Originally Posted by actionplus

Here are more videos.

Thank you for sharing! I still have to figure out how to make and publish a video here..
If covering the camera remedies the 'gass' problem, the algorithm of the vision software for the down looking camera is falling apart over grass... However, when it is working, it does a good job keeping the drone hovering in one place. Since I can not look at the source, my guess would be that they use some sort of correlation between successive image frames and just loose it over grass using a lot of CPU time, which in turn upsets everything else...
The wild swinging also happens here once in a while with the camera uncovered over concrete.
The midpoint problem might be due to the low resolution of the receiver signal sampling (+-100 points). I changed that in rev 0.05 to +-800 points.
Noticed you are flying with the indoor shell, did you change the 'cfgs[]' to reflect that?

No. I did not change the config to indoor shell. Do I just put 'false' in that line code?

I was thinking about using a little DPST switch on the arduino, where one is a pull up and the other a pull down, so that way we can switch between modes in a fly. But that will require some code changes from the master.

But I did try flying without any shell and it felt MUCH better and smoother, but didn't want to fly it too long till the drone figured out that the cam is still looking at one spot. (the tape).

On the helicommand, we place a piece of paper with a marking on it, so it thinks it is still looking at one spot and so no corrective action is taken by it. I will try placing a marker on the drones vision today and see if that helps.


As far video.
I took it on my iPhone. Download it to my Mac. Then uploaded it to YouTube.
Then pasted the links from YouTube on the thread v

Originally Posted by miru

Thank you for sharing! I still have to figure out how to make and publish a video here..
If covering the camera remedies the 'gass' problem, the algorithm of the vision software for the down looking camera is falling apart over grass... However, when it is working, it does a good job keeping the drone hovering in one place. Since I can not look at the source, my guess would be that they use some sort of correlation between successive image frames and just loose it over grass using a lot of CPU time, which in turn upsets everything else...
The wild swinging also happens here once in a while with the camera uncovered over concrete.
The midpoint problem might be due to the low resolution of the receiver signal sampling (+-100 points). I changed that in rev 0.05 to +-800 points.
Noticed you are flying with the indoor shell, did you change the 'cfgs[]' to reflect that?

I am using ver. 004 with out issue...over grass, concrete.etc...but I do have a question...how do the settings, indoor hull, outdoor hull etc affect the drone?

Originally Posted by rjrobbin

I am using ver. 004 with out issue...over grass, concrete.etc...but I do have a question...how do the settings, indoor hull, outdoor hull etc affect the drone?

What drone firmware are you running? 1.5.1?

Maybe I should reload it again as I updated it via my iPhone. Might be better doing it via a Mac/pc

Originally Posted by actionplus

What drone firmware are you running? 1.5.1?

Maybe I should reload it again as I updated it via my iPhone. Might be better doing it via a Mac/pc

well...( long answer ) I am using the latest..1.5.1...I read on one of my iPhone apps that I was upgrading the suggestion to re-download the firmware....author claimed that Parrot does not do point releases...once again...that is what I read..So I down loaded the latest code and ftp'd it to my drone....( short answer ) 1.5.1.

Originally Posted by actionplus

No. I did not change the config to indoor shell. Do I just put 'false' in that line code?

Yes.

Originally Posted by rjrobbin

I am using ver. 004 with out issue...over grass, concrete.etc...but I do have a question...how do the settings, indoor hull, outdoor hull etc affect the drone?

I think it affects the gains in the control loop, it is more sloppy if it is told that it is flying with the indoor hull.

Originally Posted by miru

I think it affects the gains in the control loop, it is more sloppy if it is told that it is flying with the indoor hull.

Interesting..thanks...this has been bugging me for a while....

Originally Posted by miru

I think it affects the gains in the control loop, it is more sloppy if it is told that it is flying with the indoor hull.

Yes, it does impact the PID gains.
http://code.google.com/p/brown-ros-p...common.h?r=964

/**
 *  \var     CTRL_DEFAULT_NUM_PQ_KP
 *  \brief   Numerator of default proportionnal gain for pitch (p) and roll (q) angular rate control loops
 */
/**
 *  \var     CTRL_DEFAULT_NUM_EA_KP
 *  \brief   Numerator of default proportionnal gain for Euler Angle control loops
 */
/**
 *  \var     CTRL_DEFAULT_NUM_EA_KI
 *  \brief   Numerator of default integral gain for Euler Angle control loops
 */

#define CTRL_DEFAULT_NUM_PQ_KP_NO_SHELL /*30000  26000 */ 20000
#define CTRL_DEFAULT_NUM_EA_KP_NO_SHELL /*7000  9000  7000 */ 8000
#define CTRL_DEFAULT_NUM_EA_KI_NO_SHELL /*4000  7000 6000 */ 7000

#define CTRL_DEFAULT_NUM_PQ_KP_SHELL /*30000 23000*/ 20000
#define CTRL_DEFAULT_NUM_EA_KP_SHELL /*9000 10000*/  9000
#define CTRL_DEFAULT_NUM_EA_KI_SHELL /*5000 9000*/   8000

It can be changed by AT*GAIN

I did my first update on the drone via my Mac yesterday. Usually via iphone and it took a few minutes to update.

Now, using ftp via Filezilla. It only took less than a minutes to transfer the file over. Then I disconnected the battery from the drone and reconnected. LED was red as usual, then back to green.

Instruction on the website said it will take about 10minutes
But mine was like normal bootup time. Did it really update it?

Thanks

Originally Posted by actionplus

I did my first update on the drone via my Mac yesterday. Usually via iphone and it took a few minutes to update.

Now, using ftp via Filezilla. It only took less than a minutes to transfer the file over. Then I disconnected the battery from the drone and reconnected. LED was red as usual, then back to green.

Instruction on the website said it will take about 10minutes
But mine was like normal bootup time. Did it really update it?

Thanks

Check with your iPhone and FreeFlight, it will tell you what version you are running.

Well. The problem is that I had 1.5.1 already in it. But I did
It via iPhone. But to make sure the install via iPhone was not causing the grass issue. I want to make sure my reinstall via Mac did actually install
And it is not the install from the iPhone. Guess there is no way of telling.

Originally Posted by actionplus

I did my first update on the drone via my Mac yesterday. Usually via iphone and it took a few minutes to update.

Now, using ftp via Filezilla. It only took less than a minutes to transfer the file over. Then I disconnected the battery from the drone and reconnected. LED was red as usual, then back to green.

Instruction on the website said it will take about 10minutes
But mine was like normal bootup time. Did it really update it?

Thanks

I did my update via the PC. It didn't take anywhere close to 10mins..more like a minute.

Originally Posted by actionplus

Well. The problem is that I had 1.5.1 already in it. But I did
It via iPhone. But to make sure the install via iPhone was not causing the grass issue. I want to make sure my reinstall via Mac did actually install
And it is not the install from the iPhone. Guess there is no way of telling.

If your iPhone has the latest FreeFlight, v1.7.1 at this time, AND you can 'telnet' into the drone, you can force an update of the firmware like this:
1) power up drone
2) once it shows the green lights telnet into it and do the following commands:
# cd firmware
# echo 0.0.0 > version.txt
# sync
# exit this should terminate your telnet session
3) cycle power to the drone, it should boot and turn on the green lights
4) get your iPhone connected to the ardroneXXX network and start FreeFlight, it will tell you that the drone is running revision 0.0.0 and there is an upgrade available, select OK and follow the procedure... when it wants you to restart the drone, cycle power to the drone and it will do the install. Do NOT quit FreeFlight during the power cycle, it expects, acually waits for, the loss of communication and reconnects once the drone is back up.
At least you know now that you are running the latest stuff on the drone.

Think my grass issue is gone after the reinstall of Drone firmware. I only got a few minutes of flight. (Windy again out here)
But looks to be very promising.