Back
[03:18:16] <rue_shop5> so I ahve, lets say 3 robot arms, I'd like to get to do something togethor
[03:18:30] <rue_shop5> I'm thinking maybe just one program that does it all at once
[03:18:45] <rue_shop5> hmm
[03:56:22] <z64555> would make things somewhat easier to coordinate, since one program would know the positions, states of the arms vs. trying to communicate with other programs
[03:57:04] <rue_shop5> yea, the only way to communicate between the progams otherwise would be local network port
[03:57:13] <rue_shop5> which comes with a tonne of overhead
[04:03:48] <rue_shop5> Tom_itx, hehaha the drywall router? yea turns out its a great speed at 24VDC with good torque
[04:04:31] <rue_shop5> 1.5A
[04:04:45] <rue_shop5> turns out, I have some 24VDC 2A supplies sitting on the windowstill
[09:23:18] <shulky> good day, good people :)
[09:23:37] <shulky> are hall sensors used in robot lawnmowers for perimeter wire detection?
[09:24:14] <shulky> ... or ... well, what is used to detect them?
[09:26:12] <deshipu> I'm not sure "robot lawnmovers" even exist outside of some early experiments
[09:26:27] <deshipu> but I'd guess they just have induction coils
[09:27:12] <shulky> well, so did i think...
[09:27:37] <shulky> but... husqvarna automower is quite an existent brand
[09:27:55] <shulky> robomow
[09:27:57] <shulky> etc.
[09:30:28] <shulky> people here (
http://forum.arduino.cc/index.php?topic=118885.0) and here (
https://www.reddit.com/r/robotics/comments/2eyhe7/idly_thinking_about_robotic_lawnmowers/ck4h6xr) mention Hall
[09:30:34] <shulky> which one is more optimal?
[09:31:01] <shulky> optimal is not comparable, so let's say "closer to an optimal solution" :P
[09:38:07] <shulky> thought maybe you guys would know :/
[09:38:35] <shulky> is there some other channel where people are more focused on a niche topic like robot lawnmowers?
[09:45:53] <shulky> li'l bumpy :)
[09:54:15] <z64555> just because the transmission is instant doesn't mean you'll instantly get a response. :)
[09:55:10] <z64555> hall effect sensors require the presence of a magnet in order to work
[09:55:22] <z64555> they literally sense the magnet field. no magnet, no field
[09:57:06] <z64555> you might be able to establish a field using a pair of wires, but I don't think it would be strong enough for a hall effect sensor to detect over the EMI from the engine
[09:57:59] <z64555> I'd try something optical, or radio based
[10:05:56] <SpeedEvil> http://www.ardumower.de/index.php/en/induktion
[10:21:34] <shulky> z64555, sorry, I won't ever be able to completely understand IRC mechanics
[10:21:57] <shulky> *I will never be...
[10:22:19] <shulky> oh, hi SpeedEvil :)
[10:22:41] <shulky> sorries to both of you for slow responses
[10:23:37] <ace4016> niche channels tend to have fewer and fewer active users and less liklihood of existence
[10:24:17] <shulky> and, forgot to say regarding IRC: I can't really tell when people are offline, AFK or idling while ignoring my stupid messages because they make no sense
[10:25:11] <ace4016> :P
[10:25:22] <shulky> there is no such thing as "UserName is typing" or "Seen HH:MM" on IRC, so I don't have a proper perception of communication over here
[10:25:27] <robopal> lets bring some AI in this channel guys
[10:25:27] <shulky> sincerely sorry for that chaps
[10:27:53] <shulky> oh, especially if sarcasm :D
[10:28:28] <deshipu> idle : 0 days 0 hours 0 mins 23 secs
[10:28:39] <deshipu> there is actually an "idle" counter
[10:30:10] <deshipu> but anyways, we do other things while looking at the channel once in a while
[10:30:23] <deshipu> also, I have no idea about a channel that would be better for this
[10:30:32] <deshipu> I suppose you will need to do your own research
[10:31:19] <shulky> indeed, reorienting the story in my work from Hall sensors to (quite generally speaking) inductive sensors
[10:34:56] <z64555> *magnetic
[10:35:47] <z64555> "inductive" sensors make little sense
[10:36:57] <z64555> well... maybe not entirely. there's such a thing as capacitive sensors, so I guess there could be inductive
[10:37:45] <z64555> but in the case of inductive sensors it wouldn't be sensing a magnetic field alone, it would also sense the existence of a ferrous object
[10:38:15] * z64555 tries to wake up
[10:38:20] <z64555> coffee hasn't hit me yet
[10:41:14] <shulky> haha, don't bother so much, I'm considering to redo it in a GPS way
[11:10:19] <Tom_itx> yeah that wouldn't work so well
[11:10:56] <Tom_itx> they are generally used for very short range
[11:11:03] <Tom_itx> like wheel sensors on cars etc
[12:21:12] <shulky> um, uh, can GPS coordinates give directions too?
[12:22:31] <shulky> if not, which device/sensor would report it most accurately?
[12:26:33] <z64555> GPS can give heading, but only if you're moving
[12:27:02] <z64555> it's recommended you use at least a magnetometer (MEMS compass)
[12:27:46] <shulky> I see, I hoped I'd get away with GPS only, thanks for confirming :P
[12:29:46] <z64555> GPS has an inherit inaccuracy built-in for civilian applications. You still have to use dead reckoning methods to get a good enough resolution
[12:31:50] <shulky> relying solely on dead reckoning won't provide current orientation if it changes without moving in a particular direction
[12:32:50] <z64555> you can get heading on a compass without moving
[12:34:20] <z64555> the compass requires software filtering to filter out the EMI and radio frequency interferance, generally meaning the compass updates at a slow rate
[12:34:29] <shulky> but not on a device (ahem, lawn mower) with only GPS built in
[12:34:42] <z64555> but, you can fuse the compass with a gyro and have a faster response
[12:35:08] <shulky> by compass you mean the MEMS magnetometer?
[12:35:11] <z64555> yes
[12:35:18] <shulky> aha, ok, sorry :P
[12:35:29] <z64555> easier to type :)
[12:35:39] <shulky> yup, agreed
[12:38:38] <shulky> is using kalman filter with only magnetometer and GPS enough?
[12:39:49] <z64555> dunno, I haven't progressed to that stage yet. :P
[12:40:06] <shulky> ah, okay then
[12:40:26] <z64555> I need to grab a new IMU with a magnetometer and barometer on it
[14:38:16] <[VEGETA]> hi, anyone here has experience in ROS?
[14:43:43] <[VEGETA]> robot operating system
[14:43:58] <[VEGETA]> i want to build a robot with kinect sensor to build 3d map
[14:44:14] <[VEGETA]> and make it available to other robots via a server
[14:47:08] <deshipu> some people here have some experience
[14:47:36] <[VEGETA]> I'd like to speak to them
[14:49:03] <deshipu> go ahead
[14:49:29] <[VEGETA]> I've said my stuff above, waiting for them
[14:49:42] <deshipu> ok
[14:50:05] <deshipu> you would be better off if you asked a specific question, though
[14:53:06] <[VEGETA]> my problem is I don't know where to start
[14:53:22] <[VEGETA]> what is the packages I must use to do the tasks I wanted
[14:53:38] <[VEGETA]> what I want is like the following
[14:53:49] <deshipu> I would probably start with the ros tutorial on their official website
[14:54:09] <[VEGETA]> a robot with raspberry pi 3 with kinect sensor, collect data and send it to server
[14:54:31] <[VEGETA]> server builds a 3d map and calculates obstacles and stuff, then send info to the robot
[14:54:39] <deshipu> http://wiki.ros.org/ROS/Tutorials
[14:54:50] <[VEGETA]> the robot itself is controlled by an ARM board running freeRTOS
[14:55:00] <deshipu> also check out
http://wiki.ros.org/ROS/StartGuide
[14:55:45] <[VEGETA]> so I also need the main ROS in the server (with the map) to send some data to other non-ros robots
[14:56:04] <[VEGETA]> especially the map and the obstacles
[14:56:30] <[VEGETA]> that is my main idea
[14:56:48] <[VEGETA]> btw, my robot will be custom robot so no specific hardware
[14:57:07] <[VEGETA]> if there are someone with good experience, please tell me what should I do
[14:57:30] <deshipu> start with the tutorials
[14:58:42] <[VEGETA]> try them in my laptop only? without pi3 or any other robot?
[14:58:58] <deshipu> yes
[14:59:07] <[VEGETA]> the subscribe, publish, message, topic,....bla bla thing is the hardest
[14:59:32] <deshipu> you have to understand it to be able to do anything
[14:59:47] <deshipu> I mean, whole ROS is based on that
[15:00:16] <z64555> all the fun stuff has a mountain of boring stuff that it stands on
[15:01:48] <[VEGETA]> I am excited to learn it, but how many months or so do I need to get started good?
[15:02:14] <[VEGETA]> I have 2 years to do my research which is what you saw up there
[15:02:16] <z64555> that really depends on how much time you're willing to spend each day to fiddle with this
[15:02:28] <z64555> ...and really outside our expertise
[15:02:37] <[VEGETA]> not to mention freeRTOS (which i didn't try yet) and making a gui with pi3 and win10iot
[15:02:57] <[VEGETA]> ok, from your experience, give me an example
[15:03:26] <z64555> look for libs or projects that you can grab or fork that'll help you focus on what you really want
[15:04:01] <deshipu> that's a good approach actually -- changing an existing project to see what happens and learn
[15:04:16] <deshipu> (not so good to build your own, of course)
[15:04:18] <[VEGETA]> if I work around 5 hrs a day on ROS, how many time do i need? < approx
[15:04:31] <z64555> m8
[15:04:43] <deshipu> nobody did such measurements
[15:04:46] <z64555> we're engineers.
[15:04:49] <z64555> not managers
[15:04:59] <z64555> :)
[15:05:08] <veverak> [VEGETA]: ros takes a lot of time
[15:05:11] <veverak> like a really a lot
[15:05:14] <veverak> but nah
[15:05:20] <veverak> somehow I still got the feeeling that it was worth it
[15:05:40] <deshipu> famous last words
[15:06:19] <[VEGETA]> vaverak, what is the duration of your learning approx? in order to build something like i described
[15:06:33] <[VEGETA]> I am convinced it is worth it
[15:06:48] <[VEGETA]> but I have limited time and there are other stuff too
[15:06:48] <z64555> you sure you're not a business major? :)
[15:07:07] <[VEGETA]> no, i am a mechatronics engineer
[15:07:22] <[VEGETA]> but, z64555, i explained why i care about time
[15:07:24] <deshipu> you will probably need to learn ros sooner or later anyways
[15:07:53] <[VEGETA]> to be honest, i am planning to apply for a research scholarship in Japan
[15:08:04] <[VEGETA]> and this is my research proposal idea
[15:08:10] <[VEGETA]> duration is 2 years
[15:08:17] <[VEGETA]> full of stuff other than ros
[15:08:28] <z64555> I'm guessing at least 2 college semester's worth
[15:08:30] <[VEGETA]> i still have around 8 months before it begins
[15:09:23] <[VEGETA]> what about freertos? i guess it is way much easier than ros xD
[15:09:41] <z64555> basically the same thing, although freertos is more general
[15:10:16] <[VEGETA]> the two OSes will be on the same robot
[15:10:35] <veverak> [VEGETA]: depends on skills you got
[15:10:36] <[VEGETA]> ROS in Pi 3, and FreeRTOS in stm32 board or something
[15:10:39] <veverak> I am "learning" ros for a year
[15:10:42] <veverak> :)
[15:10:58] <[VEGETA]> did you do any actual thing in it?
[15:11:48] <veverak> work in progress
[15:12:01] <z64555> mkay... so you've got a scout bot with the kinect lidar on it, a central brain, and a bunch of swarm bots
[15:12:16] <veverak> [VEGETA]: got Bc. thesis -> develope software for quadruped robot
[15:12:23] <veverak> deadline is in year +-
[15:12:29] <veverak> [VEGETA]: again, it depends on what you now
[15:12:46] <veverak> being linux admin + experience with RC cars + wrote a lot of python ocde
[15:12:49] <veverak> helps a lot
[15:13:35] <[VEGETA]> I know some linux basics, wrote C programs for PIC MCUs, RC cars are not hard too
[15:15:35] <veverak> looks cool
[15:15:37] <veverak> [VEGETA]: try and see
[15:15:43] <veverak> problem with ROS is that there are too many packages
[15:15:46] <veverak> :D
[15:15:54] <[VEGETA]> ^ that is the damn thing
[15:16:25] <[VEGETA]> kinect is basically an rgbd sensor... there are multiple packages for rgbd-slam
[15:16:43] <[VEGETA]> but i guess my problem is 2 things
[15:17:17] <[VEGETA]> one: ros architecture is not clear to me (topics, subscribe, message, build,...)
[15:17:38] <[VEGETA]> two: using other people's code, as I am not a good programmer myself
[15:17:42] <Polymorphism> Vegeta!! What does the scouter say about his power level!?!?!
[15:19:22] <z64555> a programmer's skill is not just based on their ingenuity and daftness, but also on their ability to take what is there and add the necessary bits to do something new or different
[15:19:39] <[VEGETA]> it is over nine thousaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaand!
[15:19:56] <z64555> 64555 level, actually.
[15:20:04] <Polymorphism> What 9000!?! there's no way that could be right...
[15:20:09] <veverak> [VEGETA]: ad. ros architecture: trial&error
[15:20:16] <veverak> just try to write simple things, play with it
[15:20:23] <Polymorphism> start with C
[15:20:24] <veverak> it's not complicated but took me while too
[15:20:26] <veverak> :)
[15:20:34] <z64555> Polymorphism: not helpful m8
[15:20:50] <deshipu> Polymorphism: you start with A
[15:20:53] <Polymorphism> deshipu, xD
[15:21:03] <deshipu> Polymorphism: then B, and then you can do C
[15:21:11] <[VEGETA]> polymorphism, you are one of those who liked DBZ like myself = a living legend
[15:21:12] <Polymorphism> right, what was I thinking
[15:21:35] <Polymorphism> [VEGETA], indeed
[15:21:56] <[VEGETA]> ROS's level is over 9000
[15:23:15] <jason_> ROS is actually pretty simple. It's really just a framework for moving data around, and then a bunch of packages that make use of the data.
[15:23:15] <Polymorphism> z64555, whats wrong with C/C++
[15:23:27] <[VEGETA]> mmm do you know a mobile robot platform with steering front wheels? instead of doing my own...
[15:23:30] <z64555> a mountain can become a hill and then a plain given time and perserverance. but it shall stay a mountain until you start digging
[15:24:02] <Polymorphism> http://lcamtuf.coredump.cx/rstory/
[15:24:16] <Polymorphism> thats a nice link
[15:24:18] <Polymorphism> for those who make robots
[15:24:31] <z64555> usually you'll have to fabricate your own robot chassis, there are a few commercially available
[15:24:58] <Polymorphism> I just bought a 3d printer and a cnc, that should be no problem
[15:25:10] <Polymorphism> but what will my robot do...
[15:26:07] <z64555> [VEGETA]: could hack a powered wheelchair, but it might be a bit too big
[15:26:14] <[VEGETA]> I mean, a chassis with good wheels and front steering motor and a good size is good enough
[15:26:59] <Polymorphism> https://www.youtube.com/watch?v=4rg2Mrgq3ss hahahahahah
[15:29:20] <Polymorphism> powered wheelchair would be nice
[15:29:26] <Polymorphism> if it drives me around
[15:29:28] <Polymorphism> I like the sound of that
[15:29:31] <Polymorphism> like dune
[15:29:56] <Polymorphism> or whatever movie that was
[15:33:35] <z64555> Baron Harkkonen in Dune had suspensors, basically like suspenders but instead of keeping your pants up it levitates you a tiny bit.
[15:34:40] <z64555> he had them because he was so grossly overweight that he could not stand on his own power, the suspensors allowed him to walk or float wherever he wanted
[15:34:41] <deshipu> so you can grow fatter
[15:35:23] <veverak> [VEGETA]: steering front wheels?
[15:35:23] <deshipu> of course today fatness is a disease of poor people
[15:35:25] <veverak> pick RC Car
[15:35:27] <veverak> :)
[15:35:45] <deshipu> I think differential steering is much easier
[15:35:48] <deshipu> tank-like
[15:36:06] <deshipu> you don't need a differential drive then
[15:36:37] <z64555> 2-wheel with caster is perhaps the easiest
[15:36:46] <deshipu> also, you can build a simple robot for experiments out of 2 servos and a pro mini for less than $10
[15:37:00] <Polymorphism> lol
[15:37:11] <deshipu> with some bumper swithces and a distance sensor
[15:37:20] <deshipu> you can do a lot with that
[15:37:33] <deshipu> learning-wise
[15:44:06] <[VEGETA]> i meant the 2 front wheels can steer rather than differential drive. meaning, controlled by dc or servo moto
[15:44:10] <[VEGETA]> *motor
[15:44:32] <[VEGETA]> i made a youtube video to generate 50hz signal using pic just for that xD
[15:45:05] <veverak> tamiya rc cars
[15:45:10] <veverak> using 'pwm servo' as steering
[15:45:17] <veverak> 'pwm motor regulator' for driving
[15:45:22] <veverak> easy to control with PWM
[15:46:09] <z64555> pathfinding routine is going to be a bit trickier than differential, but ok
[15:46:14] <deshipu> but trajectory planning...
[15:46:18] <z64555> it's your pain. :)
[15:48:53] <deshipu> paint is the sign of weakness leaving the body
[15:48:57] <deshipu> pain
[15:49:03] <deshipu> thank you Freud
[15:51:02] <[VEGETA]> <z64555> it's your pain. :) < pure evil
[15:52:13] <z64555> deshipu: that should be your motto or something :D
[15:52:43] <deshipu> z64555: unfortunately marines beat me to it
[15:52:55] <z64555> not the paint, however :P
[15:53:30] <[VEGETA]> veverak, these tamayas are nitro rc cars rigth
[15:53:36] <[VEGETA]> right*
[15:53:48] <z64555> [VEGETA]: good, bad, I'm the guy with the source
[15:53:55] <[VEGETA]> plus, can it hold 2 boards and a kinect with a servo?
[15:54:12] <veverak> [VEGETA]: nope
[15:54:17] <veverak> :)
[15:54:36] <[VEGETA]> z64555, so you have the source and we must do w/e u order to give it to us xD
[15:54:49] <veverak> [VEGETA]: electric/nitro different sizes
[15:54:49] <deshipu> [VEGETA]: really, start with simple robots and ramp up, you won't build your final robot now
[15:54:56] <veverak> anyway, deshipu is right
[15:55:05] <veverak> go small, than go big
[15:55:25] <z64555> preferabbly get something sturdy and easily repairable
[15:55:43] <deshipu> [VEGETA]:
https://hackaday.io/page/1252-small-robots
[15:56:16] <[VEGETA]> I have this one with me:
http://www.aliexpress.com/item/New-4-wheel-Motor-Smart-Robot-Car-Chassis-with-steering-engine-For-Arduino-Free-Shipping/32266961086.html
[15:56:18] <z64555> or super cheap. Basically something you can crash into smitherines and not feel much pain in your wallet
[15:56:44] <z64555> We'll call it a "trainer"
[15:56:45] <z64555> :)
[15:58:00] <z64555> dohhoho... that T-frame
[15:58:28] <z64555> I-frame, rather.
[15:58:40] <[VEGETA]> i guess that is enough for now
[16:03:18] <z64555> [VEGETA]: If you're a college student and have access to the campus machine shop, you should fab some replacement parts for it
[16:03:34] <z64555> otherwise it looks like a good trainer chassis
[16:04:47] <[VEGETA]> i have no access now... I am an engineer in a factory
[16:05:00] <[VEGETA]> i plan to go to japan via mext scholarship
[16:25:55] <Jinxit> do we have any control theory channels on freenode?
[16:29:42] <deshipu> I think freenode is more programmers than mathematicians
[16:40:10] <rue_shop5> you want to know about PID and th elike?
[16:42:53] <Polymorphism> Jinxit, are you looking for work
[16:43:40] <Jinxit> Polymorphism: hmm?
[16:43:59] <Jinxit> rue_shop5: thinking a bit more advanced, underactuated systems mainly
[16:45:23] <deshipu> there is a large group for that at MIT
[16:45:55] <Jinxit> yeah I've been following the online course, but it's from 2009
[16:46:05] <Jinxit> so I'm trying to see what the state of the art is
[16:46:43] <deshipu> Jinxit: that course is every year
[16:46:52] <deshipu> Jinxit: it was there this year on edx too
[16:47:05] <deshipu> hmm, I'm actually subscribed
[16:47:32] <deshipu> Jinxit:
https://www.edx.org/course/underactuated-robotics-mitx-6-832x-0
[16:49:02] <Jinxit> oh I must've misread
[16:49:11] <Jinxit> last time I checked there I somehow saw it as cancelled
[16:50:03] <deshipu> I subscribe each time just to watch the videos, but never have the time to do the exercises
[16:51:01] <Jinxit> they seem much more condensed
[16:51:18] <Jinxit> the 2009 ones are all 1h+
[16:51:50] <Jinxit> but yeah I'll definitely take a look at this, thanks
[16:54:47] <Polymorphism> whats your price Jinxit
[21:18:50] <anniepoo_> Jinxit - #symbo1ic might be at least a place to ask that, as might #math