Robot Location Triangulation Using IR Beacons: Improving LED Sensitivity

[TuviaDaCat]

im doing this home made project, the purpose is to build a robot which can find its location inside a room by doing triangulation to 3 Infra red beacons set inside the room.

the robot does the scanning by turning a photo sensor by a servo motor and it collects the brightest spots.

now after a lot of trial and error i have concluded to use both LED and phototransistor to find the bright spots. the photo transistor is very sensitive and gives me analog data(by POT) though it is also sensitive to bright reflections of ambient and flourescent light which makes it impossible to find the beacons. so i use a LED as a receptor because it is reactive only to the other infra LEDs as a digital component that tells me that i am in the vicinity of the beacon, and only then the phototransistor comes into action(by finding the brightest angle in that area).

now the thing is: the led is not sensitive enough, only sufficient to about 12 cm, and i need to improve the distance to two meters. i tried making an array of LEDs packed togather in a cascade, though it resulted in an odd manner which i could not explain...

so how can i improve the signal? or maybe anyone here has a better idea for a sensor to find the leds?(pass filters?)

 

[omkar13]

My dear friend,there are 2 types of leds.Edge emitting and surface emitting.Of them edge emitting leds are better.Just check the properties of it.

 

[TuviaDaCat]

but what is the chance that i will find an egde diode the same wavelength as the surface leds of the beacon?

 

[TuviaDaCat]

and still i need to know if it is wrong to use led sensors in a cascade to get better sensitivity.

 

[omkar13]

I think you might be knowing the wavelength of current leds.Just search for other type.I think we can find because the basic difference between two types is that surface emitting led emits in a circular fashion and edge one in elliptical pattern.I din't understand what is meant by odd manner.

 

[TuviaDaCat]

well i made a small test, I've put the sensor LED array infront of the beacon(which gives constant light intensity) and probed the sensor LED array with a multi-tester to see how much voltage it would give at a distance.
now at zero range it would give very little as if it was one LED, then at a small distance of 4 cm it would give 1.6 volt, which is very nice, and then further on it would fall again to very low voltage.
so in big distances it would give about the same result as a single led, which does not make any sense to me...

 

[skeptic2]

How much voltage do you get from a single LED at 12 cm? To increase that range to 2 meters you would have to amplify the voltage from the LED by (2/0.12)^2 or about 278 times or about 49 dB. This is doable.

Are you pulsing the beacons in order to distinguish them from background infrared? If not you might try that along with tuning your detector for the same frequency. There exist infrared detectors, all in one package, with very high gains for use with remote controls for televisions which would probably give you more range than you need. I've seen them with gains as high as 86 dB.

 

[TuviaDaCat]

i have already thought about pulsing the diodes, thing is, i do not know how to POT for photo-transistor which resistance goes up and down. i must know the exact resistance, for each beacon i am searching for the angle of the maximum light intensity to know that i am on the center of it.
is there a way for find light intensity for a pulsing light?

edit: I've just tried it at two meters, it would give 3mV and gives nothing for ambient light when covering the led with a tube.
so I am thinking of using a comparator of the signal with 2 mV... i hope it would do the trick...

 

[skeptic2]

I presume your photo-transistor has a collector resistor so you're getting a voltage out at the collector. When you pulse the beacon, the signal strength is proportional to the peak to peak voltage at the collector. Since you can AC couple the varying AC you don't have to worry about your bias voltage affecting the results.

Most commercial controls use a frequency close to 39 kHz. At that frequency or lower you could easily use an opamp to get the desired amplification.

 

[TuviaDaCat]

I presume your photo-transistor has a collector resistor so you're getting a voltage out at the collector. When you pulse the beacon, the signal strength is proportional to the peak to peak voltage at the collector. Since you can AC couple the varying AC you don't have to worry about your bias voltage affecting the results.

Most commercial controls use a frequency close to 39 kHz. At that frequency or lower you could easily use an opamp to get the desired amplification.

hmmm, thing is I am testing the light intensity through RC time(varying resistance, constant capacitance). now if i use a high pass filter i would need to measure the voltage instead(again, with RC time, only now the voltage it discharges from will very) i would need to worry a about the resolution, which may vary from 2mv to 1000mv...

im thinking about using both LED and the photoresistor, it would only demand of me to amplify the signal of the LED sensor as a Digital input, the POT for the brightest spot, knowing that i am looking at the beacon, and not some reflective object

 

[skeptic2]

http://davebodnar.com/railway/Pulsed_IR_Article/

 

[TuviaDaCat]

the reciever in the link is already designed for a certain frequency, i need to build a pass filter

edit: or maybe a can salvage a remote control! thank you =O