Efficient Sensor for Candle Detection in Firefighting Robot Competition

[alphadog0309]

i am looking into participating in a local firefighting robot competition and i have many of the parts necessary, one thing i have yet to decide on is my method of finding the candle which is placed in a random room. Originally i had intended to use an infrared sensor with a filter, but i was unsure on whether it would be able to detect a candle from about 3-4 feet away... anyone have any experience with these?

 

[mgb_phys]

Something like the passive IR sensor from a home security alarm should work. Ideally you want a couple with some sort of metal 'lens hood' so you can get a better idea of direction

 

[alphadog0309]

by couple do you mean like 2 or 3 sensors to detect the light??

 

[mgb_phys]

by couple do you mean like 2 or 3 sensors to detect the light??

Yes facing in different directions so you know whihc way to turn and can subtract the background level for the room. It's tricky to make optics at these wavelengths so you want a metal tube to limit the angle each sensor can 'see'

 

[alphadog0309]

alright so here is the basic premise of the task:

http://www.youtube.com/watch?v=6tl9...irefighting/index.htm&feature=player_embedded

im planning to use an arduino control board with a motor controller and build an infrared sensor that will pick up the light along with a sonar sensor to detect distance from the walls, a light sensor to detect the white lines, and motor encoders for precise turning and distance measuring

my physics teacher (and robotics advisor) just got two of these (http://www.allelectronics.com/make-a-store/item/IRD-10/INFRARED-DETECTOR-MODULE/-/1.html)... should they work?

 

[Topher925]

What you want is a UVTron or a sensor designed to pick up UV radiation from something burning.

http://www.acroname.com/robotics/parts/R66-R2868.html

Hamamatsu sells these as entire packages which output a PWM proportional the amount of UV light they receive.

http://www.acroname.com/robotics/info/articles/trin2001/trin2001.html

 

[alphadog0309]

What's the range on one of those things? I'm looking into something with about 2-3 ft sensitivity

 

[Mike_In_Plano]

Cool Project!

I'd suggest some melexis sensors from future electronics:

http://www.futureelectronics.com/en...umberUpshifted|*MLX90614*|1|,Ny:True,Nea:True

I was used to working with much larger, more expensive sensors that had germanium lenses. Though not as tightly calibrated or directional, the melexis parts are easily 1/10 the price and easier to interface. Also, they didn't have the handling issues that come with germanium (every fingerprint - a lasting memory...).

I'd put the little guys on a polygon, with one every sixty degrees. This way, they overlap sufficeint for you to spot the places between. That will give you 6 channels to buffer and acquire.

With the original melexis parts, I used chopper-stabilized op amps from Linear Tech to buffer them. they part easily spotted me from 5-6 feet away. With something as hot as a candle, the long wave IR output should be enourmous! (Refer to Boltznam's distribution).

Good luck and have fun,

- Mike

 

[alphadog0309]

awesome! thanks for the help i bought a few sensors off a surplus store that look similar... thanks!