Go Down

Topic: Need general idea or concept for Quadcopter (Read 1 time) previous topic - next topic

sbright33

Working on an autonomous Quadcopter.  It's stable up and down using pressure or ultrasonic.  Faces the same direction using a compass.  Plenty accurate when it's 300ft up using GPS.  Can use laser or IR remote to find its landing spot within inches.

How do I keep it hovering in one place without slowly drifting over an arbitrary location when it's low 10ft above the ground?

GPS drifts around in a 50ft circle.  It does better in the short term doing nothing to correct drift, ignore GPS, just stay level, without wind.  That's my best result so far.  Do nothing!  Any suggesting for improvement would be appreciated.

Already have gyros and accelerometers. Tried video the way the Brookstone copter does.  Any other ideas?
If you fall... I'll be there for you!
-Floor

Skype Brighteyes3333
(262) 696-9619

sbright33

Let's say I could keep it level using some sort of theoretically perfect mercury switch.  I'm happy even without this switch, when there is no wind.  A gusting or steady wind of 5mph completely ruins this method.  In a matter of a minute it's 100ft away.  What I need is a better GPS.  Loran would be an excellent choice if it still existed.  What if I had an identical GPS on the ground to compare it to.  Keep the difference between them the same.  I have heard this wouldn't work because they might be tracking different satellites?

How about this idea?  Assume the GPS is stable for 30 seconds.  Assume the wind is consistent.  Calculate the elevator and aerilon offset to keep it still.  Then ignore GPS?  Keep those 2 settings constant.  Allow the pilot to override this with a simple button push, causing the GPS to be activated again for 30 seconds.  This would fix the problem of variation/error during the first 30 seconds period.  It would also correct for any changes in wind, when you press the button.

I need to think outside the box.  Maybe I'm too close to the problem.  Anyone?


If you fall... I'll be there for you!
-Floor

Skype Brighteyes3333
(262) 696-9619

Aussie_V8

I would think if you coded it properly, that a 4 axis or 6 axis accelerometer would work just fine?

sbright33

I don't think so.  I only have a $50 sensor.  There's a lot of high frequency vibration.  Without this sensor it moves VERY slowly in 1 direction if there's a slight breeze.  I think a $5000 accel might be more successful. 
If you fall... I'll be there for you!
-Floor

Skype Brighteyes3333
(262) 696-9619

sbright33

An example of outside of the box thinking:
Drop a lead weight attached to light fishing string 10' ?
Measure the angle of the string?

Nobody has any ideas for me?  I'm going to try a better GPS module.
If you fall... I'll be there for you!
-Floor

Skype Brighteyes3333
(262) 696-9619

KirAsh4

Erect three poles in a triangle pattern with the copter in the middle.  Bounce lasers from the copter to the poles and make sure it stays in one spot.

Hey, you did ask to think outside the box. :)

Chagrin

The way I've seen this done is using an "optical flow sensor" (an optical mouse sensor) with a lens to set the focus at a greater range.


PeterH


The way I've seen this done is using an "optical flow sensor" (an optical mouse sensor) with a lens to set the focus at a greater range.




Optical motion sensing is the only way I can think of without using an external transponder. Using an optical mouse to do this is a very elegant solution, if you get it working.
I only provide help via the forum - please do not contact me for private consultancy.

The Clever Monkey


Working on an autonomous Quadcopter.  It's stable up and down using pressure or ultrasonic.  Faces the same direction using a compass.  Plenty accurate when it's 300ft up using GPS.  Can use laser or IR remote to find its landing spot within inches.

How do I keep it hovering in one place without slowly drifting over an arbitrary location when it's low 10ft above the ground?

GPS drifts around in a 50ft circle.  It does better in the short term doing nothing to correct drift, ignore GPS, just stay level, without wind.  That's my best result so far.  Do nothing!  Any suggesting for improvement would be appreciated.


As you note, GPS is not granular to the level you need. A trick with using GPS is to use one or more known ground location(s) that is/are fixed for reference, and triangulate off of that.
I yield() for co-routines.

PeterH


As you note, GPS is not granular to the level you need. A trick with using GPS is to use one or more known ground location(s) that is/are fixed for reference, and triangulate off of that.


If you're talking about differential GPS, that helps compensate for the deliberate errors introduced to degrade the GPS accuracy, but doesn't do anything to improve the resolution. I think it's the resolution which is the problem here.
I only provide help via the forum - please do not contact me for private consultancy.

sbright33

The resolution is 1ft.  If you average it over time you get <10ft error from day to day.  But at any instant it wanders in a 50ft circle 90% of the time.  What do you mean by ground locations?  My own?  Or do you mean WAAS?  95% of the time it's supposed to be within 1m.  Unfortunately I don't get good coverage here in Maine in the woods. 

Reading about the optical flow sensor from DIYDrones.  Very cool!  But I cannot find any Arduino code by Google.  Does anyone have any code I can read?  What is the range over grass for example?  5m?

If you fall... I'll be there for you!
-Floor

Skype Brighteyes3333
(262) 696-9619

kg4wsv

Quote
If you're talking about differential GPS, that helps compensate for the deliberate errors introduced to degrade the GPS accuracy,


No, SA doesn't exist any more.

dGPS helps compensate for the error introduced by atmospheric distortion.

-j

brydenrh

I'm an arduino noob, so excuse me if this idea is unrealistic, but a wii remote has a built-in positioning/IR tracking. I'm not sure if anyone has gotten it to work with an Arduino, but it certainly works with PCs. You could set up a 2 or 3 IR light beacon somewhere on the ground and the copter could orient itself based on the wii remote output.

Go Up