Back
[00:20:16] <rue_house> ok, I'v just failed to objectify the program
[00:38:41] <Anniepoo__> you treated your program as a person?
[00:48:50] <AlienCat> the problem with subjective programming
[00:51:34] <rue_house> I want to make an object for the cnc machine and cant work out file io
[01:27:19] <rue_bed> think I worked out a model
[06:29:28] <scotch> hi
[06:29:57] <scotch> We got a six legged walking robot and want to climb an incline of 35 degrees
[06:30:17] <scotch> For this we need to make sure all feet have propper ground contact
[06:31:05] <scotch> The robot features a 6 DOF force/torque sensor in every hip, which allows us to estimate the forces on the food contact point
[06:31:13] <Jak_o_Shadows1> neat
[06:32:14] <scotch> We now went for an approach, to compute the expected forces, using the body configuration and then use a PID controller per leg to equalize expected measured forces
[06:32:32] <Jak_o_Shadows1> Is it working?
[06:32:47] <scotch> for now this sort of works, as our formula seems to be wrong
[06:32:56] <Jak_o_Shadows1> aha. Is the PID controller working?
[06:33:10] <scotch> the pid ist well tested an known to work
[06:35:37] <scotch> The formula for computing the expected force on the legs seems to be off
[06:35:50] <scotch> did any one do something like this before ?
[06:38:10] <Jak_o_Shadows1> No, but I probably have enough background to look at it a little
[06:38:12] <Jak_o_Shadows1> ping me tomorrow
[09:43:51] <artg02> hi all :)
[09:47:04] <artg02> The following document is of my university, the first part is about "system classification"
[09:47:13] <artg02> but I am totally not following there it is about
[09:47:18] <artg02> any help here?
[09:47:30] <artg02> I tried searching but nothing looks anything like this
[09:47:55] <artg02> https://www.expirebox.com/download/6390f7e80ed1e58c591181a4deb34d1b.html
[09:48:34] <deshipu> artg02: how are we supposed to know, you didn't even tell us what university it is
[09:49:03] <artg02> university of twente if that is of any importance
[09:49:23] <artg02> the problem is that the professor is not a good one and I was hoping someone recognized some of it
[09:49:48] <artg02> Just some keywords or name of formulas could help for better searching
[09:50:01] <deshipu> artg02: go to a different university then
[09:50:03] <artg02> I am not asking for a complete explanation of the whole document
[09:50:10] <artg02> deshipu: .........
[09:50:16] <deshipu> that has a better professor
[09:50:34] <artg02> professors are different on each subject
[09:50:46] <deshipu> you mean they are not cloning them?
[09:50:48] <artg02> sometimes its a battle sometimes a walk in the park
[09:50:59] <artg02> deshipu: could you just take a look please?
[09:51:44] <deshipu> what is your question exactly?
[09:52:29] <artg02> Where can I find more information about system classification?
[09:52:39] <deshipu> perhaps it would be easier for you to understand if you attended the lectures?
[09:52:40] <artg02> that is related to the topics mentioned in that pdf
[09:52:54] <deshipu> I think the lectures would be a good start
[09:53:22] <deshipu> failing that, any book on control theory
[09:54:10] <deshipu> you can start with wikipedia:
https://en.wikipedia.org/wiki/Control_theory
[09:54:36] <deshipu> especially point 10 there
[09:55:09] <veverak> :)
[09:55:19] <artg02> interesting point is that in the wiki they talk about the names of the system
[09:55:34] <artg02> and in the document they talk about "system of type 0" or 1
[09:55:52] <artg02> as I told before the professor is not good so attending is getting more stupid
[09:55:59] <artg02> I am trying to get help from other sources
[09:56:11] <deshipu> tell him then that he's not doing his job
[09:56:16] <artg02> I am trying to be proactive here
[09:56:17] <deshipu> he's getting paid for it
[09:56:19] <deshipu> I'm not
[09:56:39] <artg02> well if you help me get a sufficient on this topic I pay you 100 bucks
[09:56:44] <artg02> how does that sound
[09:56:47] <veverak> O_o
[09:56:48] <deshipu> I didn't attend those lectures and I have no idea what nomenclature he uses
[09:56:51] <veverak> o_O
[09:57:15] <deshipu> artg02: make that offer to the professor
[09:57:26] <artg02> btw, telling a professor he is not doing a good job is not a good thing for a grade
[09:57:46] <veverak> you have to tell it the right way
[09:57:50] <deshipu> if you only care about the grades, you can buy the diploma from China
[09:57:53] <veverak> than the faculty will defend you
[09:57:57] <veverak> if not
[09:58:02] <veverak> that university is shitty
[09:59:02] <deshipu> artg02: seriously, he has to have "open door" hours, just come to him and ask him for help
[09:59:09] <deshipu> artg02: I'm sure he will help you
[09:59:25] <deshipu> artg02: and you are helping the future students this way too
[09:59:42] <veverak> yup
[09:59:43] <deshipu> just don't tell him "your lectures suck"
[09:59:54] <deshipu> tell him "I'm having problems understanding this"
[10:00:14] <rue_house> tell him "we all are having problems understanding this"
[10:00:15] <artg02> I am having trouble understanding it and thus trying to understand it here
[10:00:28] <rue_house> whats it?
[10:00:31] <artg02> It is clearly related to robotics
[10:00:40] <artg02> as the whole study is closely related to it
[10:00:51] <rue_house> whats it?
[10:00:53] <deshipu> rue_house: categorization of systems for control theory
[10:01:00] <rue_house> hmm
[10:01:05] <artg02> Just taking a look at the document and giving some help wouldn't hurt right?
[10:01:38] <deshipu> artg02: I would gladly do that, but that document is pretty useless withthout the accompanying lecture
[10:01:55] <rue_house> what catagories did they make up for that?
[10:01:56] <deshipu> artg02: and also you still didn't ask a specific question that we could answer
[10:02:10] <deshipu> rue_house: "type 0", "type 1", etc. :)
[10:02:24] <deshipu> rue_house: gotta love the academics
[10:02:30] <rue_house> yea
[10:02:37] <artg02> deshipu: yeah exactly type 0 and type 1, where can I find more info about that
[10:02:43] <artg02> wiki is not of much help
[10:02:48] <artg02> type 0 or type 1 is not mentioned once
[10:02:51] <rue_house> I love it when you get to the real world and find out everyone uses different equations
[10:03:09] <deshipu> artg02: that's because he probably made up those names, or took them from a very specific book
[10:03:19] <veverak> artg02: go to university catalog
[10:03:22] <veverak> find details about course
[10:03:24] <rue_house> and if you sit down long enough, you find out that its the same, but focused on different variables with different constants mixed in and different units
[10:03:28] <veverak> usually, there should be books that were used
[10:03:40] <veverak> rue_house: more or less yeah
[10:04:01] <rue_house> with different variable names for the same things
[10:04:15] <deshipu> rue_house: names, ha! you mean letters
[10:04:20] <rue_house> yup
[10:04:35] <deshipu> and sometimes just some weird squiggles
[10:04:39] <veverak> letters are not important and can be replaced!
[10:04:46] <veverak> veverak: back to study!
[10:04:51] <veverak> veverak: I know I should but ...
[10:05:01] <artg02> the books that were used were "Continuous and discrete signals and systems" and "feedback control of dynamic systems"
[10:05:03] * deshipu slaps veverak with a large quadruped
[10:05:07] <artg02> both do not mention anything about it
[10:05:17] * veverak catches quadruped and wants to play with it
[10:05:20] <rue_house> so what are the attributes of type 1?
[10:05:28] * deshipu detonates the quadruped
[10:05:33] <rue_house> constrained rotation?
[10:05:35] * veverak cries
[10:05:48] <veverak> artg02: than I would suggest asking another people that took that course
[10:05:50] <veverak> again
[10:05:57] <deshipu> rue_house: he linked a doc:
https://www.expirebox.com/download/6390f7e80ed1e58c591181a4deb34d1b.html
[10:06:02] <deshipu> rue_house: have fun
[10:06:11] <veverak> we don't have crystall balls to deduce what exactly type0/1 means
[10:06:23] <deshipu> I think it's all about which derivative they use in the loop control
[10:06:24] <artg02> so as I understand here, names like type 0 and type 1 are not very widely used and may only used on this unveristy?
[10:06:52] <deshipu> artg02: we don't know if they are used only on this university, but we haven't heard them before
[10:06:57] <rue_house> just with him
[10:07:07] <deshipu> artg02: but I'm a math major, not and engineer, so...
[10:07:33] * veverak is on IT university, not robotics
[10:07:44] <veverak> but hey, thinking about changing university after bc degree
[10:07:57] <deshipu> veverak: come to ETH!
[10:08:59] * veverak would need place to sleep
[10:09:18] <rue_house> that just goes over PID in components
[10:09:47] <rue_house> not type 012
[10:10:30] <deshipu> they say "type 0 systems will always have steady state error"
[10:10:30] <veverak> hmm
[10:10:36] <rue_house> there will be orders of system responce
[10:10:41] <rue_house> thats P
[10:10:45] <rue_house> P is type 0
[10:10:52] <rue_house> proportionate control
[10:11:05] <rue_house> drive = control - position
[10:11:15] <rue_house> er
[10:11:20] <deshipu> and watch your fingers
[10:11:25] <rue_house> I suppose negitive feedback is more stable eh?
[10:11:33] <rue_house> drive = position - control
[10:11:49] <deshipu> every system stabilizes eventually
[10:11:53] <deshipu> as a pile of junk
[10:12:07] <rue_house> no it dosn't
[10:12:12] <veverak> damn it
[10:12:17] <rue_house> if you have too much P gain you can oscillate
[10:12:18] <veverak> deshipu: Zurich seems costly
[10:12:18] <deshipu> right, soe keep orbiting
[10:12:23] <deshipu> some
[10:12:30] <deshipu> veverak: they have stipends
[10:12:38] <rue_house> take a mic in a stadium with the volume up too high
[10:12:44] <veverak> up to 40% of costs
[10:12:45] <rue_house> EEEEEEEEEEEEEEEEEEEEEEEEEERRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRR
[10:12:53] <deshipu> veverak: and if you live here, everything else becomes cheap
[10:12:54] <rue_house> thats feedback with a gain of mroe tha 1
[10:12:54] <artg02> rue_house: so basically those parts are from the PID part and learning that will clarify the first part?
[10:13:09] <rue_house> dunno, ask this groupd about PID
[10:13:12] <deshipu> rue_house: well, sooner or later it will run out of juice
[10:13:21] <deshipu> rue_house: or get englufed by expanding sun
[10:13:24] <rue_house> ( P PI PD PID )
[10:13:50] <rue_house> integrators usually go out of control for me, so I like PD
[10:14:02] <veverak> even that 40% would not safe it
[10:14:04] <veverak> daaaamn
[10:14:21] <rue_house> digital implementation is full of traps like windup and saturation
[10:14:22] <deshipu> veverak: yeah, you would need to work for a bank
[10:14:39] <rue_house> and dont forget the evil ROLLOVER
[10:15:01] <deshipu> bethoven
[10:15:05] <rue_house> :)
[10:15:20] <rue_house> dont let your drive signal rollover, esp if its signed
[10:15:36] <veverak> damn it
[10:15:38] <veverak> hmm
[10:15:48] <veverak> they finished IoT esp board at hackerspace
[10:15:54] <veverak> should I stay or should I go?
[10:15:57] <veverak> veverak: you should study!
[10:16:22] <rue_house> this pdf is from your course?
[10:16:51] <rue_house> what does it mean to be using a python editor to edit its own source?
[10:16:57] <rue_house> is this bootstrapping?
[10:17:26] <deshipu> rue_house: the inverse
[10:17:54] <rue_house> :/ the cat wont let me have my chair back...
[10:17:59] <veverak> :D
[10:18:06] <deshipu> furry terrorists
[12:52:11] <tsglove> If an android smartphone is used for running ROS in a platform... how would I "interact" the smartphone with things like the motor controller?
[12:52:29] <tsglove> Via the smartphone's usb connection? Or does it go via bluetooth?
[12:53:27] <AlienCat> bluetooth would be fore nifty
[12:54:25] <tsglove> AlienCat, I am a n00b... so that´s why ask. So it would be smartphone --> bluetooth adapter? roughly?
[13:02:59] <AlienCat> probablly need some adapter yeah or chip. Actually I have no idea
[13:06:56] <tsglove> AlienCat, kk... thank you.
[13:07:42] <tsglove> I don´t even know if an Android Mobile would be the best solution. I want to make a small four wheeled robot to roam through a small backyard.
[13:07:57] <tsglove> I thought o fusing the android mobile because of the camera, wifi, bluetooth, etc.
[13:08:03] <tsglove> *thought of using
[13:26:48] <SpeedEvil> wifi could work - ESP8266 or similar
[13:45:14] <tsglove> SpeedEvil, esp82xx... will check that out.
[13:46:04] <tsglove> And I´m mentioning an Android smartphone just because I´m at a loss. I mean, what would be another option? A small-motherboard computer running linux? Maybe a couple of Arduinos plugged into her?
[13:46:28] <SpeedEvil> yes
[13:46:38] <SpeedEvil> raspberry pi, or any of the manyu similar boards
[13:58:15] <tsglove> super... on my way to check those out
[16:17:01] <rue_house> tsglove, cool
[16:17:06] <rue_house> what kinda robots you done so far?
[16:18:02] <rue_house> who wants to help me make a remote prescence robot?
[16:18:19] <rue_house> if you make the robot, I'll be remotly prescent..
[17:05:28] * DJuego says bye!
[18:36:52] <jrowe47> ok, so I've got my 3d print filament for parts - anyone have a recommendation for a conductive ink or paint to heat a nylon actuator?
[20:27:13] <rue_shop3> heheh
[20:31:58] <anniepoo> I've been looking at a number of places that are winding copper magnet wire around the nylon before the second coiling
[20:33:26] <jrowe47> im thinking i might rethink my whole idea about making this an actuator
[20:33:51] <jrowe47> and instead, design a bot with something more like muscles
[20:34:09] <jrowe47> hollow enclosures around the nylon, then pump coolant through
[20:34:17] * anniepoo is here, showing off her biceps
[20:34:30] <jrowe47> lol
[20:34:36] <jrowe47> copper magnet wire?
[20:34:39] <anniepoo> yes
[20:34:48] <jrowe47> oh
[20:34:49] <anniepoo> looks like 32 ga or so
[20:34:51] <jrowe47> neat, ok
[20:35:13] <jrowe47> i was just reading up on how to wire resistors to both ends
[20:35:21] <jrowe47> to heat up the wire more efficiently
[20:36:20] <jrowe47> was also thinking that a dual coolant / heat fluid system might make the whole thing more efficient
[20:36:58] <jrowe47> and with 3d printing, i can have arbitrary channels that would be a pain in the ass with just tubes
[20:37:13] <jrowe47> starting to see why these havent taken the world by storm yet
[20:38:57] <jrowe47> ~1.5% efficient conversion of electricity to linear motion, heat management is a pain, etc
[20:39:53] <anniepoo> yes, they def. have drawbacks
[20:40:54] <jrowe47> having a coolant / heating fluid system would definitely allow it to be far more efficient, but then the restriction is pumps and so forth
[20:41:04] <jrowe47> it almost imposes a biological type design paradigm
[20:42:42] <anniepoo> lets see... I put a cat in the middle of the robot, hooked up to a dialysis type loop,
[20:43:10] <jrowe47_away> haha
[20:43:29] <jrowe47_away> i was just thinking that a cat's muscle system would be a good place to start
[20:46:09] <anniepoo> https://www.youtube.com/watch?v=iMMGfzYXwAU
[20:46:27] <anniepoo> this machine's a pretty good illustration of how to make them
[20:46:46] <anniepoo> machine itself seems overkill unless you're making a LOT
[20:53:02] <anniepoo> he's using compressed air to cool
[20:53:19] <anniepoo> http://iskanderus.ru/wordpress/
[20:54:07] <anniepoo> he's making a wicked inMoov robot.
[20:55:49] <jrowe47_away> ty - gonna ponder on this for a bit, come up with a test case, something to play with
[22:56:32] <z64555> jrowe47_away: graphite coating, perhaps