#robotics Logs

Jan 30 2019

12:09 AM rue_mohr: ai-thinker sounds familiar
12:19 AM rue_mohr: oh, ai thinker made the esp8266 carriers I got
01:23 AM SleepyTurtle: hi, does anyone have experience with the book "probabilistic robotics"?
01:23 AM SleepyTurtle: I have a couple of questions about the robot perception chapter
01:23 AM SleepyTurtle: thanks
03:19 AM deshipu: SleepyTurtle: it's possible nobody here has read it, but you can try asking the questions anyways
03:29 AM SleepyTurtle: it's about the learn_intrinsic_parameters algorithm
03:30 AM SleepyTurtle: p160
03:31 AM SleepyTurtle: p131 in this pdf https://docs.ufpr.br/~danielsantos/ProbabilisticRobotics.pdf
03:33 AM SleepyTurtle: it says that the intrinsic parameters affect the expectations and vice versa, which is only true of sigma and gamma. Does this mean it's taking the mixing parameters into consideration when calculating p_hit, p_short... ?
03:34 AM SleepyTurtle: also, I understand that Z is a vector of observations, is |Z| supposed to be the l1 mean?
03:37 AM SleepyTurtle: thanks, I've just been called so I might not answer in 15~20 min
04:37 AM SleepyTurtle: I'm back, had a surprise meeting
07:11 AM rue_bed: morning
07:11 AM veverak: morning
07:18 AM deshipu: good afternoon
08:57 AM rue_mohr: sleepys querstion sounded more like AI
09:03 AM deshipu: ai ai ai ai ai ai, puerto rico
02:47 PM nexm: hi all
04:07 PM Mikeee is now known as SWLIPDSINBFASOG
04:35 PM SWLIPDSINBFASOG is now known as Mikeee
04:54 PM nexm: I have a metal gear servo with slight backlash, anybody got an idea what cheapish servos have little to no backlash?
05:02 PM deshipu: none
05:07 PM Mikeee: if you want to reduce backlash, you should move the gears slightly closer to eachother.
05:07 PM Snert_: let me re-enginner the servo case...brb :)
05:13 PM Mikeee: That's not a difficult task, your joke is out of place
05:29 PM nexm: aren't the gears held on shafts? moving them slightly would require to make holes at different places, no?
05:32 PM deshipu: nexm: one simple way that might work sometimes is to add a spring (or rubber band) pushing the servo in one direction, so that any wiggle room is removed
05:32 PM nexm: thx, ill try that
05:32 PM deshipu: it works especially well for robots that have to fight gravity anyways, because you can make the spring work against the gravity, to make the servo's work a little easier
05:41 PM Mikeee: nexm just one shaft, and only slightly offset. 10 seconds with a drillbit to oblong the hole, then a dot of epoxy to push it closer to the other gears
05:43 PM Mikeee: if you have a shop press, you can also slightly compress the gear so it expands its circumference
05:52 PM nexm: thanks, I'll keep that in mind if I try to fix a larger servo, this one is tiny
06:04 PM Mikeee: ah yeah I was thinking it'd be a standard size servo, about the size of your thumb
06:04 PM Mikeee: if its tiny, it may be difficult to do
06:57 PM nexm: Mikeee: would that also work for joints (acrylic rods with a hole), that revolve around a shaft (=screw)?
07:42 PM flav0r: HEY HEY !!
08:34 PM rue_shop1: hey whats up
09:31 PM durrfbot is now known as durrfnotbot