People who know that I’m somewhat robot obsessed sometimes assume that I must like watching shows like Robot Wars or BattleBots. I have watched a few episodes of Robot Wars, but I’m not a fan. Here’s why.
There are no robots
Although the shows claim to feature robots, they don’t. What they do feature are large and fancy radio-controlled vehicles. “Radio Controlled Vehicle Wars” is not nearly as catchy a title though. You might claim that these vehicles look like robots and, to a certain extent, they do. However, there is no automony. Take away the human at the controls and all you have left are a heavy duty chassis, some armour and some actuators. Leave one of these “robots” powered on by itself and it would do precisely nothing (and that would not make very exciting viewing).
Back in the 19080’s I found the book How To Build Your Own Self Programming Robot by David L. Heiserman in my local library. It was a year or two after my parents bought a ZX Spectrum and I knew just enough about electronics and computers to appreciate it. This year I came across My Robots: Musings of a Hobby Robot Pioneer which I read out of a sense of nostalgia (it’s partly the author reminiscing about his work and partly extracts from his books and articles - it’s somewhat self-serving and I wouldn’t reccommend it unless you’re partly interested in David L. Heiserman). Anyway, David L. Heiserman is pretty clear about what he considers to be a robot: it needs to be autonomous and capable of altering its own behaviour. He is scathing of both the type of remote controlled vehicles featured in Robot Wars but also of robots where the human operator is replaced by a computer running a straightforward program. I suspect that Mr Heiserman would not consider NAO to be a proper robot for example.
However, the difference between NAO and Rodney is, in my opinion, just software. There is no reason why NAO could not be programmed to exhibit the sort of learning exbitited by Rodney. In fact given the more capable processor and networking that NAO is endowed with you could create something very interesting indeed. The fact that most NAO developers don’t is immaterial.
I would need to re-read “How To Build Your Own Self Programming Robot” to be sure, but I think that Mr Heiserman implemented a simple form of what we would today call reinforcement learning although I suspect he arrived at it via experiementation.
While I, personally, consider NAO to be a robot I do agree with Mr Heiserman that autonomy and an element of learning or adaptation is required to make robots more than amusing automatons. Doesn’t stop me enjoying watching NAO dance though :-)
It’s all destruction
I’m a sucker for science-fiction and action movies, fighting, explosions and all but I find it a bit sad that the only mainstream shows claiming to feature robots (even if they don’t) are all about destruction. Jay-Jay, a fellow member of the, now defunct, NAO developer program competed in robotics competitions in Korea featuring tasks (such as obstacle courses) that stretched robots and their programmers to their limits but in a non-destructive way. There was also the DARPA Robotics Challenge. Finally, there is RoboCup. These all show that TV producers could make shows of a more constructive nature if they chose to.
Links and stuff
You can find some of David L. Heiserman’s writings online: such as What is a Robot - A Matter of Semantics and Good grief! Now there’s Rodney and Buster.