GENERAL VIEWS: For the past 4 years, I have always been very excited by the Trinity College Home Fire Fighting Robot Contest. I have attended the contest for 3 of the past 4 years, and for the first 3 years I was interested in entering, but never seemed to have the time to pull together an entry.

Well, in Y2K, I decided it was time. So, I started with looking at all the competitors I had seen in the past three competitions, and started to build a plan to attack the “house” and find the candle.

STEP #1 - PUTTING OUT THE CANDLE: I started to think about how to put out the flame. I have seen C02 solutions from tanks, soda bottles, and balloons. But none of them seemed simple or foolproof.

I saw solutions that shot water, but they seemed messy, and they seem to require pretty decent accuracy to hit the flame.

I even looked at things like a candle “snuffer”, like you see used in church. I thought that might be interesting...but that requires alot of accurate positioning.

In the end, I went with what seems to be the favorite solution, and a very simple one....a fan to blow out the candle.

But I decided that I had seen alot of “wimpy” fans that also had problems....and that I would not fall prey to that problem. I decided to put a “strong’ motor on my design, so I opted for an Electric RC plane motor. (a cheap one for about $6.00 surplus)

STEP #2 - THE DRIVE SYSTEM: My next question, was “How am I going to drive this robot”. I set down a set of requirements that would help me come to a solution.

• It must be able to turn in its own body area
• It must be able to run in a straight line without a large amount of control overhead.
• It must be able to turn 90 degree accurately and without great control overhead.

These were the basic specs I put forward, and these were what was going to help me decide on a drive system.

I decided on a stepper motor design, which turned out to be a mistake. (It would have been fine had I picked motors with enough power.)

The pictures of the robot pictured currently have stepper motors. These will have to be changed, and I am working on a solution as I speak.

STEP #3 - CANDLE DETECTION: Again, there are many ways to skin the cat here as well, but I had decided that the only sure fire detector that worked well was the Hamamatsu Flame Detector. Everyone that I had spoken to using this solution was very happy with it. Many said “it was the only part of my development that worked right out of the box”.

I wanted this sort of results, and I was also told that if you jumpered it to the most sensitive position (undocumented) it would detect a candle in a room from outside the door! (and I saw several entries doing just that last year....detecting the candle as they drove by a room)

The  only question was, how do I get this sensor to look out three sides of the robot in order to detect candles in side rooms, as well as in front of the machine.

Well, the answer came from positioning the Hamamatsu sensor via a servo. That servo gives me the ability to scan the sensor around through 180 degrees!

It also means I don’t have to deal with mirrors or other light bending mechanisms.

STEP #4 - VEHICLE NAVIGATION: Dead reckoning is not a very good idea, but I chose stepper motors because I wanted the robot to be able to position itself as accurately and consistantly as possible. But I knew I could not rely on this to get everything done.

So I decided that taking a local reference would be the best tack. I decided that understanding the hallways and rooms, and then deciding where I was by measuring my distances from the walls would work well.

I needed a fairly fast and easy meathod, and I chose the Sharp GP2D02. This sensor was controlled via 2 lines, and was an all digital interface. It was simple enough to use with the Basic Stamp, or a very capable PIC.

But these devices are not cheap, and I wanted to be able to measure in many different directions to do things like detecting corners of walls and such. (also to scan for the distance from the candle)

I chose to mount a SHARP sensor on a servo in the rear and in the front. I can then move the servo under program control and measure any distance within the 180 degrees in front and in the rear.

STEP #5 - ROBOT BRAINS: My preference here was for a simple development system, but with the horsepower and flexibility to do anything I needed to.

Basic Stamps were quickly discounted because they are so limited and slow. I an expernced with PIC microcontrolers,, so I naturally gravitated to them.

Believe it or not, I love to program in Pascal....and since there were about three different Pascal programming Languages, I decided on the PIC. (JAL, SIL, and P2C)

I chose P2C for many reasons, and it only costs $50.00. (the other two are free)

More robot information

• Robot Schematics
• New Servo Motor Controller Schematics (to replace the steppers)

[Home] [Resume] [Gov. -n-Politics] [Contact Me] [My Diet] [My Work] [My Family & Me] [Hobbies] [My Cars] [Volunteer Work]