Ideas from Other Competitions
The competitions are the best place to see other ideas for the robots in action. yadda yadda yadda, put something here.
Sensing the Candle: One simple idea utilized by many robots effectively was the use of an IR sensor hooded by a black film container with a hole in the bottom. The sensor was threaded through the hole and left at the bottom of the film container so that it was sheilded from ambient light and could zero in on the candle most efficiently. Light sensors were also used in this capacity though did not seem to work as well. One robot used several IR sensors on a large swiveling [HALEY: sp?] panel, which worked extremely well but was apparantly very difficult to program.
Sensing the Walls: Most robots used range sensors, and they were quite effective. One robot used sonar sensors, which worked even better, but this was the robot using fuzzy logic and much of its success can be attributed to its programming.
Structure of the Robot: Here is where the real diversity of ideas came in. One robot, from a high school team, was completely round in shape and nicknamed "Oscar" due to its trash-can-like look. It swiveled easily and had a good range of motion, but was not as effective at finding the candle, probably because it moved so well and this made it difficult to stop exactly where it sensed a flame.
Another robot from the high school level utilized an XY wheel configuration that was actually an adapted remote controlled car from Radio Shack. The builders did not know how the wheels changed direction, saying "It just came that way," but the robot was extremely fast and navigated the ninety degree turns of the maze extremely well. The main drawback of this design was its need for power as it had to use remote sources to run the car's wheels.
Also seen at the competition were several robots that used [HALEY: need technical name for "Those nifty wheeley track things like they use on tanks"] This helped synchronize the wheels and helped out with the robot's ability to stop and resume quickly. If a dead reckoning program is used, this is the way to get the wheels together for slightly better accuracy.
All in all, robots that used an XY configuration worked well. Some were better than others. Still, the robot that actually won the competition was of the typical car-type wheel configuration.
Extinguishing the candle: Most robots used fans, but the fans varied in their actual ability to blow out a flame. Some were placed at angles entirely wrong and therefore failed, and others were large and cumbersome and managed to get themselves into a wrong angle by the time the robot navigated the maze. One had very large blades and could produce a very strong wind, but since the candle was placed so close to a wall, the fan blade caught on the wall and could not function. Keep size of the blades in mind when building the fan as well as motor strength. The fan we made with Greg in the machine shop works extremely well.
Programming ideas: Not as many teams were willing to outline their programming strategies. One robot used fuzzy logic. If anyone feels the need to try this, the robot was so good at finding and extinguishing candles that it won first place. Navigation is definitely most recommended rather than hard coding because robots that were hard coded tended to get lost before the second room of the maze if they happened to be the tiniest bit off at the beginning. Navigating robots could at least find the candle, even if their strategies for extinguishing it failed.