#robotics Logs

May 05 2020

04:30 AM goppo: anyone know if stereo cameras just output two images or do they compute the depth map on its hardware as well?
04:59 AM goppo: right
05:05 AM mun: speaking of stereo cameras, how are the IR texture projectors create the dots? is it some sort of filter over an IR source?
05:07 AM veverak: Jak_o_Shadows: actually, intel realsense has ASIC for it
05:07 AM veverak: (of course it was fpga at some part in development)
05:08 AM Jak_o_Shadows: That's pretty cool, that's a pretty dedicated part
05:08 AM veverak: yes
05:09 AM veverak: and intel has factories/experience to do it :D
08:22 AM rue_bed: veverak, did you get it to compile?
08:25 AM veverak: nope
08:25 AM veverak: busy all day
08:25 AM veverak: still missing something
08:36 AM rue_mohr: libsdl2-dev, and libsdl2-gfx-dev
08:36 AM rue_mohr: is all its using iirc
08:36 AM veverak: I installed all arch has on this topic
08:36 AM veverak: so there is some bug or something
08:53 AM rue_mohr: whats the error your getting now?
09:42 AM veverak: still the same
09:42 AM veverak: hmm
09:42 AM veverak: /usr/include/SDL/SDL_gfxPrimitives.h:38:10: fatal error: SDL.h: No such file or directory
09:42 AM veverak: 38 | #include <SDL.h>
09:43 AM veverak: but /usr/include/SDL/SDL.h exists
09:52 AM rue_mohr: oooh
09:52 AM rue_mohr: interesting
09:52 AM rue_mohr: so close and yet so far
09:52 AM veverak: yeah
09:52 AM rue_mohr: off to work, bye!
09:52 AM veverak: bb
01:54 PM Maya-sama is now known as Miyu
03:16 PM Miyu is now known as hackkitten
05:03 PM rue_mohr: veverak,
05:03 PM rue_mohr: which sdl do you have installed then
05:04 PM rue_mohr: libsdl1.2-dev: /usr/include/SDL/SDL.h
05:04 PM rue_mohr: libsdl2-dev: /usr/include/SDL2/SDL.h
05:04 PM rue_mohr: that explains it
05:04 PM veverak: not really
05:04 PM rue_mohr: doyou want to try installed SDL1.2?
05:04 PM rue_mohr: yes, see the dirs
05:05 PM rue_mohr: you ahve sdl2 its looking for sdl1.2
05:05 PM veverak: /usr/include/SDL/SDL_gfxPrimitives.h:38:10: fatal error: SDL.h: No such file or directory
05:05 PM veverak: 38 | #include <SDL.h>
05:05 PM veverak: there is SDL.h in /usr/include/SDL/
05:05 PM rue_mohr: usr/include/SDL/SDL_gfxPrimitives.h
05:05 PM rue_mohr: not
05:05 PM rue_mohr: usr/include/SDL2/SDL_gfxPrimitives.h
05:05 PM rue_mohr: you have a broken sdl1.2 install it seems
05:05 PM rue_mohr: libsdl1.2-dev
05:05 PM rue_mohr: see if you can install that
05:05 PM veverak: I have both sdl/sdl2
05:05 PM veverak: can't
05:06 PM veverak: it's debian package
05:06 PM veverak: I told you I am no arch multiple times ;P :)
05:06 PM rue_mohr: ok, then its just a bit more fun to figure out
05:06 PM rue_mohr: try chaning my includes to <SDL2/SDL.h>
05:07 PM rue_mohr: try locate SDL.h see what comes up
05:07 PM rue_mohr: have you ever compiled an SDL anything?
05:07 PM veverak: nah
05:07 PM veverak: this is funny
05:07 PM veverak: 7 #include <SDL/SDL.h>
05:07 PM veverak: works
05:07 PM veverak: this is from your file
05:08 PM veverak: the include from SDL_gfxPrimitives.h is broken
05:08 PM veverak: anyway
05:08 PM veverak: rue_mohr: you using sdl2 ?
05:08 PM rue_mohr: well I ahve both isntalled, but my source implies that I'm using 1.2
05:09 PM rue_mohr: ok, so, gfx
05:09 PM rue_mohr: <SDL/gfx.h>
05:09 PM veverak: wont' work with SDL2
05:09 PM rue_mohr: hmm I already put SDL in the path
05:09 PM rue_mohr: so, I wonder why the compile paths are broken
05:09 PM rue_mohr: wee need a -I
05:10 PM veverak: lol
05:10 PM veverak: -isystem /usr/include/SDL/
05:10 PM veverak: HACKED
05:10 PM veverak: wooow
05:10 PM veverak: this is supersmooth
05:10 PM veverak: rue_mohr: cool
05:10 PM rue_mohr: :) yea
05:10 PM rue_mohr: ok, so
05:11 PM rue_mohr: right
05:11 PM rue_mohr: heh
05:11 PM rue_mohr: thanks
05:11 PM rue_mohr: now you can look at my math if you like
05:11 PM rue_mohr: I can do up a solver for 3d with it
05:11 PM rue_mohr: if you want
05:11 PM veverak: yeah 3D is easy with this config
05:12 PM veverak: and this seems as what I would expect
05:12 PM rue_mohr: maybe you can use it for sanity testing what your working on
05:12 PM veverak: doubles are actually needed?
05:12 PM rue_mohr: no
05:12 PM rue_mohr: I just had math problems with things in general and started throwing doubles at everything
05:12 PM veverak: (so far I test the IK with FK, seems ok)
05:12 PM rue_mohr: as is that
05:12 PM rue_mohr: thats FK test of the IK
05:13 PM rue_mohr: funny as must of it is the FK
05:13 PM rue_mohr: :)
05:13 PM rue_mohr: most of that code
05:13 PM rue_mohr: the IK is just those 4 lines
05:13 PM rue_mohr: I'll leave you to play, I have to put a new PULL CORD ON A LAWN MOWER
05:13 PM veverak: :D
05:14 PM veverak: (playing with my code now tbh)
05:14 PM veverak: (really want to finish the unfolding)
05:18 PM veverak: hmm
05:19 PM veverak: rue_mohr: thinking that your thing could be used in university course
05:33 PM rue_mohr: sure , rmember credits
05:34 PM rue_mohr: not like I ever end up putting any in my source
05:34 PM rue_mohr: its the library I wrote for my robots
05:34 PM rue_mohr: funny enough I'v not gotten around to ik for them yet
05:44 PM veverak: fu me
05:44 PM veverak: IK works perfectly in random tests
05:44 PM veverak: but not in the real robot
05:44 PM Tom_itx is now known as Tom_L
05:44 PM zhanx: ????
05:44 PM zhanx: veverak, you need to explain your reasoning on that one
05:45 PM veverak: zhanx: random tests: I have test model that is moved to default pso for the arm
05:45 PM veverak: and random positiones are generated relative to the default pos
05:45 PM veverak: (positions of tip)
05:45 PM veverak: +-0.025 in all axes
05:45 PM veverak: I do same stuf in visualizer with the robot of the robot I actually have
05:46 PM veverak: goes into random positions
05:46 PM zhanx: weird
05:46 PM veverak: yes
05:46 PM veverak: time to sync model of real robot with test robot
05:46 PM zhanx: i use IK now for my bots
05:46 PM zhanx: works fine for me
05:46 PM veverak: what kind?
05:47 PM zhanx: there is more than one kind?
05:47 PM zhanx: if so i am just calculating end point location at max 3 dof
05:49 PM veverak: ah, got generic solver based on jacobian
05:50 PM veverak: NOW TESTS FAIL
05:50 PM veverak: good, sanit restored
05:50 PM veverak: *sanity
05:53 PM veverak: so, only difference in models?
05:53 PM veverak: real models has non-zero-quaternions in the model shape
05:53 PM veverak: dimensions were same
05:54 PM veverak: seems like I have math error
06:18 PM veverak: (of course there was error in my math)
06:41 PM rue_mohr: ok, well, you seen some of what mine can do