Variable thrower

[bbradf44]

We have a motor driven thrower arm driven by a 2 cim gearbox on jags. We are trying to make it so a camera of some sort reads the distance and adjusts the power of the thrower accordingly. How would we go about doing that in both electrical and programing?
We have an axis camera, kinect, rasbery pi(were not entirely sure how to legally use it so instructions for that would be helpful as well) and an encoder on the gearbox running the thrower

[Jeffy]

Originally Posted by bbradf44

We have a motor driven thrower arm driven by a 2 cim gearbox on jags. We are trying to make it so a camera of some sort reads the distance and adjusts the power of the thrower accordingly. How would we go about doing that in both electrical and programing?
We have an axis camera, kinect, rasbery pi(were not entirely sure how to legally use it so instructions for that would be helpful as well) and an encoder on the gearbox running the thrower

My understanding of how a camera distance sensing works:
1. Use a light to light up the retro reflective material.
2. Use camera and software to detect and measure the size of the retro reflective material
3. use this data to triangulate your distance

OR I suggest you use an ultrasonic sensor
The process
1. mount sensor high
2. read voltage from signal
3. use calibration factor to know how far you are.
These sensors are pretty darn accurate.
I would find another supplier than robotmarketplace because of ship time, but other than that, this is a great sensor.

OR
Create a mechanism with the so called "sweet-spot" so that you can use the same power, but score from many locations with the same exact shot pattern.

[rwkling1]

I would say (like Jeffy did above), make a sweet spot for your shooter, I realize you guys made this device so you can shoot from multiple places, but you can get the same effect by creating a good arc from your ball. It simplifies everything and yet is very effective

[bbradf44]

Because of weather we actually created our thrower in our 45lb withholding allowance. We weren't expecting to be able to vary the power but after finishing and testing today we found it has far more speed and power than we thought. Do you know of any way we can use a kinect and rasbery pi?

[geomapguy]

Originally Posted by bbradf44

Because of weather we actually created our thrower in our 45lb withholding allowance. We weren't expecting to be able to vary the power but after finishing and testing today we found it has far more speed and power than we thought. Do you know of any way we can use a kinect and rasbery pi?

It'd be hard to put together quickly.....when's your first regional?

[bbradf44]

We have until Peachtree. I wasn't sure if we would be able to pull it off but we have some very skilled programmers and Linux guys so I figured it was worth a shot. We also need to know how to get the pi to communicate with the crio legally

[bbradf44]

We are also planning to use a different laptop for our driver station so if it would be easier to put our classmate on the robot and do it that way we can look into that. I don't know how to do that legally either though. This is the first time my team has experimented with raspi or on board computers for processing

[geomapguy]

Originally Posted by bbradf44

We have until Peachtree. I wasn't sure if we would be able to pull it off but we have some very skilled programmers and Linux guys so I figured it was worth a shot. We also need to know how to get the pi to communicate with the crio legally

You could have the pi send I2C(not sure if Pi does this), analog voltage proportional to a distance or even use DIO to use timing and see how long the time between pulses is (similar to what a true ultrasonic does)..

many options....I'm not too good with network tables but you could also use network tables as an option..

EDIT: and yes you could put your classmate on the robot

[bbradf44]

How would we put the classmate on the robot? Is there a paper or something I can go to where it explains it in detail?

[geomapguy]

Originally Posted by bbradf44

How would we put the classmate on the robot? Is there a paper or something I can go to where it explains it in detail?

Not sure but if you want to do RPi processing....I think Yash101 has created code to determine distance to the retroflective tape

http://www.chiefdelphi.com/media/photos/39569

Driver side code might be a bette roption

[Jeffy]

Using camera (or kinect) vision is seriously over complicating this problem. You only need to know distance to the wall. So, don't use something designed to give you distance, shapes and color.

If you are going for the hot goals though, vision could be helpful. I bet you could detect them without it though.

[geomapguy]

Originally Posted by Jeffy

If you are going for the hot goals though, vision could be helpful. I bet you could detect them without it though.

*cough* kinect *cough*

[bbradf44]

Originally Posted by Jeffy

Using camera (or kinect) vision is seriously over complicating this problem. You only need to know distance to the wall. So, don't use something designed to give you distance, shapes and color.

If you are going for the hot goals though, vision could be helpful. I bet you could detect them without it though.

All I need is distance to the wall. But we don't have funds to get a fancy range sensor so we were thinking we could get that data from the kinect. We actually don't need any video at all

[Jeffy]

Originally Posted by geomapguy

*cough* kinect *cough*

I don't get it. The kinect is surely an option. But as I said, I think it is a more complex solution than is needed.
Are we having a conversation, or are we just saying words?

Originally Posted by bbradf44

All I need is distance to the wall. But we don't have funds to get a fancy range sensor so we were thinking we could get that data from the kinect. We actually don't need any video at all

The sensor is not "fancy" by any means of its implementation. And some ultrasonic sensors cost as little as 5 dollars.

[geomapguy]

Originally Posted by Jeffy

I don't get it. The kinect is surely an option. But as I said, I think it is a more complex solution than is needed.
Are we having a conversation, or are we just saying words?



The sensor is not "fancy" by any means of its implementation. And some ultrasonic sensors cost as little as 5 dollars.

You don't have to use the kinect in a complex situation for hot goal detection...you can just use human eyes

And yes many teams forget that we got a $35 voucher to a sonar sensor company...maxbotix