on the panel will be executed. It is easy for any child to perform.
The inspiration of this method supports the actDresses concept that tackles the physical programming issue in more detailed way that is closer to Human dressing and appearance concepts.
Human Clothing and Robots?
To enable children to still pursue their joy in interacting with toys and personalizing them, and in the same time to allow them to interact with smart technology, physical programming means are needed to be available and concurrently simple. This way, children would be smarter and able to create versatile interaction scenarios with their imagination[3].
In the human world, clothing is a widely used tool by children to personalize their toys. Clothes are daily used to reflect our own modes and behaviors according to our culture. Similarly, in the robot world, Roomba and Pleo have their own share of garments and accessories culture on the web.
Clothes have a basic modularity property. Where different items could be combined together. This property relates clothing to the high level programming aspects where libraries, interfaces and modules are used together[4]. Differing from other visual programming methods, actDresses incorporates physical codes like clothing garments that are integrated with the robot itself. While the strings system codes are separated from the object they are programming. actDresses involves three scenarios in interacting with the robots: Pleo, GlowBots, Roomba.
Pleo in Fashion
Dressing up Pleo causes it to shift between different modes: from watching dog when wearing the dog bracelet, to walking when a shoe is on, ending with the sleeping baby dinosaur when wearing its pajama. Changing the “persona” of Pleo has very rich range of options. Putting a warm jacket on it might also changes its behavior, in addition to many different factors that could affect its behavior and mode. A smart Pleo seems to be available. A robot with recognition ability to the clothes that it is wearing is such a huge step. However, this new feature doesn't really express an increased intelligence level of Pleo. It is merely the physical programming of the behavior that simply uses RFID technology with RFID tags to whatever garments available accompanied with temperature and sound sensors. The group of added hardware allows Pleo to conclude what is the matching mode it is supposed to be in and simply shift to it.
GlowBots with Accessories
Seeing a glowbot following square paths while it has a square tag on its head is not a strange thing to see. Glowbots are small robots that display visual shapes, generate sounds and move in certain paths. In actDresses special physical ‘amulets’ are used to change the behavior of these tiny bots as the image illustrates below:
Roomba and Tags
With some comic magnetic patches the vacuum robot could change its behavior. It could hide under a couch when “shy” tag is attached to it. It could also swirl silently on the carpet.
Design Challenges
Designing physical manipulation to robots technology is interesting and very desired approach from users. At the same time it brings a large range of limitations and restrictions. First of all, since mapping one physical cue changes the mode of the robot, then attaching more than one garment may end in a state of confusion of modes or failure in achieving what is desired. Just attaching the garments of “watch-dog” and “sleeping” to Pleo will cause confusion in reaction. Scalability of the system then is restricted.
If one could simply press a button to put Pleo to sleep, it would be easier than dressing it in a pajama. Other simpler options might be available like adding switchers that can simply changes the mode of the robot. Technically wise this is cheaper and more efficient. One drastic downside for this option is that it will lower the level of interaction and believability with the “live robot”. Therefore the amusement of the user and learning of children would be affected.
Closet for Your Robots
Developing new ways to control the behavior of robots is a very promising approach to satisfy the increasing demand of the public to have more interaction with their robots. The physical visual programming offers rich approaches in personalizing the technology, which leads to more affective mutual interaction. Therefore, garments, accessories and jewelries are the items expected to exist in a robots wardrobe in the future.
Article Answers:
Can a robot sleep when it is wearing its pajama? Yes, it can.
Can Roomba feel shy and hide? Yes, it can.
Can I have my own Pleo that looks and acts differently from my sister’s? Yes, you can.
Can Pleo dress up according to its mood? Not yet. In this article we don’t imply that the behaviors of robots are generated by their free will. They are only reacting to user’s actions.
References:
[1] Fernaeus, Y. and Jacobsson, M. (2009). Comics, Robots, Fashion and Programming: outlining the concept of actDresses. Proceedings of TEI'09, Cambridge, UK, ACM
[2] Patten, J., Griffith, L. and Ishii, H. (2000). A Tangible Interface for Controlling Robotic Toys. Proceedings of CHI'00, Sunrotary, ACM.277-278
[3] Fernaeus, Y., Kindborg, M. and Sholz, R. (2006). Rethinking children's programming with contextual signs. Proceedings of IDC'06, Tampere, Finland, ACM Press.121-128
[4] Fernaeus, Y. and Jacobsson, M. (2009). Comics, Robots, Fashion and Programming: outlining the concept of actDresses. Proceedings of TEI'09, Cambridge, UK, ACM