paparazzi-devel
[Top][All Lists]
Advanced

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

RE: [Paparazzi-devel] Re: Paparazzi-devel Digest, Vol 43, Issue 5


From: akochevar
Subject: RE: [Paparazzi-devel] Re: Paparazzi-devel Digest, Vol 43, Issue 5
Date: Fri, 12 Oct 2007 06:29:08 -0700
User-agent: Web-Based Email 4.11.5

David,
In short, no this sensor is not all you need.  The reason for this is that it is an inertial based sensor.  It will output both rates and accelerations, but give you no information on attitude.  Without attitude the aircraft wont know which way is up as we like to say.  People who use these sensors apply various filters, the most common of which is a kalman filter to take the data from these noisy and drifting sensors, and try to get an accurate attitude model.  They are prone to getting confused simply because the new position is always based on the previous one, so care must be taken when writing the filter, and to put limits and bounds on the inputs from the sensors. 
 
In the case of hte IR, although "Crude" as you say, they output an angle.  This absolute attitude means the plane can not get confused of 1 blip of data is bad, because it will just use the next.  These sensors have proved they work over water, in deep cold, and even in snow I think(iceland team chime in here).  So although crude, the proof is in the demonstration success.  Look at the record of MAV competitions the last 3 years and you will see its a IR based airplane on top almost every time.  Not bad for some FMA hobby crap huh?
 
Anton
-------- Original Message --------
From: "David Conger" <address@hidden>
Date: Thu, October 11, 2007 11:25 pm
To: address@hidden

Greetings again,

I realize even after building the tiny to "fly" the plane I will need some way for the autopilot to know the attitude and pitch/yaw/roll that is going on and correct...IR sensors to me seem super crude.

Has anyone seen these: http://www.analog.com/UploadedFiles/Data_Sheets/ADIS16350_16355.pdf

Seems to have everything necessary in one place. Would this work with a Tiny? Anyone willing to hint me to the right resource to learn how I would do that? I'm assuming tiny would just need to "read" the outputs from this and simply apply the same PID logic correct?

I'm guessing in short time these parts will come down in price considerably.

Also, I have been lucky. Someone I work with has offered to build me 2 Tiny's. I work for a large company that has labs with all the equipment. So, he's an awesome resource for me. Seems he found several issues with either the Mouser Parts and the Circuit Board I'll share when he debriefs me on completion of the first board. I also hear there is a company nearby that will build/fabricate circuit boards for me from designs. I will ask him about the tiny 2.1. So, a big leap forward on my end. Is the Eagletree design ready for fabrication that's in CVS?

Best Regards,
David Conger



_______________________________________________
Paparazzi-devel mailing list
Paparazzi-devel@nongnu.org
http://lists.nongnu.org/mailman/listinfo/paparazzi-devel

reply via email to

[Prev in Thread] Current Thread [Next in Thread]