Michaud’s paper addresses a specific difficulty with providing children with an adaptable, interactive toy while at the same time creating a robust toy that will not break with constant use and abuse from children. The paper speaks about a prototype that was created, “Roball”, a ball-shaped robot that moves by making its external spherical shell rotate. The reason for the design is because of the well known fact that children are extremely difficult on their toys as well as providing an interesting design that the children can play with. The Roball has the ability to navigate autonomously and purposefully.
This article brings up interesting points with how children interact with different types of toys. Toys based upon electronic parts are susceptible to breaking more than a standard plastic toy. Electronic toys are essential to a healthy development for children do to the interaction aspect. Interactive toys can help the child to develop the social and cognitive skills that are severely missed from sitting in front of a television. With how children react to the Roball, one can see that the robot could essentially grow with the child becoming a sort of best friend. The potential for devices like this is remarkable. The barriers for these devices are of course robustness of the design, to withstand abuse as well as a reasonable cost for market viability.
As part of the social interaction with the children, it was interesting in how the Roball would react to certain states it was in with the child. For instance, if the Roball asked to be spun and the child shook the Roball, the Roball would respond accordingly, letting the child know that it was not happy. I find the social interaction fascinating because even though the Roball does not appear as a human form, the children still respond to the requests as if it was a live creature. Another way that the development team modeled Roball after a living creature is how they mapped common robot functions to human functions. For instance, if the Roball was running low on batteries, it would act hungry. If the Roball was stuck, it would act like it was in distress. If the Roball was not actively playing, it would simulate a sleep state and wake up when the child interacted with it.
The part of the architecture that should be reevaluated is the behavioral model that they chose. They utilized the Subsumption architecture, while a three layer approach would have been more scalable for their implementation. For the purposes of the Roball as a toy, the Subsumption architecture worked extremely well, acting as a living creature. Adding this software module to any interactive toy would greatly increase the interactivity from the child because of the humanlike qualities the robot exhibits. As the materials for toys get cheaper and stronger, there will more room for electronics and therefore more interactive and adaptable intelligent robots.
Reference:
- Michaud,François. “Roball, the Rolling Robot.” <http://robotics.usc.edu/~maja/teaching/cs584/papers/roball.pdf>
