One thing I like most about robots, is the possibility to make things in the physical world that are impossible to do so for humans. This is based upon the fact that - in theory - every physical constraint can be disappear by accurate constructing of certain body parts.
This one here for example would require tons of practice for a human to do so. It is a video of a Beluga (Weißwal), that blows rings of air into the ocean. They are beautiful to look at and not less than the counterpart for having rings of smoke in the air. Anyone already constructing this? Make it a friendly underwater cannon! (Here is another video)
Maria Ralph and Medhat A. Moussa from the the University of Guelph in Ontario made a study about "Talking to Robots". They wanted participants, that they should begin to talk to a robot-arm in order to let the robots collect things that are normally around in the household like spoons and keys.
The individuals were given a list of a handful of simple phrases that they could use to verbally operate the robot, and they were also permitted and encouraged to develop new phrases that they thought might help the robot perform its task.
Instead of using speech recognition software, the researchers trained human operators to translate the participants' words into movements via a graphical user interface. When participants made up new commands, the operator prompted them to define these new commands with a series of simple commands from the list.
The researchers figured out that the participants tend to use simple phrases more than complex ones. The teaching progress of the human - robot interaction was similar to the teaching progress of small children.
People also tended to humanize their language compared with the simple commands. For example, simple commands such as "move left" became "move closer to me" or "move this way."
The participants also tended to encourage the robot when it was doing well, providing feedback much like humans give to children. For instance, they used phrases such as "you're almost there" and "you've got it" for correct motions, "that's it" for successful moves, and "that's wrong" for incorrect moves. When using the same commands in sequence, sometimes people left out the actual commands, and replaced them with words such as "again" or "keep going."
It is also noted, that each participant developed an unique language to talk to the robot. The design of a robot to understand humans should take this into consideration.
Alan Winfield, a robotics researcher at the University of the West of England in Bristol, has fresh wisdom about swarm robotics at hand. He told one of the writers from the New Scientist:
"Software bugs are emergent properties - but today we think they're bad and must be fixed. We need to know how to design the kind of emergent 'bugs' that produce swarm intelligence in nature."
Software development could face a significant shift if we change the meaning of what software really is. Evidentially all software development methods at present try to reach totally bug-free code. We never learned, at least not systematically, to build systems where we can explore and utilize bugs. The only things I can think of pointing towards this direction are glitch and circuit bending. Maybe more! Feel free to comment!
Benoit Espinola wrote an answer to our Hypercube posting. He analyzes, whether the Hypercube would be possible. A real hypercube would not be possible, because of te 4th dimension. The results of his analysis:
What we see (in the hypercube animation) is a hypercube that remains still in our three dimensions and that is manipulated in the 4th dimension. To have that resultant movement, we have to be able to manipulate the 4th dimension of the hypercube, if we suppose that we reach to build a hypercube, we would be able to manipulate the 4th dimension... To do so, we need hyper dimensional engines. Really complicated in my eyes...
Far more essential is the closing statement of his article.
But what forbids us to inspire ourselves with the hypercube transformations to find new ways to move? Nearly nothing, except the physical constraints and costs.
BEAM is a school of robotics, starting from simple reflexes, in a 'bottoms-up' approach. The majority of BEAM robots are non-computerized (although simple CPUs can be used to drive them, in a 'horse and rider' sort of way). Unlike many traditional processor-based robots, BEAM robots are cheap, simple, and can be built by a hobbyist with basic skills in a matter of hours. Because of this, BEAM is an excellent way of getting started in robotics, and of learning about electronics.
First. I think that the key for robots making a vivid impression are not gestures, but mainly reaction. If you threw something at the robot and he will react, than we will get the impression of something that lives. That gets me to the second point.
Second. If we create robots, and that is what we do, when will be the point where we say: hey, that bot is alive? When is the point when we give birth to that robot? This really is something else than playing god, but at least I think that this point is worth discussing although we know that robots never will be life.
This weblog is dedicated to fresh and alternate views on robotics. We will focus especially in music robots, toy robots, play and game and robots as well as the culture of robots. We are the the first robot-blog from Germany - and maybe from Europe - that write regularly on this topic! This blog is made by the same people around the Digital Tools Magazine. We will have some nice surprises later this year, so stay tuned!
