Developer
Adam Rivera
  • Milwaukee, WI
  • United States
Share

Adam Rivera's Friends

  • Andy Sousa
  • Chris Norman
  • Erik Grinnell
  • james johnson
  • shanthosh ravikumar
  • Steve Maller
  • Jason Wennberg
  • cirilo
  • Damon Amberstone
  • Brian Prange
  • Shyam Balasubramanian
  • Reuben
  • Luc van Helfteren
  • Nathan Reim
  • Tyson Baldwin

Adam Rivera's Groups

Adam Rivera's Discussions

FS: PhotoShipOne 2X Gimbal

Started this discussion. Last reply by Adam Rivera Apr 2, 2012. 1 Reply

I am selling my photoshipone 2x gimbal: http://photoshipone.com/products/2x/I will be very honest. I was not pleased with the tolerances of this…Continue

Tags: gimbal, fs

XBee 900Mhz Telemetry Stopped Working

Started this discussion. Last reply by Mattia Parise Mar 14, 2012. 10 Replies

Hello all,My telemetry stopped working recently. When I attempt to connect using the Mission Planner at 57600 baud rate, it times out. When I look at the command prompt output it looks like it isn't…Continue

Tags: telemetry not working, telemetry, xbee

ArduCopter FlightGear HIL

Started this discussion. Last reply by Dan Banks Mar 23, 2013. 6 Replies

Does anyone have experience with running FlightGear as a HIL simulator for ArduCopter?FlightGear launches and I see the quad on the runway, but when I use my radio to arm the motors and take off,…Continue

Tags: flightgear, simulator, HIL

Simple mode controls reversed

Started this discussion. Last reply by RickP Jan 12, 2012. 2 Replies

Flying latest stable release.I armed my octo with a 3d gps fix and started to fly in simple mode. This is one of my first times flying in simple mode so I was just trying to get a feel for it and…Continue

Tags: arducopter-modes, simple, simple-mode

 

Adam Rivera's Page

Profile Information

About Me:
I am a Development Manager and Technical Architect for GMR Marketing in New Berlin, WI.

I am also the President/Owner of Sentient Drone, LLC. We offer Droidworx systems and custom solutions. http://sentientdrone.com
Please tell us a bit about your UAV interest
Autonomy interests me.
Hometown:
Milwaukee

Adam Rivera's Blog

NEW FEATURE: Automatic Compass Declination

Posted on March 31, 2012 at 7:38am 29 Comments

 

Hello DiyDroners,

 

I recently completed the implementation of a new feature for ArduPlane & ArduCopter. I have to thank Jan Scherrer for the idea and especially thank Andrew Tridgell for pushing me to improve the original implementation. The new feature is automatic…

Continue

RESOLVED: ArduCopter 2.5.3 Yaw Issues

Posted on March 26, 2012 at 9:00pm 80 Comments

I just finished a long debugging session with Andrew Tridgell and he has discovered the cause of my yaw (and attitude) issues. In case you don't know Andrew Tridgell, he is the genius developer who made the the latest release fly so well with his DCM improvements.

I think the majority of…

Continue

Object Tracking - Real Test (teaser)

Posted on February 15, 2012 at 10:00pm 18 Comments

Note: This video is available in 1080p. Use it!

 

I finally have an update on the object tracking software I started back in these two posts:…

Continue

Rate PI Autotuning - A third attempt

Posted on February 3, 2012 at 11:30pm 17 Comments

Hi again all, Two blog posts in two days... i'm on a roll. I just finished the next set of revisions to my autotuning code. I have shifted from tuning the stab gains to tuning the rate gains. This is more dangerous but also more rewarding.

Here are…

Continue

Comment Wall (4 comments)

At 3:46pm on April 12, 2012,
T3
William Premerlani
said…

Hi Adam,

Regarding your question about auto PID tuning, I am doing some relevant work, but taking a slightly different approach. The idea is to determine the dynamic model of the aircraft (LaPlace transform) from flight data. Once you know the dynamic model, you can synthesize the ideal controller (its not a PID), and select optimal values of the gains. The method should work for both fixed wing and rotor craft.

So far, I have been able to determine the dynamic model of my draganflier quad, offline, and built the ideal controller for it, with the optimal parameters. The next step is to extend the method to work recursively in real time.

Right now, my work in this area is on the back burner, in favor of work that I am doing on acceleration compensation. Tridge has implemented it for ArduCopter. He says that it improves loitering performance. But after about 10 minutes, it goes unstable, for reasons that neither he nor I understand yet. Therefore, the technique is not yet ready for release. So, when I get a chance, I am going to implement it on my end and see if I can figure out what is going on.

By the way, Adam, a few years ago I did some consulting work for Johnson Controls. I stayed at the Phister Hotel. One summer, during ItaliaFest, I almost wound up in jail in Milwaukee. Would you care to hear the story?

Best regards,

Bill Premerlani

At 4:27pm on April 16, 2012,
T3
William Premerlani
said…

Hi Adam,

Ok, here is how I almost wound up in a Milwaukee jail....

A friend of mine was trying to drop me off at the Pfister Hotel one evening during ItaliaFest, arriving along a superhighway. The exit off the highway was barracaded, so my friend dropped me off at the barracade, which was manned by 3 policeman. As I walked along (quite safely, because the exit was blocked), the police asked me what was I doing. I said I was walking to the Pfister and pointed to it, it was only about 100 feet away. They said that I can't do that, no walking on the highway. I said, well there are no cars....but what do you want me to do?

They said, either get your friend to come back (not possible, this was before cell phones) or you can go to jail.

So, stalling for time, I say, "Well, I don't like either of those choices, how about I stay here with you guys and help direct traffic".

Generally, the 3 of them did not care for my attitude about that, but one of them crossed his eyes and thought about it for a few minutes (I think he was a cheese-head, at least he struck the same pose as the guy on the billboard), and then said, "Why no, you cannot do that, you do not have a safety vest and a flashlight". Really, that is what he said.

The other 2 were a bit brighter than the 3rd one, and a bit meaner, too. They did not seem to like my back-talk, I thought maybe I had just made a big mistake and was going to get sent to jail for jay-walking.

They demanded to see some ID. I showed them my New York State driver's license, which turned out to be a get-out-of-jail-free card in Milwaukee!!

When they saw my license, they assumed that I was from New York City, I guess. They said to me, "OK, we understand now. You are one of those f^$%ing a%$holes from New York. Why, you New Yorkers are almost as bad as those flat-landers from Illinois."

They gave me back my license, they stopped the next shuttle bus, put me on it, and told the driver to drop me as far from the Pfister Hotel that he could. So, I had a long walk back to the Pfister that night, and could not even get a taxi, because there were pedestrians everywhere. Plenty of time for me to think about how unpredictable my own tongue was.

And, by the way, if you are a cheese-head, please forgive me for anything that I said that might be misconstrued as insulting to cheese-heads.

Best regards,

Bill Premerlani

At 12:46pm on April 18, 2012,
T3
William Premerlani
said…

Hi Adam,

I used a spreadsheet to compute the dynamic model of my draganflier. Here is the discussion, which includes a link to a description of the theoretical basis. Here is the description of the spreadsheet. Here is the spreadsheet. The method will work no matter what the pilot and the controls are doing. The only source of error is any output that is not a result of the inputs. In other words, vibration. I had to do some pre-filtering of the data to remove vibration.

Best regards,

Bill

At 3:21pm on May 1, 2012,
T3
William Premerlani
said…

Hi Adam,

Yes indeed, I have given some thought to applying the "offset learning algorithm" to determining accelerometer offsets, but it is not high up on my list of things to do for a couple of reasons (besides the fact that I have a lot of other things on my list).

The math for the offset learning algorithm was developed for vectors that have constant magnitude. But the vector output of the accelerometers is not constant in magnitude, so the math will have to be extended. I am not sure if that can be done without making assumptions or approximations.

Also, right now, in UDB-MatrixPilot land, the residual offsets are very small to begin with, so I am afraid any learning algorithm might make the situation worse.

That said, the idea is still on my list, it is certainly worth pursuing, especially since some pilots are not very careful about setting up to avoid offsets.

Best regards,

Bill

You need to be a member of DIY Drones to add comments!

Join DIY Drones

 
 
 

DIY Drones Monthly
Newsletter

Groups

Season Two of the Trust Time Trial (T3) Contest 
A list of all T3 contests is here. The current round, the Vertical Horizontal one, is here

© 2016   Created by Chris Anderson.   Powered by

Badges  |  Report an Issue  |  Terms of Service