Paul Riseborough
470098e182
EKF: update default tuning parameters
...
Parameters needed to be changed when covariance prediction error was corrected.
The magnetic earth and body field process noise now use separate parameters to help with tuning
2016-04-05 22:23:37 +02:00
Paul Riseborough
eaf94935f0
EKF: Fix bug in initialisation of height and magnetic field
...
This prevents zero data being used to form the initial height and magnetic field.
Do not start sampling initial values until non-zero time values are retrieved from the buffer.
2016-04-05 22:23:37 +02:00
Paul Riseborough
03eac2f25e
EKF: Apply covariance prediction derivation changes
2016-04-05 22:23:37 +02:00
Paul Riseborough
ac71ec5d37
EKF: Fix scaling error in covariance prediction derivation
...
The covariance prediction derivation was using the noise 1-std value instead of the noise variance. This resulted in the values of IMU process noise required to tune not matching the observed covariance growth.
2016-04-05 22:23:37 +02:00
Roman
65d491cafc
ekf2: update airspeed fusion (not yet active)
2016-04-05 21:37:36 +02:00
Roman
42a5be5afe
replay cleanup
2016-04-05 21:24:11 +02:00
Roman
f3e147f57b
make replay faster
2016-04-05 21:24:11 +02:00
Roman
1ce99e3b96
ekf2 replay: some cosmetics
2016-04-05 21:24:11 +02:00
Roman
bd4a0e30de
ekf2 replay: allow user to change parameters
2016-04-05 21:24:11 +02:00
CarlOlsson
8678a939e2
removed reinitialization of intermediate variables
2016-04-05 14:18:47 +02:00
Julian Oes
7352dc6f2e
commander: workaround to wait 1s before param save
...
Without this wait, we end up saving the old parameters on Snapdragon.
2016-04-05 13:02:53 +02:00
Julian Oes
60d66dc23f
param_shmem: before saving update the param
2016-04-05 13:02:53 +02:00
Julian Oes
57c134a23d
DriverFramework: updated submodule
...
This fixes the clock used for pthread_cond_timedwait properly.
2016-04-05 09:30:33 +02:00
Julian Oes
1afb7af2bb
cmake: don't mess with git submodules everywhere
...
It's enough to have the check_submodules.sh script. The rest needs to
go.
2016-04-05 09:22:37 +02:00
Roman Bapst
2632c930f7
Merge pull request #85 from CarlOlsson/add_error_reporting_tas
...
added fault status reporting
2016-04-04 18:52:06 +02:00
Roman Bapst
96b9674cb6
Merge pull request #84 from CarlOlsson/change_to_true_airspeed
...
change name to true_airspeed
2016-04-04 18:49:04 +02:00
CarlOlsson
eee6f1048f
Changed vtaspred threshold
2016-04-04 16:27:06 +02:00
CarlOlsson
3ad5b52aea
removed comment
2016-04-04 16:26:46 +02:00
CarlOlsson
616725410d
updated calculation
2016-04-04 16:26:13 +02:00
CarlOlsson
ec9902d802
added fault status reporting
2016-04-04 16:21:12 +02:00
CarlOlsson
1ea26b406a
change name to true_airspeed
2016-04-04 16:10:52 +02:00
Julian Oes
e5c64c7629
DriverFramework: update submodule
...
This fixes a bug that the mainapp on Snapdragon uses 100% CPU.
2016-04-04 16:01:44 +02:00
Julian Oes
2b95b3dff1
sdlog2: fixed wrong memset
2016-04-04 13:02:36 +02:00
tumbili
94aaf0d298
increase sleep time in accel calibration routine to make accel calibration work on snapdragon
2016-04-04 11:17:26 +02:00
suiauthon
e006189ce5
changed counting semaphore to binary
2016-04-04 10:06:59 +02:00
Mark Whitehorn
2f057939ad
fix bug in handling of missing default current scaling parameter
2016-04-02 21:40:34 +01:00
Lorenz Meier
0d94800b50
VTOL: Add missing fields to log. Fixes #4080
2016-04-02 21:39:03 +01:00
sander
cbe7bd2198
Set virtual waypoint for VTOL_TAKEOFF item
2016-04-02 21:29:27 +01:00
sander
4fc24dce60
Set correct setpoint type for vtol transitions
2016-04-02 21:29:27 +01:00
sander
dcacefb0d3
Ignore position for VTOL_TAKEOFF
2016-04-02 21:29:26 +01:00
sander
a9511478f2
Don't allow yaw updates during transition
2016-04-02 21:29:26 +01:00
sander
f9037b9ca7
Force vtol by default and correct param name
2016-04-02 21:29:26 +01:00
sander
4171e89137
Code style and commenting
2016-04-02 21:29:26 +01:00
sander
b922ceecfd
Rename VT_FORCE_VTOL to VT_NAV_FORCE_VTL
2016-04-02 21:29:26 +01:00
sander
26c8df0eb9
Sanitize vtol_takeoff behavior
2016-04-02 21:29:26 +01:00
sander
1a20de7908
Fix aligning to next waypoint for regular transition command
2016-04-02 21:29:26 +01:00
sander
3f256a778d
apply vtol_force_vtol
...
remove debug output
2016-04-02 21:29:26 +01:00
sander
b99e3e0ae2
Wait for VTOL transition command to complete
2016-04-02 21:29:26 +01:00
sander
240c8fbee1
Fix VTOL_LAND to have correct SETPOINT_TYPE
2016-04-02 21:29:26 +01:00
Andreas Antener
aafeab63d5
do not try to calculate FOH altitude when distance and acceptance radius are 0
2016-04-02 21:29:26 +01:00
sander
31862be420
Workitem handling for land
2016-04-02 21:29:26 +01:00
sander
903a0cd689
Vtol takeoff/land handling in work items
2016-04-02 21:29:25 +01:00
sander
8d8c3f9683
Implement forced VTOL landing
2016-04-02 21:29:25 +01:00
sander
a713fd4197
Implemented VTOL_TAKEOFF and VTOL_LAND commands
2016-04-02 21:29:25 +01:00
Lorenz Meier
e832f7dbf5
Fix reporting on broadcast address
2016-04-02 00:04:26 +02:00
Julian Oes
f4f1f5f027
commander: raise stack of low prio thread
...
The accel calibration sometimes lead to a crash. Raising the stack size
resolved the issue.
2016-04-01 15:12:08 +02:00
tumbili
08be5b3f3a
fixed formatting
2016-04-01 13:54:47 +02:00
tumbili
4da93a316d
ekf2 replay logic:
...
support recent change in ekf2 module which made it only publish the output
when an ekf update was occuring. in case the ekf2 module does not update
the replay module will be informed and will thus not wait for an update
but continue preparing the next sample of sensor data
2016-04-01 13:54:46 +02:00
suiauthon
b3818860de
changes in includes for baremetal
2016-04-01 13:50:29 +02:00
Julian Oes
994947ea84
Makefile/cmake/Tools: add easy Snapdragon upload
...
This adds a a target `make eagle_default` to build both the POSIX and
the QURT side in one command. Also, it adds an upload target for both to
push the files over adb to the device. This doesn't just push the
executables and lib files, but also the startup config files.
2016-04-01 11:36:27 +02:00
Paul Riseborough