Re: [Paparazzi-devel] Change in altitude handling?

[Ben Laurie]

On 27 March 2014 16:10, Gautier Hattenberger
<address@hidden> wrote:
> Hi,
>
> I only flew the ardrone two times in autonomous mode (last week), so I can't
> really pretend that I recommend my settings, but if you want to compare with
> the default setttings, I'm attaching my airframe file. The link below shows
> the result of the demo flight I made with it.
>
> https://www.youtube.com/watch?v=X2SUFWoKBKw

It looks to me like the sonar goes negative when it gets no reading -
wouldn't it make sense to have the sonar code ignore values below
zero?

Also, you mention the baro is really noisy - which curve is that? The
only one I could find (INS_REF:baro_qfe) is completely flat.

>
> Gautier
>
>
> 2014-03-27 6:24 GMT+01:00 Stephen L Hulme <address@hidden>:
>
>> Hi Gautier
>>
>> You may have noticed I posted a similar Ardrone2 issue:
>> http://lists.gnu.org/archive/html/paparazzi-devel/2014-03/msg00209.html
>>
>> I was using the latest master repo Sonar defines:
>>
>> <section name="INS" prefix="INS_">
>>   <define name="SONAR_SENS" value="0.00034"/>
>>   <define name="SONAR_MAX_RANGE" value="2.2"/>
>>   <define name="SONAR_UPDATE_ON_AGL" value="TRUE"/>
>>   <define name="SONAR_VARIANCE_THRESHOLD" VALUE="2.0"/>
>> </section>
>>
>> Can you post a recommended section setup for the SONAR sensor?
>>
>> Also it would be helpful to know about the AGL related define if one flies
>> in an area where the SRTM elevation data reflects a large variance with GPS
>> derived HMSL and hence is not functional and not used.
>>
>> Unfortunately I am currently away from home and cannot recheck my log
>> data.
>>
>> Regards
>>
>> Stephen
>>
>>
>>
>> On March 27, 2014 12:13 PM,Gautier Hattenberger wrote:
>>
>> Hi,
>>
>> I think I found the problem. It is not related to the waypoints or the
>> command, but rather from the estimation and the sonar sensor.
>> From the few flights I have done with the ARDrone, it seems that when
>> the sonar is out of range or can't get a "correct" range, it returns 0.
>> You can see it clearly in the log (INS_SONAR message). As you have only
>> set the SONAR_MAX_RANGE, the estimator only discard to high values, but
>> it takes all the 0 even when in flight. The result is that the estimator
>> always see you close to the ground (below the tracked altitude) and the
>> aircraft keep trying to climb.
>> Here are my suggestions:
>> - set INS_SONAR_MIN_RANGE to a small value above 0 (I have set it to 0.01)
>> - the threshold on sonar variance is probably useless since the baro is
>> really noisy (it is actually looking at the variance of the difference
>> between sonar and baro to see if the "change" in the direction)
>> - you can remove INS_SONAR_SENS, it is set outside of INS now (in
>> boards/ardrone/navdata.c)
>>
>> I hope it will help you to solve your problem.
>>
>> Gautier
>>
>> Le 26/03/2014 22:53, Ben Laurie a écrit :
>>>
>>> Data file is rather large, due to me taking quite a while to find the
>>> copter after the crash... I'll send it offlist.
>>>
>>> On 26 March 2014 20:59, Gautier Hattenberger
>>> <address@hidden> wrote:
>>>>
>>>> yes, .data file is needed to replay the flight, maybe we'll see if the
>>>> waypoints have been incorrectly moved when doing geo init.
>>>>
>>>> Gautier
>>>>
>>>> Le 26/03/2014 21:21, Ben Laurie a écrit :
>>>>
>>>> On 26 March 2014 14:50, Gautier Hattenberger
>>>> <address@hidden> wrote:
>>>>
>>>> Hi,
>>>>
>>>> There are some recent changes in ins_reset functions, but I made a
>>>> successful flight last week with an ARdrone.
>>>> What is your flight plan ?
>>>>
>>>> Flight plan is the standard rotorcraft_basic plan.
>>>>
>>>> current and last working version ?
>>>>
>>>> Current version is ad7994c7efc81b725823efff44e9ee9b83fe76bc, previous
>>>> working version was last year some time before xmas (can probably work
>>>> out exact version if needed). Been rather busy this year, was my first
>>>> chance to fly!
>>>>
>>>> maybe your
>>>> flight log would help.
>>>>
>>>> log attached - do you need the .data file, too?
>>>>
>>>> Gautier
>>>>
>>>> Le 26/03/2014 15:17, Ben Laurie a écrit :
>>>>
>>>> Nearly lost my ardrone today ... latest master (plus my local stuff,
>>>> which is mostly just calibration) from 3 days ago. Took off, was fine.
>>>> Changed to "stay_p1" and it immediately started climbing, fast.
>>>> Switched to "land" before it went completely out of sight, which
>>>> luckily it did kinda do, though it ended up more-or-less out of
>>>> control and quite some way away where it hit trees. Possibly because
>>>> it thought it was landed well above ground level? Dunno.
>>>>
>>>> Anyway, it looks like the waypoints are set to 200m above ground level
>>>> (207.78, to be precise) ... which is not what they used to do when GPS
>>>> lock is achieved and they all relocate.
>>>>
>>>> Did something change?
>>>>
>>>> BTW, I should have video, might be quite amusing :-)
>>>>
>>>> _______________________________________________
>>>> 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
>>>>
>>> _______________________________________________
>>> 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
>