Mirror/ardupilot
5
0
Fork 0
mirror of https://github.com/ArduPilot/ardupilot synced 2025-01-21 08:08:29 -04:00
Code Issues Packages Projects Releases Wiki Activity
3,562 Commits 71 Branches 511 Tags 761 MiB
41b435a487
Commit Graph

1281 Commits

Author SHA1 Message Date
Jason Short
b094f6ce52 added set integrator 2012-02-23 09:38:08 -08:00
Andrew Tridgell
ef90f336d3 DCM: changed the sense of floating point range comparison 
This allows us to detect NaN, otherwise NaN values were considered 
'in range'
 2012-02-23 11:41:26 +11:00
Andrew Tridgell
57ffd00b69 DCM: range check the matrix before calculating pitch 
The asin() in the pitch calculation can only take values between -1
and 1. This change ensures that the value is in range, and if it isn't
then we force a normalization. If that fails we reset the matrix
 2012-02-23 08:16:08 +11:00
Andrew Tridgell
2bb58db3f7 DCM: reset more values on renorm blowup 
when DCM blows up, we need to reset a lot more variables to ensure
that any NaN values don't persist
 2012-02-23 08:16:08 +11:00
Andrew Tridgell
156d98cb9f DCM: remove the taylor expansion optimisation for renormalisation 
The sqrt() costs about 44usec on a 2560, which is small enough for us
not to worry about the speed.

This also changes the range of values where we declare a blowup to
much less likely, which means we can cope with larger delta_t glitches
 2012-02-23 08:16:08 +11:00
Andrew Tridgell
53b87229ff AP_Param: added ParamToken type for variable list traversal 2012-02-19 17:05:28 +11:00
Jason Short
1e37a0ec17 added a limit to range output 2012-02-18 21:08:17 -08:00
Randy Mackay
33aa6fe13b Optflow - added reference to FastSerial to resolve compile error on Arduino 1.0. 
also removed reference to DCM (no longer required) and removed some old code.
 2012-02-18 18:09:40 +09:00
Andrew Tridgell
9793d04ed1 SITL: increase the amount of noise in the simulated ADC 
this increases the noise to 2 bits, which actually can have the effect
of improving accuracy, as it leads to better averaging
 2012-02-18 19:27:15 +11:00
Andrew Tridgell
d52cb7e574 DCM: only add in centripetal accel if we have GPS lock 
if we don't have a GPS or the GPS doesn't have a good lock then we
can't rely on the ground speed for adjusting the acceleration vector
 2012-02-18 19:27:15 +11:00
Andrew Tridgell
baac883eb1 IMU: move _sensor_cal to general IMU class 
this makes it available both in shim and INS subclasses
 2012-02-18 10:51:57 +11:00
Andrew Tridgell
882cd4ea83 AP_Param: moved AP_Vector3f and AP_Matrix3f declarations to AP_Math.h 
this avoids us needing AP_Math.h in every utility sketch and example
 2012-02-18 10:51:57 +11:00
Andrew Tridgell
4b15b63d21 AP_Param: fixed copy-assignment operators for AP_Param vectors 2012-02-18 10:51:57 +11:00
Andrew Tridgell
7f3858cb7f AP_Param: added a set_and_save_ifchanged() method 
this can be used to avoid the scan() in more frequenctly saved
variables, such as the MAVLink stream rates in APM
 2012-02-18 10:51:57 +11:00
Andrew Tridgell
80a6773b8b AP_Param: added an initialised() method 
this will be used by the compass code
 2012-02-18 10:51:57 +11:00
justinbeech
62b104cbfa Found free extra 256 bytes of RAM eliminate _clz use from libgcc 
Believe it or not, changing / 2^31 to >>31 saved 256 bytes in the "d" segment.

The reason is that GCC version prior to 4.3.5 does not have a count_leading_zeros (clz) assembler macro, so it uses a 256 byte lookup table called _clz
The _clz table gets pulled in if you do 64 bit division.
This tiny change is the only place that we do long long division.
Changing to a shift saves 256 bytes of ram.
 2012-02-17 09:46:05 +00:00
Jason Short
24ce02c6a4 made I term return in same pattern as D term 2012-02-16 22:09:22 -08:00
Andrew Tridgell
79d622939a ADC: on channel overflow we should not zero last_ch6_micros 
this happens every 64 seconds because of unused channels on the
ADC. Zeroing this creates a bad delta_t value for the DCM code.
 2012-02-16 17:48:36 +11:00
Jason Short
026e4a6567 change constant to float 44330.0 2012-02-15 09:10:14 -08:00
Michael Oborne
2f81776b0d eedump ap param in perl 2012-02-15 21:31:58 +08:00
Andrew Tridgell
a48a118128 SITL: adjust the SITL compass code to have saner z values 2012-02-15 22:34:16 +11:00
Andrew Tridgell
62326c0f72 Compass: enable compass offsets for HIL compass 
this makes the null_offsets algorithm have an effect for the HIL
compass that is used for SITL, which makes for easier testing of
compass calibration
 2012-02-15 22:33:50 +11:00
Andrew Tridgell
048aed5427 fixed dump size 2012-02-15 10:44:09 +11:00
Andrew Tridgell
089db43183 show group_element too 2012-02-15 10:42:21 +11:00
Andrew Tridgell
ea6a46fc9b AP_Param: added eeprom_aaparam.c dump utility 2012-02-15 10:41:06 +11:00
Andrew Tridgell
4398171374 SITL: catch floating point errors 
if we get a FPE then log a message and exit
 2012-02-15 08:50:45 +11:00
Andrew Tridgell
245d265c6a DCM: added a SITL_debug() for DCM normalisation errors 
I have seen normalisation errors during simulation runs, but have not
yet tracked down the cause. This log message may help track things
down.
 2012-02-15 08:50:45 +11:00
Andrew Tridgell
601a991a46 added SITL_debug() macro 
this can be used to print a message into the logs in the SITL
simulator. This will be used for critical conditions.
 2012-02-15 08:50:45 +11:00
Andrew Tridgell
085cc00c33 RC_Channel_aux: fixed a uninitialied variable error, and save a bit of stack space 2012-02-15 08:50:45 +11:00
Andrew Tridgell
43c56be70c SITL: add some magnetometer offsets 
this ensures the compass null_offset algorithm has some work to do in
autotest
 2012-02-14 11:26:42 +11:00
Andrew Tridgell
f95d7cadc5 fixed SITL GPS heading to be 0..360 2012-02-14 11:26:42 +11:00
Andrew Tridgell
6a42598ade use stdint types in the compass code 2012-02-14 08:35:03 +11:00
Andrew Tridgell
be5ffdedbb AP_Param: allow dumping of AP_Vector3f 
this means the 'show' command in the CLI shows the compass offsets now
 2012-02-14 08:34:50 +11:00
Andrew Tridgell
d90dc4d893 GPS: change back to 4Hz for MTK GPS 
this will hopefully fix the NO_GPS problems that have been regularly
happening in recent months. I will watch the logs carefully to see if
the problem recurs
 2012-02-13 22:50:20 +11:00
Andrew Tridgell
804332c340 AP_Baro: fixed some signed/unsigned warnings 
and a potential bug on timer wrap
 2012-02-13 21:39:18 +11:00
Andrew Tridgell
74bbc72b11 move AP_Var code and example into libraries/AP_Var 
you can use AP_Var by defining USE_AP_VAR
 2012-02-13 16:22:53 +11:00
Andrew Tridgell
23ed5c2cee fixed imax load/save in PID 2012-02-13 16:22:53 +11:00
Andrew Tridgell
37cabb4cfa fixed dead zone load/save 2012-02-13 16:22:53 +11:00
Andrew Tridgell
150a67c262 fixed PID initialisation 2012-02-13 16:22:53 +11:00
Andrew Tridgell
47807ec68d RC_Channel_aux: catch invalid function settings 
a bad function setting would overwrite the heap
 2012-02-13 16:22:53 +11:00
Andrew Tridgell
1f05e7dafb fixed types in GPS detect code 
using int everywhere is wasteful and non-portable
 2012-02-13 16:22:53 +11:00
Andrew Tridgell
d55a0c3a71 added AP_Param::show_all() 
this moves the logic into common code
 2012-02-13 16:22:52 +11:00
Andrew Tridgell
9175ebe33a change back to 256 max variables in _var_info[] 
this keeps the header to 24 bits, and allows the token to be 16
bits. If we need more variables we can put them in a nested group.
 2012-02-13 16:22:52 +11:00
Andrew Tridgell
9c5a5473ab added indexes to group info structures 2012-02-13 16:22:52 +11:00
Andrew Tridgell
f2a6a31645 AP_Param: added group idx, and change packing of header elements 
adding a idx element to the GroupInfo will make it less likely that a
developer will change the IDs of group elements, and make it easier to
see that these IDs are important for identifying a variable in EEPROM

The header packing has changed to make it 24 bits on all platforms,
which allows us to lower the EEPROM variable max size to 1024 bytes
again
 2012-02-13 16:22:52 +11:00
Andrew Tridgell
678a1df3be fixed support for AP_PARAM_SPARE 
spare group elements are needed if we remove a group parameter
 2012-02-13 16:22:52 +11:00
Andrew Tridgell
accaf9bb30 fixed build on arduino 1.0 2012-02-13 16:22:52 +11:00
Andrew Tridgell
db96303fa0 convert APM_PI library to AP_Param 2012-02-13 16:22:52 +11:00
Andrew Tridgell
452749149f convert AC_PID library to AP_Param 2012-02-13 16:22:52 +11:00
Andrew Tridgell
f9eed9bd9e fixed missing return in D value 2012-02-13 16:22:52 +11:00
Andrew Tridgell
f760b3c717 restore _fCut that was lost in AP_Param development 2012-02-13 16:22:51 +11:00
Andrew Tridgell
84aef17bb0 fixed initialisation of RC_Channel class 2012-02-13 16:22:51 +11:00
Andrew Tridgell
5418302b38 AP_Param: fixed desktop build 2012-02-13 16:22:51 +11:00
Andrew Tridgell
d658bc6f9d desktop: more progmem string functions 2012-02-13 16:22:51 +11:00
Andrew Tridgell
95e107ab75 AP_Param: added generic cast_to_float 2012-02-13 16:22:51 +11:00
Andrew Tridgell
9509c69f42 AP_Param: disable debug ready for production 2012-02-13 16:22:51 +11:00
Andrew Tridgell
5d03f13aa5 AP_Param: make ptype in first() and next() optional 2012-02-13 16:22:51 +11:00
Andrew Tridgell
e01af5885f adapt Mavlink_compat.h to AP_Param 2012-02-13 16:22:51 +11:00
Andrew Tridgell
bf4a77da81 convert libraries to new typesafe macros 2012-02-13 16:22:51 +11:00
Andrew Tridgell
e8c4ecf9fa added nested group support and validation 2012-02-13 16:22:51 +11:00
Andrew Tridgell
94a86204ee added size of Vector6f 2012-02-13 16:22:51 +11:00
Andrew Tridgell
12dac42174 modify RC_Channel for AP_Param 2012-02-13 16:22:51 +11:00
Andrew Tridgell
e5a0074312 modify the PID library for AP_Param 2012-02-13 16:22:51 +11:00
Andrew Tridgell
e5f3864f81 modify the IMU code for AP_Param 2012-02-13 16:22:51 +11:00
Andrew Tridgell
8c58d95187 modify the compass code for AP_Param 2012-02-13 16:22:51 +11:00
Andrew Tridgell
497a400841 hacks to get AP_Param to compile 2012-02-13 16:22:50 +11:00
Andrew Tridgell
5b842a447f new AP_Param subsystem 2012-02-13 16:22:50 +11:00
justinbeech
69f1613f2c Uncommented cold temp compensation after testing. Fixed read throttle so read rate ends up nailing 100hz, not 90hz. minor commenting improvements. 2012-02-13 00:00:06 +00:00
analoguedevices
0ea6ecf8d1 Removed outdated comment: 
"	// deadzone moved to
		//control_in = (abs(control_in) < _dead_zone) ? 0 : control_in;"
 2012-02-11 16:51:02 +00:00
Michael Oborne
8ec192cf45 fix uint > int issue - for hil 2012-02-11 11:35:59 +08:00
Pat Hickey
83326a9d62 APM_RC_APM2: Add support for CH_10 and CH_11 
* CH_10 is on APM2 pin A10 (based on A1..A8 convention)
* CH_11 is on APM2 pin A11

* Only the code in enable_out, disable_out, and OutputCh needed to change.

* CH_10 and CH_11 always have an output period of 20ms (50Hz).
 2012-02-06 22:11:01 -08:00
Randy Mackay
6877b8b3d6 AC_PID Library - added example sketch 2012-02-05 17:35:15 +09:00
Randy Mackay
90b39609c6 APM_PerfMon - removed because it's too big to use with our limited memory now anyway! I have a backup in case we need to resurrect it in the future. 2012-02-05 11:11:59 +09:00
Jason Short
af79eb273f Added D term to APM_PI - need to refactor this as a parent class and two child classes to save code space, remove dupes 2012-01-28 21:21:43 -08:00
Randy Mackay
9c26e2369e Arduino 1.0 - GPS_IMU - small fix to make it compile. Does anyone use this lib?! 2012-01-28 17:56:56 +09:00
Randy Mackay
8533aaf5d9 AP_PID, AP_RC_Channel, FastSerial - small changes to make example sketches compile again 2012-01-28 17:54:20 +09:00
Randy Mackay
b7ce7bfe18 Arduino 1.0 - small fix to make AP_Baro_MS5611_test.pde compile under Arduino 1.0 2012-01-28 13:21:35 +09:00
Randy Mackay
cf25bbfa74 AP_ADC library - tiny fix to make example sketch work after some changes to AP_PeriodicProcess library 2012-01-28 13:12:19 +09:00
Randy Mackay
ed19c25a97 Arduino 1.0 - changed all #includes of "WProgram.h", "wiring.h" and "WConstants.h to "Arduino.h". 
Modified FastSerial's write function to return size_t (number of bytes written).
 2012-01-28 12:25:47 +09:00
Randy Mackay
a8b3d2f063 OpticalFlow - fixed minor compiler warnings and moved chip select to A3 for APM2 2012-01-26 23:35:49 +09:00
Randy Mackay
8dee001657 Sonar - changed default pin to A0 2012-01-26 22:50:48 +09:00
Randy Mackay
99506accf4 Optical Flow - added overflow check 2012-01-25 21:45:52 +09:00
Pat Hickey
9195006541 APM_RC examples: add calls to enable_out after init 2012-01-24 09:09:08 -08:00
Andrew Tridgell
9c1e257826 fixed example in PID library 2012-01-23 08:47:46 +11:00
Pat Hickey
602d56304a APM_RC: PWM outputs are disabled by default on init 2012-01-22 12:03:11 -08:00
Pat Hickey
276913a950 APM_RC enable_out and disable_out added to APM1 and APM2 
* Outputs still enabled by default
 2012-01-22 12:03:11 -08:00
Jason Short
94e7ba8df2 Removed Gyro filter, In the end it doesn't help and I think it could actually hurt gyro integration. 2012-01-20 22:52:29 -08:00
Michael Oborne
6b949511ca change rc norm_output to scale evenly across the entire range 2012-01-20 16:24:55 +08:00
Andrew Tridgell
0dc0d9ddf0 added a simple DCM library example 2012-01-18 21:25:19 +11:00
Andrew Tridgell
1f43cf6579 fixed Oilpan test for new library structure 2012-01-18 20:31:58 +11:00
Doug Weibel
e22a83eec0 Add a CLI setup action to reset the magnetometer offsets 
Since the magnetometer offsets are not available through the MAVLink parameter interface (since they are an AP_Var vector) this little feature allows them to be reset from the CLI.  Useful if you somehow get bad offsets or if you change magnetometer.  If you have a bad set of large offset values I have seen issues with the nulling algorithm have trouble converging to the proper values.  I have never seen it have trouble converging from 0/0/0, so this could be a useful feature from time to time.
 2012-01-15 16:15:08 -07:00
Mike Smith
9c5a776456 Teach the Makefile about the new Arduino sourcefile suffix 2012-01-15 11:51:38 -08:00
Jason Short
69ffc7e115 updated channel notes 2012-01-15 10:24:15 -08:00
Andrew Tridgell
897d9a1c0b baro: fixed an integer overflow issue at high altitudes 
the averaging array was using 16 bit numbers, but we are storing
numbers with 19 significant bits. That caused overflow at high
altitude, and some very interesting altitude graphs!

Thanks to Michael Oborne for spotting this in a log
 2012-01-14 19:51:35 +11:00
Randy Mackay
bcd63a9807 OpticalFlow - bug fix - highest resolution was incorrectly set as 1200 instead of 1600 2012-01-14 12:34:29 +09:00
Andrew Tridgell
15d446bde2 DCM: prevent a segmentation fault when compass is not enabled 
when compass is disabled _compass is NULL
 2012-01-13 15:48:07 +11:00
Doug Weibel
9846822748 Bug fix for compass. 
This is a fix for an interesting bug when a DCM matrix reset was added to the ground start.  This bug only showed up if (A) a ground start were performed after an air start or due to use of the "Calibrate Gryo" action, (B) if the current orientation were sufficiently different from 0/0/0, and (C.) if the particular magnetometer had sufficiently large offsets.  Why did resetting the DCM matrix to 0/0/0 pitch/roll/yaw at ground start cause a bug?  The magnetometer offset nulling determines the proper offsets for the magnetometer by comparing the observed change in the magnetic field vector with the expected change due to rotation as calculated from the rotation in the DCM matrix.  This comparison is made at 10Hz, and then filtered with a weight based on the amount of rotation to estimate the offsets. Normally it would take considerable time at normal in-flight rotation rates for the offset estimate to converge. 

If a DCM matrix reset occurs when the offset nulling algorithm is up and running, the algorithm sees the DCM reset as a instantaneous rotation, however the magnetic field vector did not change at all.  Under certain conditions the algorithm would interpret this as indicating that the offset(s) should be very large.  Since the "rotation" could also have been large the filter weighting would be large and it was possible for a large erroneous estimate of the offset(s) to be made based on this single (bad) data point.

To fix this bug methods were added to the compass object to start and stop the offset nulling algorithm.  Further, when the algorithm is started, it is set up to get fresh samples.  The DCM matrix reset method now calls these new methods to stop the offset nulling before resetting the matrix, and resume after the matrix has been reset.
 2012-01-12 14:44:24 -07:00
Andrew Tridgell
49cf409c23 desktop: floating point precision changes from Justin Beech 2012-01-11 21:31:59 +11:00
Jason Short
e2830ca147 fixed type error 2012-01-10 23:43:44 -08:00
Jason Short
354d9588bf sensorSign swicth 2012-01-10 23:43:44 -08:00