Submitted by Bangarang on January 30, 2009 - 10:48pm.
In 1996, I was watching the news and heard about two soldiers injured in a mine blast in Bosnia. The land mine explosion injured one soldier's leg and right foot. The second serviceman suffered shrapnel wounds to his lower body, but was able to walk away from the scene. I came to understand they were conducting an inspection of a mine field with the Bosnian Serb army.
Almost as soon as I heard this I couldn’t help but think that there had to be a better way of doing this kind of thing. I began to think that someone should design something that could scope out a dangerous mine field before any human set foot on it. Maybe design something that could travel all over a field, even setting off unexploded charges, yet cheap enough to be no big loss when blown up. Maybe even be cheap enough to release a whole swarm of them over a few acres of land.
To build a bunch of robots with tiny computers telling them what to do didn’t seem very cost effective to me. Even an expert on the subject stated that:
"the cost of even limited-application robots is beyond what demining projects can afford, compared to the typical salaries of human deminers. Progress in robotics can change this situation, but the gap remains wide."
Even remote control vehicles would require too much man power in order to swarm a whole area in a short time. In order to do clear a mine field you would have to come up with something that could randomly cover an area, avoid obstacles, and be very, very cheap- requiring limited human interaction.
That's when I started thinking about a computer-less robot.
If you look at the picture of the ant below, you can see one lone ant simulating random movement in a designated area. Now, with even limited intelligence such as obstacle avoidance, you can see that one ant, even though he is moving around quite a bit, won’t cover the area very effectively:
What you need is a group of ants wandering around the designated area randomly. Much more ground can be covered, in less time. Of course, in order to do this you need a mobile, inexpensive, cpu-less bot so that a swarm technique can be cost effective:
These random wanderings, using a swarm approach to any task that would require ground cover such as search and rescue, or triggering surface mines, can actually spread over a given area quite effectively. Remember the swarm of ants in the above picture? Take a look at how much of the surface is covered by their wanderings
This is of course, is only one of the reasons such a design would be useful. I have always had a desire that amateur robot hobbyists move beyond glorified remote control cars and start putting some legs on their creations. This really was made evident for me after reading Sojourner, the book by the designers of the first Mars Rover. I came to understand that a wheeled robot while quite functional, had some limited applications. A walking robot, while initially more difficult, could ultimately overcome these limitations. Maybe someday, with the advancement of walking robotic techniques and design, a road-less location would not have to go unexplored.