#robotics Logs

Dec 19 2018

12:00 AM zhanx: i need to make rue_mohr a cook book
12:05 AM rue_mohr: I have a lot of thigns to work out if I'm gonna live 10000 years
12:06 AM rue_mohr: including how to live and not just stay alive
12:06 AM rue_mohr: but also how to stay alive
12:06 AM zhanx: living is one thing yes
12:06 AM rue_mohr: prioritization, even when it hurts
12:07 AM rue_mohr: in 10000 years, there be a hell of a lot of hurtin
12:09 AM zhanx: in 10000 years i will be a history footnote
12:10 AM rue_mohr: lots of things will be
12:11 AM rue_mohr: 8000-10000 tho
12:12 AM rue_mohr: dunno when the light from the supernova makes it here
12:12 AM rue_mohr: better be spectacular
12:15 AM teej: Supernova.
12:17 AM rue_mohr: oops I said that all in the wrong channel...
12:18 AM rue_mohr: here's hoping I dont need a spaceship, cause it looks like I'm gonna end up building a hous
12:18 AM rue_mohr: see if I'd started that underground lair YEARS ago, I'd not need to be building a new shop
12:20 AM rue_mohr: hmm, actaully the last plan was to build a robot that naturally just made huge caverns and leave it running somewhere
12:20 AM zhanx: ok batman
12:20 AM rue_mohr: hey, no building permit, no square footage limit
12:21 AM rue_mohr: this is stupid, I'm waiting for a printer to finish one lot of things so I can start another one
12:21 AM rue_mohr: 'cause I dont like the colour of filament thats in the other printer
12:22 AM rue_mohr: will I be reduced to creating printers for each colour of filament I get?
12:23 AM rue_mohr: what do you want for supper rue?
12:23 AM rue_mohr: to be a robot!
12:23 AM rue_mohr: *sigh*
12:23 AM rue_mohr: I dont want to eat
12:23 AM rue_mohr: you need to eat
12:23 AM rue_mohr: it just upsets my stomach and I dont get good sleep
12:23 AM rue_mohr: you still need to eat
12:24 AM rue_mohr: how about a peanut?
12:24 AM rue_mohr: and wake up at 1am starving again?
12:24 AM rue_mohr: I could keep other peanuts by the bed?
12:24 AM rue_mohr: not a good plan
12:25 AM rue_mohr: its 10:24
12:25 AM rue_mohr: mhm
12:25 AM rue_mohr: you should have had supper ages ago
12:26 AM rue_mohr: does this mean were into emergency options?
12:26 AM rue_mohr: not cholocolate!
12:26 AM rue_mohr: .. sure, how about veg soup
12:27 AM rue_mohr: its not gonna be a good morning
12:27 AM rue_mohr: its ok! look! bag of peanuts!
12:27 AM rue_mohr: whaaaaa?
12:27 AM rue_mohr: not good plan
12:28 AM rue_mohr: aaaaaaaaa?
12:29 AM rue_mohr: you didn't tell zhanx about turning that alg into a med scanner
12:29 AM rue_mohr: keep 'er on the todo list
12:29 AM rue_mohr: we cant to everything tho
12:30 AM rue_mohr: we can try
12:30 AM rue_mohr: were not enough peoples
12:30 AM rue_mohr: I dont think anyone is on your track dude
12:30 AM rue_mohr: yea I know
12:30 AM rue_mohr: sorry
12:30 AM zhanx: rue ranting again
12:30 AM rue_mohr: but we can still get helps!
12:30 AM rue_mohr: sorry
12:30 AM rue_mohr: srry z
12:30 AM zhanx: its fine
12:31 AM zhanx: wait till this bot hits you
12:31 AM rue_mohr: sometimes the echo is the only one there
12:31 AM zhanx: there is a reason i am testing the plastics
12:31 AM rue_mohr: I need more energy
12:31 AM rue_mohr: I need to know whats wrong
12:31 AM rue_mohr: we think something is wrong
12:32 AM rue_mohr: cant prove it
12:32 AM rue_mohr: I can
12:32 AM rue_mohr: you didn't build the scanner tho
12:32 AM zhanx: i need a scanner for my brain
12:32 AM rue_mohr: its not hard
12:32 AM zhanx: I did make the scanner code is not done
12:32 AM rue_mohr: well, skull would be
12:32 AM zhanx: I keep changing my mind on it
12:33 AM rue_mohr: rue_bed, I'm falling sleep, pls start supper
12:33 AM rue_mohr: zhanx, we need that kalman filter
12:33 AM rue_mohr: light src on / light src off delta = image
12:33 AM rue_mohr: heh.
12:33 AM rue_mohr: well, maybe he will do it for us eh?
12:33 AM rue_mohr: na
12:33 AM zhanx: i need to have the pain to stop so i can
12:34 AM rue_mohr: pain is distracting
12:34 AM zhanx: today was a bad day for the pain
12:34 AM rue_mohr: tell him how to overmodulate the nerves
12:34 AM rue_mohr: there is no input mechanism tho
12:34 AM rue_mohr: hmm
12:35 AM rue_mohr: what do you think it looks like
12:35 AM rue_mohr: dunno
12:35 AM rue_mohr: we should check the cerebelum mapping again
12:35 AM rue_mohr: it still dosn't make sense thats its on the sensory
12:36 AM rue_mohr: those nerves cross over, I'm sure of it
12:36 AM zhanx: yes it does
12:36 AM zhanx: prevents further injury
12:36 AM rue_mohr: but zhanx its output timing
12:36 AM zhanx: and this year i pushed myself hard
12:36 AM zhanx: i am paying the price for it
12:36 AM rue_mohr: he has a bad disc
12:36 AM rue_mohr: or a few
12:37 AM zhanx: knees, shoulder also
12:37 AM zhanx: turns out the body is not made to lift weights like i did for years
12:38 AM rue_mohr: ewch, arthritis...
12:38 AM rue_mohr: disc injection?
12:38 AM zhanx: doc said he had seen worse cartilage in a shoulder than mine, but it had some
12:38 AM rue_mohr: they dont know how to
12:39 AM zhanx: replacement
12:39 AM rue_mohr: can you hydraulic a joint apart?
12:39 AM rue_mohr: technically , yes
12:40 AM rue_mohr: technically?
12:40 AM zhanx: i suppose
12:40 AM rue_mohr: if you scar the surfaces your screwed
12:40 AM rue_mohr: what if they are already
12:40 AM rue_mohr: then you prolly cant get them apart anyho
12:40 AM rue_mohr: thas the morining 'wont move' thing
12:40 AM rue_mohr: yea
12:40 AM rue_mohr: oh
12:41 AM zhanx: or the evening used it too much and it stopped
12:41 AM rue_mohr: what promotes tissue regrowth
12:41 AM zhanx: my hands do that, no grippy thing
12:41 AM rue_mohr: eating supper!
12:41 AM rue_mohr: oh yea
12:41 AM rue_mohr: soup
12:42 AM rue_mohr: ok
12:42 AM zhanx: eat
12:42 AM rue_mohr: gonna wake up starving
12:42 AM rue_mohr: again
12:42 AM rue_mohr: ugh
12:42 AM rue_mohr: what wasn't even a small meal
12:42 AM rue_mohr: yea your gettin fat too
12:42 AM rue_mohr: thats not fat...
12:42 AM zhanx: hey now that is mean
12:43 AM rue_mohr: yea, lets throw snowballs at him
12:43 AM zhanx: ice balls here, its -15f
02:48 AM SleepyTurtle: Hi!is there a definition for the distance between two poses?
02:53 AM deshipu: hundreds
02:54 AM deshipu: you can pick any
03:02 AM ib0t: SleepyTurtle: normally you have a two dimensional distance, with one component for the linear distance and one for the angular
03:03 AM ib0t: if you want to have only one number you have to combine them with weights which depends on your application
03:04 AM ib0t: and it depends on the unit of your angular values and the linear
03:09 AM ib0t: oh and the angular distance is not so easy
03:10 AM ib0t: because for the linear you can say I have to travel x meters to reach my target
03:14 AM ib0t: for the angular you have to say I have to rotate x degrees around a special axis
03:16 AM ib0t: https://en.wikipedia.org/wiki/Angular_displacement#Three_dimensions
04:04 AM SleepyTurtle: deshipu, ib0t: thanks, I wanted a scalar value, it's for a rough cost estimate for navigation. I was curious about whether there was a standard way of finding the norm of a pose, but it makes sense for position and orientation to be weighed differently depending on the particular problem.
04:04 AM SleepyTurtle: Sorry for the late reply, I had to go afk for a while.
04:05 AM deshipu: SleepyTurtle: you can use any norm in the configuration space, or, if it's an arm robot, you can check the position of the tool
04:05 AM deshipu: SleepyTurtle: or a sum of distances of ends of all legs, etc.
04:06 AM deshipu: depending on what you actually need to do
04:07 AM SleepyTurtle: it's a mobile robot for 2.5 nav, so I'll just do position difference and weigh the orientation change around the Z axis.
04:08 AM ib0t: is the robot able to turn on place?
04:08 AM SleepyTurtle: it's diff drive
04:11 AM ib0t: okay than it is easy to say which distance one wheel has to drive to turn the robot from one angle to an other
04:12 AM ib0t: it is half of the wheelbase multiplied with the orientation change around the z axis
04:13 AM ib0t: so a good measurement in your application for the distance is the linear distance plus the angular distance around z axis multiplied with the wheelbase
04:13 AM ib0t: with angles given as radiants
04:14 AM ib0t: than the distance encodes how far one wheel has to go
04:14 AM ib0t: which is linear to how much work the motors has still to do
04:14 AM SleepyTurtle: I see, that's a great way of weighing rotations, making the cost be motor usage
04:15 AM SleepyTurtle: thanks
06:04 AM veverak: any ideas about this on walking robot?
06:04 AM veverak: :)
06:04 AM veverak: I have poses of it's body
06:04 AM veverak: (again, for cost estimate for navigation)
06:11 AM deshipu: well, the movement of the tips of the legs is a good estimation
06:15 AM veverak: deshipu: that's before the movement of itps of legs is planned
06:15 AM veverak: so yeah, 'estimate'
06:16 AM veverak: deshipu: https://imgur.com/a/jLsaLSH
06:16 AM deshipu: I'm sure there will be good correlation
06:16 AM veverak: at first, path for the robot as whole is picked (green line)
06:16 AM veverak: and leg movement is planned on the path
06:16 AM veverak: *than leg movement is planned...
06:17 AM deshipu: I really need to get back to my bots
06:17 AM veverak: so, when I pick the path for the robot as whole, I would like some estimate of distance between two poses
06:17 AM veverak: :)
06:17 AM veverak: whole idea of the thing is to 'estimate path for body that is most propably OK'
06:20 AM deshipu: yeah, and then figure out where to put the legs
06:21 AM deshipu: you know that real spiders and insects don't do anything like that?
06:21 AM deshipu: they just kinda wing it
06:23 AM veverak: maybe? :)
08:12 AM ldlework: rue_mohr: what?
08:33 AM ib0t: veverak: why not using the same estimation than with the diff drive, calculate the way one of the legs has to go
08:51 AM ib0t: I think for distance in navigation in a 2.5D environment it is never to bad to use as distance linear_distance + robot_radius * angular_distance
08:51 AM ib0t: because normally the effort to turn a robot depends on its dimensions
09:02 AM veverak: hmm, that could work
09:09 AM rue_mohr: ldlework, I lost the context
12:29 PM prasket: Anyone around into 3D printable robots? I am trying to decide between InMoov or ASPIR projects.
12:32 PM therealdurrf: im printing robots atm but they are tiny robots
12:37 PM prasket: oh yeah? Maybe I should start smaller lol. Have a link to the project you are working with?
12:38 PM prasket: InMoov seems much fuller featured but it also seems stale, ASPIR is pretty new but not a lot of software support for it yet.
12:42 PM deshipu: starting smaller is definitely a good idea
12:45 PM veverak: yup
12:50 PM prasket: I mean that does make sense, anyone have suggestions on smaller 3d printable robots to start with? I saw this link https://all3dp.com/1/3d-printed-robot-print-robots/ but only InMoov and ASPIR seemed to be the only user printable ones.
12:59 PM deshipu: does it need to be a humanoid?
01:00 PM veverak: tote!
01:00 PM veverak: :)
01:00 PM prasket: not at all
01:01 PM prasket: I liked the little spider with cam on the link I shared but was not able to find parts to print etc :(
01:04 PM prasket: and I would prefer to 3d print most of it and then buy the servos, arduino/Rpi, boards, etc
01:06 PM veverak: you can't effectively print servos/arduino/rpi
01:17 PM prasket: not yet at least :D especially on a Prusa i3 MK3 :D
01:21 PM veverak: :)