IR Distance Sensor (not a sharp! wahoo) for my xmos project

So I had an idea (because i i like building my own stuff) of an IR range finder very similar in fact to odbots attempt at a laser rangefinder combined with CtC's ir beacon. So a narrow beam IR LED is mounted on a servo or stepper (or several if more power is needed). A IR transistor (one of those TV remote receivers) is mounted next to it at a known distance inside a tube similar to that in CtC's design. The servo or stepper scans from pointing towards the sensor (near perpendicularly) outwards near parallel while the led is being pulsed at the appropriate frequency. When the sensor detects an IR signal the servo or stepper stops swinging. Because this angle is known because of the use of the servo, and the distance between the sensor and the led is known the simple math of distance/cos(angle) the distance from the sensor to the object can be detected. the reason for the tube around the sensor is to only allow the IR signal to reach the sensor when the light from the led is directly in front of the sensor. This was a bit confusing. Time for a picture

IR_sensor.jpg

What do you think?

Sounds like a sweet idea.My
Sounds like a sweet idea.
My one recommendation is that instead of stopping when the IR signal is first detected, keep scanning until the IR signal is no longer received. Then you can find the average of the ‘first detected’ and ‘last detected’ angles, and use that in your triangulation calculations. This should help with accurately detecting the proximity of targets that reflect the IR a bit more erratically.

A good plan! I had thought

A good plan! I had thought of a similar thing with a cheap laser diode like Oddbots laser rangefinder but having the sensor/laser mounted on a servo. I scrapped the idea because I didnt think the servo would be able to make fine and accurate enough movements, especially as the distance to the object increases, the important range or angles gets smaller.</p><p>Ill be interested to see how it turns out.

i like that idea (@TeleFox)
thats a good idea, find the two angles and find the average of the two for the center.

Not the first, but interesting
Parallax has been doing similar for years with it’s boe bot kits but shorter range. Sounds interesting but sharp still rules. Why re invent the wheel? It’s all in the software.