Sharp IR distance sensors

[Gdeaver]

As far as autonomous navigation I've seen allot of posts on encoders, gyro's, accelerometers, Banner sensors but, nothing on Sharp IR distance sensors. Today I've been playing with a Sharp GP2Y0A02YK 5' max distance sensor mounted on a Vex servo with a square bot and encoders. Using only Easy C 1.1 I've got the square bot driving straight, random drive with object avoidance, Wall hugging with just the IR and several other autonomous strategies. The Sharp IR sensors put out a analog voltage proportional to distance. A simple and quick ADC gives an accurate distance measurement with a very narrow field of focus. Mounting the IR sensor on a servo allows a scan and mapping of the area in front of the robot. Have any other teams tried these sensors? I haven't seen any at competitions or remember any posts.
So here's a link to a site that sells them
http://www.hobbyengineering.com/SectionS.html
Any comments on their application on FRC robots?

[Rickertsen2]

Originally Posted by Gdeaver

As far as autonomous navigation I've seen allot of posts on encoders, gyro's, accelerometers, Banner sensors but, nothing on Sharp IR distance sensors. Today I've been playing with a Sharp GP2Y0A02YK 5' max distance sensor mounted on a Vex servo with a square bot and encoders. Using only Easy C 1.1 I've got the square bot driving straight, random drive with object avoidance, Wall hugging with just the IR and several other autonomous strategies. The Sharp IR sensors put out a analog voltage proportional to distance. A simple and quick ADC gives an accurate distance measurement with a very narrow field of focus. Mounting the IR sensor on a servo allows a scan and mapping of the area in front of the robot. Have any other teams tried these sensors? I haven't seen any at competitions or remember any posts.
So here's a link to a site that sells them
http://www.hobbyengineering.com/SectionS.html
Any comments on their application on FRC robots?

I used a few of these once in a non-FIRST bot that wonders around the room avoiding obstacles. I also made a wall follower out of one of our practice bots using these. They work well from my experience. The only drawback is that they are not directly porportional to distance, but for most applicaitons this does not matter, and if it does, it they can be linearized fairly easily.

[Mike]

As I understand, the Sharp sensors have a maximum range of 30 inches... 3.5 feet. When you have a robot moving 7 feet per second (conservative estimate), that means you have one half a second to "see" the object, make the calculations to find out which way to turn to avoid the object, and then actually make the motor adjustments to turn, and then actually turn. Ultrasonic sensors, however, can have a range substantially bigger. The SRF08 ($56.99) has a range from 1 inch to 18 feet.

This may work if you were to stop once you have detected an object, and then turn around your axis and continue forward, but that would be an awfully slow and inefficient way of doing things. This makes the sensors not very useful for object avoidance. It may be of good use for say, not ramming straight into a wall. This would be executed by mounting it on a servo, then once an object is detected, scan left and right to make sure the object is a wide flat area (and not say, a pole.) It would also be good for wall-following, but (at least in this years game) you would follow the wall for about 2 seconds and then ram straight into a side goal.

It all comes down to what you are trying to accomplish. When you are playing on a pre-determined surface (such as the FIRST field), I find an encoder/gyro combination to be a lot better. If you are in unfamiliar territory, then an ultrasonic/infrared sensor will be a better choice.

[TimCraig]

Originally Posted by Mike

30 inches... 3.5 feet

[Gdeaver]

The listed sensor measures out to 5'. The other closer range sensors could be used to detect when a game piece was in grasping range and position.

[devicenull]

We used them last year to measure the height of a component on our robot. As I recall, they worked quite well, except when we managed to loosen one of the solder joints

[Mike]

Originally Posted by TimCraig

Haha, this is why I should stop making long posts that early in the morning.

[Andrew Rudolph]

There is more than one kind of sharp sensor, heres a quick list of whats offered at acroname.com....

Sharp GP2D02 IR Ranger
Very popular 8-bit IR Ranger with range of about 3"-30".
Sharp GP2D120 IR Sensor
Reports distance as an analog voltage with a range of 1.5" to 12".
Sharp GP2D05 IR ranger
A high precision object sensor.
Sharp GP2D12 Detector Package
Reports distance as an analog voltage with a range of 4"-30".
Sharp GP2D15 Detector Package
Digital version of the popular Sharp IR detector preset to 10".
Sharp GP2Y0D02YK IR Sensor
Digital version of the popular Sharp IR detector preset to 31.5".
Sharp GP2Y0A02YK IR Sensor
Reports distance as an analog voltage with a range of ~8"-60".

We used GP2D12's on 1083's robot in 2004 for autonomous, we would drive at the wall, use the sensors to triangulate us following the wall, then encoders would track how far we had traveled so the bot would stop and punch the ball off the Tee. They worked pretty flawlessly once we got the numbers right and shielded them enough from the surrounding light.

[Gdeaver]

Andrew's post was what I was hopping to see. The Sharp IR sensors compliment encoders and gyros. The also can replace limit switches. The are relatively cheap and easy to interface. Hopefully some teams took note.

[ktrussell]

I've seen mention that one "MAY" need to shield the sensor from the ambient light. Those of you who used the Sharp sensors previously, did you have to provide special shielding from light? The year that FIRST had the IR beacons, we had our robot tracking to the beacons very well on practice day but on Friday, when they turned on the bright lights for the TV cameras, our receivers were blinded. Will the Sharp distance sensors experience anything like this?

[yongkimleng]

From my own use in the past, one point is not to have direct strong light fall onto the sharp IR sensor, even though it has a recessed receiver. Other than that, it does its own ambient light compensation quite well.
Refer to the item's datasheet for more information.
P.S. the soldier joint on its connectors are very weak.

[dtengineering]

We've had very good luck with a wide range of Sharp IR sensors.

Just make sure that if you are planning to use them this year for detecting any parts of the rack that you test them using diamond plate. We found that the diamond plate was so reflective that if we were at anything other than exactly 90 degrees to the plate that the IR would simply bounce off... and not return to the sensor.

We're thinking sonar is the answer for detecting diamond plate, but will use the IR for monitoring the height of our lift.

Jason

[ktrussell]

Thanks for that tip, Jason. I saw in the datasheet that mirrors could cause a problem. I had assumed the diamond plate would be OK. I was wanting to use IR for that to avoid the more likely chance of interference from others Ultrasonic. I plan to experiment with both, however, today as a matter of fact. Now I've got to find some diamond plate..I wonder if aluminum foil would be a good substitute.

[ktrussell]

I tried aluminum foil wrapped around cardboard and found exactly the problem Jason describes. As I sweep the foil through angles off 90 degrees I find various spots where the object seems to disappear to the sensor. I would say it isn't totally useless by any means for detecting the diamond plate (well, my simulated diamond plate, at least) but one will need to be aware of its faults.