DEV call 13 Jun 2016 4pm PST = 13 Jun 2016 2300 UTC

151 views
Skip to first unread message

Craig Elder

unread,
Jun 13, 2016, 7:02:09 PM6/13/16
to drones-discuss
Copter 3.4 Update 

Vision Update

Plane 3.6 Update

handling of safety switch and FMU outputs (and the associated patch from Michael duBreuil)

ADSB

Linux from Gustavo who cannot attend today

BMI160 is basically done... I'm just waiting for the hardware to arrive so I can test it. I've been doing some refactors related to the AP_HAL::Device interface. I'll test it soon and send to upstream as well.

Using Ardupilot on manned aircraft.

R&D Log analysis


Tom Pittenger

unread,
Jun 13, 2016, 7:03:30 PM6/13/16
to drones-discuss
Also: wind in SITL takes very long time to change when SIM_WIND_DIR is changed

--
You received this message because you are subscribed to the Google Groups "drones-discuss" group.
To unsubscribe from this group and stop receiving emails from it, send an email to drones-discus...@googlegroups.com.
For more options, visit https://groups.google.com/d/optout.

Craig Elder

unread,
Jun 13, 2016, 7:20:42 PM6/13/16
to drones-discuss
trad-heli test flight results with 3.4
found and fixed two heli bugs on the weekend

Craig Elder

unread,
Jun 14, 2016, 4:54:27 PM6/14/16
to drones-discuss

Copter 3.4 Update

testing 3.4rc1

http://discuss.ardupilot.org/c/arducopter/copter34


an issue found with over powered copters taking off with less than 21% throttle, it doesn’t think it’s taken off until it hits ~ 21%

Questions as to changing copter load either between flight, or during flight.

Issues on calibrating ESC’s
Issues on Compass calibration on Pixracer.
Safety switch issue on Pixracer.

Leonard has been doing lots of work on the quaternions, looking good :) improved yaw / pitch / Roll recovery and prioritisation.

Discussion as to limiting accels on trad heli to reduce the risk of cutting the tail boom off in a recovery…

Trad Heli
Tridge tested two helis on the weekend, both petrol, both flew well.
Two bugs found,

https://github.com/tridge/ardupilot/commit/94acb1079377e98aaf2f27db6bdacc5cb3ab2dd6
System was using an incorrect min / max (constraint proposed to be removed)


https://github.com/tridge/ardupilot/commit/ba66b52394b38677cffb2ccce7705ecf7f24c8ea

Calculation of Motor speed load…..

One heli didn't auto disarm… it started to roll on landing (possibly following GPS…..)
.

Vision Update


TX1 is looking good! Comms with pixhawk are good, visual follow me, precision landing.
Looking at SLAM, for internal mapping with a single camera.

LSD… Large scale direct Monocular Slam…. Is another method.

Looking for SLAM experts…

Needing to add the “Pose” into the EKF…

Discussion as to how this data should be brought into the EKF.

SLAM requires a large covariance matrix, so it runs on a companion computer (maybe with ROS)
Providing the EKF with a local position.
Some things to think about,...
Timestamping… to know about delay.
Need 5ms or better timestamping…. (Tridge discussing the difficulties of getting this right)
The  higher the Baud Rate, less buffer time is required. This may help on getting the time sync correct.

Re SLAM, this is the slam algorithm that Kabir recommends: https://github.com/uzh-rpg/rpg_svo

Plane 3.6 Update

Desired recovery feature…

When switching to “Auto” from inverted…. That the system can recover.
When at 90 degrees, we need to use some knife edge logic to allow the rudder and elevator to work correctly.
(back to quaternions)
Discussions as to what airframes can handle, and the fact that different airframes recover in different ways.
Discussion as to a “recovery” mode. Maybe a recovery technique enum. (pick the solution of your airframe.)

Quadplane


A bug at the end of transitions is smoother after getting a bug fixed on to outgoing transition.
On the incoming transition, there is a pitch up, (stall?) Tridge is investigating.

Other than that, the weathervaning is working well.
The use of the forward motor to assist in holding in hover…


S.Bus has been dropped to 150 Hz due to a weakness that does not allow securesync of frames at high framerates. (incorrect frame could end up with the wrong motor getting signal….)


https://github.com/ArduPilot/ardupilot/pull/4122

Discussion as to how to deal with a damaged Airspeed sensor.  https://github.com/ArduPilot/ardupilot/issues/1628

How to then add dual airspeed sensors.

There are stages of flight that this is ignored in, to avoid false positives.



handling of safety switch and FMU outputs (and the associated patch from Michael duBreuil)

Do we want to go back to Aux chan’s being active during safety switch mode.
Default will be to set all off. And user can set as needed.


ADSB

https://github.com/ArduPilot/ardupilot/issues/4309
When there are more than 25 aircraft, it clags up the link.
Peter and Tom will look into this further.


Ping 2020 model is a full transceiver  it is FCC and FAA approved. Discussions as to testing this.


Linux

from Gustavo who cannot attend today

BMI160 is basically done... I'm just waiting for the hardware to arrive so I can test it. I've been doing some refactors related to the AP_HAL::Device interface. I'll test it soon and send to upstream as well.

Using Ardupilot on manned aircraft.

We have had recent requests for support using ArduPilot in a manned aircraft.

We state on http://firmware.ardupilot.org/

Never use the software or hardware for manned vehicles. The software and hardware we provide is only for use in unmanned vehicles.

R&D Log analysis

Discussion as to the value of different tools that are available for finding issues….
Replay, High speed logging, and RMS.


For Vibe analysis:  Turn on full rate accel data, then run the FFT data.  Must use good SD card.

Reply all
Reply to author
Forward
0 new messages