[Top][All Lists]

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

Re: [Paparazzi-devel] Talking about Camreras and Wind

From: Christophe De Wagter
Subject: Re: [Paparazzi-devel] Talking about Camreras and Wind
Date: Thu, 29 Jul 2010 23:27:52 +0200

Sorry about the confusion.

We always fly with the -DSTRONG_WIND, even with not strong winds...

If the wind is strong, you should make sure that the wind estimate is "set" manually prior to takeoff. Once in flight the wind is estimated by the groundstation and sent to the plane. But especially during a take-off, the wind is not known yet to the groundstation while the groundspeed is reduced due to the climbing, so the plane is very prone to get into trouble (flying backward). Since it takes a while for the groundstation to have a good estimate of the wind, it is (always) better to manually guess the wind and set it. (as long as the direction is good, the amplitude is of less big importance: set it to something in the range 0.5 x V-CUISE to VCRUISE: if there is more wind than VCRUISE you shouldnt fly, and less than half VCRUISE there is no need to set is.)

On Thu, Jul 29, 2010 at 1:28 PM, Stephen L Hulme <address@hidden> wrote:
Dear Christophe
Thank you for the informative reply.
But I think I am still missing something.   
When you say "I would say: if the wind comes close to your cruise speed.... you should set it prior to takeoff"
Do you mean set the flags such as:
ap.srcs += nav.c fw_h_ctl.c fw_v_ctl.c
or something else?. And if the wind is low remove the flags?
On Wed, 28 Jul 2010 22:16:27 +0200 Christophe De Wagter wrote:
Dear Stephen,

The wind.xml is redundant as the wind parameters are always present in the gcs - settings - flight parameters.

Procedure for strong winds:

I would say: if the wind comes close to your cruise speed.... you should set it prior to takeoff. In practice we never set it but always start with at least 1 standby circle. 

Without the "strong wind" procedure when the ground speed becomes zero or negative (e.g. when you climb, that reduces your groundspeed), the MAV keeps turning even when the nose is pointing in the right direction and hereby is blown away. The strong-wind procedure is no miracle but simply adds the estimated wind to the ground speed when estimating angles. Even if the wind estimation is not accurate but just points in more or less the correct quadrant, it makes a huge difference. When the groundspeed becomes close to zero, the strongwind function reduces the nav gain (as a small heading difference makes the plane move laterally which is seen as a (huge) 90 degree error, while the real heading error is only a few degrees). When the ground speed becomes negative, the strongwind function uses the estimated heading instead of the ground track. This enables you to fly backward. Using the formation flight routines in paparazzi you can make a "hovering controller" for planes.


On Wed, Jul 28, 2010 at 12:12 AM, Stephen L Hulme <address@hidden> wrote:
Hi Everyone
Re: Camera:
I have a similar software problem with a Canon IXUS 80IS and CHDK.
Following a very kind donation from John Stowers of a Booz main, IMU and power board I have successfully launched a 40cm MK quadrotor.
I wish to carry the mentioned cradled Canon camera and remotely trigger still shots using the USB remote configuration.while simultaneously transmitting  video over a 5.8 GHz transmitter.
To date I am having great difficulty getting the video to display beyond three minutes without the default script appearing on the video image.
Any help with a working script would be appreciated.
Pascal previously indicated in a PM that one of Antoine's quadrotors was running this setup.
Re :Wind
After crashing my Funjet last summer I came back to test my old the airframe.xml code for a rebuilt version and found the settings file wind.xml has been removed. 
Christophe De Wagter maybe you could kindly set out the proper procedure to run the strong wind function?  

Paparazzi-devel mailing list

reply via email to

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