Cyfoam, ctrl effectiveness scheduling, JSBsim #2145
Conversation
EwoudSmeur
commented
Oct 30, 2017
EwoudSmeur
commented
- Added Cyfoam airframe file
- Added control effectiveness scheduling function
- More advanced scheduling function should be added in the future
- Added simple JSBsim files for Cyclone for a crude simulation
| g1g2[i][j] = 0.0; | ||
| } else { | ||
| g1g2[i][j] = (g1g2_hover[i][j] * (1.0 - ratio) + g1g2_forward[i][j] * ratio); | ||
| } |
There was a problem hiding this comment.
why the if here? if g1g2_hover and _forward are 0. the result will be 0 anyway...
| */ | ||
| extern void ctrl_eff_scheduling_periodic(void); | ||
| extern void ctrl_eff_scheduling_periodic_a(void); | ||
| extern void ctrl_eff_scheduling_periodic_b(void); |
There was a problem hiding this comment.
are these defined anywhere, don't see it in this PR.
| } | ||
| } | ||
|
|
||
| #if EFF_SCHED_USE_FUNCTION |
There was a problem hiding this comment.
why is this in an ifdef? Either the module is included and running because it is included in the airframe (in which case this must be defined else pprz will not compile because the function is not defined) or its not included in the airframe and this entire file is not compiled.
conf/modules/tlsf.xml
Outdated
| @@ -17,6 +17,6 @@ | |||
| <define name="MAX_FLI" value="$(TLSF_MAX_FLI)"/> | |||
| <file_arch name="tlsf_malloc_arch.c"/> | |||
| <include name="$(PAPARAZZI_SRC)/sw/ext/tlsf"/> | |||
| <file name="tlsf.c" dir="$(PAPARAZZI_SRC)/sw/ext/tlsf"/> | |||
| <file name="tlsf.c" dir="../../tlsf"/> | |||
There was a problem hiding this comment.
This doesn't look quite right. As far as I understand it this does "work" but maybe keep the old line commented here for future people (just in case this causes issues in the future).
There was a problem hiding this comment.
this was working for sure, can it be related to the vpath tool (again) ?
There was a problem hiding this comment.
it works, it just looks really weird, I would have guessed it would be something like ../ext/tlsf but that didn't work. I am fine with it like this though. Indeed it needs to be changed due to the vpath tool.
There was a problem hiding this comment.
I have tried again with the original line, and it does continue compiling, but it gives the following error in the beginning:
../../sw/tools/find_vpaths.py arm-none-eabi-gcc /home/ewoud/paparazzi_check/var/aircrafts/CYFOAM/ap_tmp.list /home/ewoud/paparazzi_check
arm-none-eabi-gcc: error: ERROR:: No such file or directory
arm-none-eabi-gcc: error: found: No such file or directory
arm-none-eabi-gcc: error: more: No such file or directory
arm-none-eabi-gcc: error: than: No such file or directory
arm-none-eabi-gcc: error: once:: No such file or directory
arm-none-eabi-gcc: error: /home/ewoud/paparazzi_check/sw/ext/tlsf/tlsf.c/home/ewoud/paparazzi_check/sw/airborne/arch/chibios/mcu_arch.c: Not a directory
This if statement triggers 34 times for the first srcs[i]:
(line 58) if os.path.isfile(path.join(vpath, f)) :
I don't really understand what the vpath file does...
There was a problem hiding this comment.
Ah this looks like #2182 !
Running it from the command line helped to look at the prints and errors:
python3 sw/tools/find_vpaths.py arm-none-eabi-gcc /home/ewoud/paparazzi_check/var/aircrafts/CYFOAM/ap_tmp.list
There was a problem hiding this comment.
find_vpaths again. Does it get better with the latest fix in master? (just rebase master on top this PR)
There was a problem hiding this comment.
It does get better! Thanks @podhrmic !
| <module name="ins" type="float_invariant"> | ||
| <define name="INS_PROPAGATE_FREQUENCY" value="500"/> | ||
| <define name="INS_FINV_MAG_ID" value="MAG_HMC58XX_SENDER_ID"/> | ||
| </module> |
There was a problem hiding this comment.
This airframe looks reversed...
There was a problem hiding this comment.
what is the issue here exactly ?
There was a problem hiding this comment.
The airframe order is reversed to the conventional order (firmware, control, sections, modes).
EwoudSmeur
force-pushed
the
cyfoam_master
branch
from
b4cba33
to
36081be
Compare
conf/airframes/ENAC/cyfoam.xml
Outdated
| </firmware> | ||
|
|
||
| <modules main_freq="512"> | ||
| <!--<module name="gps" type="ubx_ucenter"/>--> |
There was a problem hiding this comment.
We are phasing out separate modules sections, can you move this to the firmware above (like you already did with your other modules?) additionally, I think the order the modules appear is important for how they are run so consider that, usually, the "subsystems" go first then the more mission specific modules.
| STABILIZATION_INDI_G1_PITCH, STABILIZATION_INDI_G1_YAW, STABILIZATION_INDI_G1_THRUST | ||
| }; | ||
|
|
||
| float g2_both[INDI_NUM_ACT] = STABILIZATION_INDI_G2; //scaled by INDI_G_SCALING |
conf/airframes/ENAC/cyfoam.xml
Outdated
| <define name="BODY_TO_IMU_PSI" value="180." unit="deg"/> | ||
| </section> | ||
|
|
||
| <section name="AUTOPILOT"> |
There was a problem hiding this comment.
This is usually at the bottom of the airframe. I think we should discuss what the preferred order is for the future... I think it is beneficial to have a standard but that standard hasn't been clearly defined. @gautierhattenberger is this defined anywhere?
There was a problem hiding this comment.
There is no definition of the order of the main nodes in airframe files at the moment. Even the DTD file is not reflecting anything in particular. Also the parser don't really care. We could edit somewhere a "recommended practice" guide. My preferred order is firmware, then "special nodes" (servos, commands, ...) and finally 'sections'.
- Added Cyfoam airframe file - Added control effectiveness scheduling function - More advanced scheduling function should be added in the future - Added simple JSBsim files for Cyclone for a crude simulation
EwoudSmeur
force-pushed
the
cyfoam_master
branch
from
df170ba
to
97ef94f
Compare
|
@EwoudSmeur @kirkscheper it looks fine now |
* Cyfoam, ctrl effectiveness scheduling, JSBsim - Added Cyfoam airframe file - Added control effectiveness scheduling function - More advanced scheduling function should be added in the future - Added simple JSBsim files for Cyclone for a crude simulation * only calculate pseudo inverse when using it
