paparazzi-devel
[Top][All Lists]
Advanced

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

Re: [Paparazzi-devel] New IMU in YAPA


From: Hector Garcia de Marina
Subject: Re: [Paparazzi-devel] New IMU in YAPA
Date: Tue, 19 Feb 2013 00:29:05 +0100

Why if only matters the quality of the algorithm or the filter, in civil aviation, military systems and space crafts use so expensive (hundreds of thousands dollars) inertial sensors if a fifty bucks IMU can be equally efficient?

Because in the worst case scenario, they can not afford loosing a lot of money, experiments or even lives. The price of a tactical IMU compared with the cost of the Space Shuttle is not comparable for instance, and they can (could actually ) cover the all reentry only
integrating gyros for instance.

The flight duration of a quadcopter is at most one hour as long as I have seen by internet and typical flights are below 30 minutes. But if a quadcopter was going to still for more than 6 hours, also the same IMU would be sufficient for insuring the right attitude determination?

Basically the maths of the filter (in the ideal case which is close to the actual one) tell you that the error in your estimation is bounded (covariance) with a 99% of confidence. As Christophe pointed out, for a quad, a few degrees in the attitude is not
a significant problem. For fixed wing, it should not be neither if you are only interested in flying and take pictures. However, if you are more interested in flying "perfectly" trimmed, or you want to estimate your AoA with accuracy, or focusing on an "optimal problem", then probably you want to aim at more expensive IMUs.

That always depends on your requirements.

Héctor





On Mon, Feb 18, 2013 at 11:01 PM, Christophe De Wagter <address@hidden> wrote:
For quadrotors a few degrees of attitude error are usually no problem (e.g. because the trim attitude to fight e.g. wind is much more than that and the outerloop takes care of it, and because there are no sustained rotations, and because magneto bias is a much larger problem than accel/gyro biases etc)

For fixedwing that do sustained turns, a g-load-dependent bias and vibration-dependent bias on cheap gyro's combined with large accelerometer vibrations is not so trivial. And while mechanical damping is always a good start, you can reduce the feedback gains of the AHRS filters as the gyro's get better. For those who can afford it, it sounds like a sensible thing to do.

-Christophe 


On Mon, Feb 18, 2013 at 10:40 PM, Mauro Garcia Acero <address@hidden> wrote:

Dear Hienrich,

 

I had bought AD products without any restriction until now.

 

Thanks for all your interesting remarks and comments.

 

Best regards,
Mauro.

 

 


From: paparazzi-devel-bounces+m.garcia=address@hidden [mailto:paparazzi-devel-bounces+m.garcia=address@hidden] On Behalf Of Heinrich Warmers
Sent: lunes, 18 de febrero de 2013 21:04


To: address@hidden
Subject: Re: [Paparazzi-devel] New IMU in YAPA

 

Hello Mauro,
 The Sensor
ADIS16488 cost 1,690.24 Dollar The Data a quite good. Comparable IMU`s with filter cost about 3000 Dollar and have GPS  inputs. Please look for the ADIS16445 400 Dollar.
But for normal flight with quadrrotors and fixed wing the 7-50 Euro solutions works perfect in Papparazzi..
The only field in witch precise sensors a necessary i think are wind turbulence measurements with angle errors in 0,1° region.

Can you by  the AD sensors  in Europe?
I  found that it becomes more and more difficult to by strategical  US sensors for military use free.
Therefore i want to by sensors from ST or Bosch in future.

I assume that  there is a misunderstanding how the solution in paparazzi work: Not the rate bias is a problem. The bias is calculated and corrected by the use of  the earth gravity measurement with the accelerometers.  So we have a static error < 1° and that is tolerable for autopilots.
Boch sells in future a accelerator sensor with 14 bit resolution. This may be brake down the error. 
When dynamically the error grows up to 2-5° for  times < 5s is is also not a problem for the autopilot.

Regads

Heinrich Warmers

Mauro Garcia Acero schrieb:

Hello Heinrich,

 

very interesting point you have here.

 

Have you made any experiments about vibrations with fuel engines? Typical gas engines may produce annoying vibrations even if silent blocks are installed in the engine block as long as I understood from people using this kind of aircrafts.

 

The flight duration of a quadcopter is at most one hour as long as I have seen by internet and typical flights are below 30 minutes. But if a quadcopter was going to still for more than 6 hours, also the same IMU would be sufficient for insuring the right attitude determination?

 

Why if only matters the quality of the algorithm or the filter, in civil aviation, military systems and space crafts use so expensive (hundreds of thousands dollars) inertial sensors if a fifty bucks IMU can be equally efficient?

 

Anyway, as Hector suggested, integrating the ADIS16488 IMU into paparazzi would be a positive evolution for integration of any other AD sensor because most of them use the same protocol. So even if at the end, it is proved that the integration of a high end IMU does not improve so much the overall behavior of the autopilot, a whole new brand of inertial sensors would have been added into paparazzi integration.

 

Coming back to my first question, there is an "easy start" for integration of this IMU other than creating a new module? I think that reusing some existing module may be more efficient in creating the source code.

 

Best regards,
Mauro.

 


From: address@hidden [mailto:address@hidden] On Behalf Of Heinrich Warmers
Sent: domingo, 17 de febrero de 2013 17:52
To: address@hidden
Subject: Re: [Paparazzi-devel] New IMU in YAPA

 

Hello,
what is the reason for hi-end IMU's on the yapa ?
When i measure the vibrations on my multikopters i get levels of  0.4 g .
The rate sensors are not so sensible for vibrations since modern have vibrations frequencies over 20kHz..
The vibrations have most first an second orders of the revolutions of the propellers when you generate a  spectrum.
A good pressure sensor with 24 bit resolution save the problems with the altitude measurement of the GPS (Wrong values up to 80m if roll angles > 70°. are flown). When i calibrate the ID500 after power on i reach a bias value of 0.3 to 0.5 °/s with the HB-mini.
So the cheap rate sensors build not a problem. The MPU6050 cost about 7 Euros 1-2°/s  and fly quadrcooters and normal wing excellent.
The problem is the filter.  The DCM filter was the first filter in paparazzi witch estimates the centripetal forces. With this filter you can fly helixes and circles as long you want. A improvement  for implement magnetometer values and offset angles would be fine.

To make a true INS you need a FPU and a closely coupled GPS and about 3-5  years of time to write a new dissertation of sensor fusion.
And the result ?  To have true values for 20 seconds instead of 4?

Please think first on the near problems 1. save fly with a GPS lost. 2. filters for the acceleration measure values. 3. true altitude measurement without GPS 3. true airspeed measurement 4. damping the vibrations etc.

Regards

Heinrich Warmers






Mauro Garcia Acero schrieb:


Hello everyone,

 

After some discussion with TU Delft people, I will try to integrate the new ADIS16488 from Analog Devices into the YAPA in order to verify the behavior of paparazzi with a high end MEMS IMU:

 

<From the AD site>

6°/hr in-run bias stability

0.3°/√hr angular random walk

0.01% nonlinearity

 

Because this IMU communicates through SPI port, do you have any suggestion about how could I go quicker in the development of the new module? I was thinking to reuse one existing IMU module already communicating through SPI. But I have just begun to analyze the doc of the web site, so, I'm open to any kind of possibility.

 

Thanks in advance,


 

 
 
 



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



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

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



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




--
Héctor


reply via email to

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