So... i like robots and want to learn more about them. I'm a EE/CPE and i focused on controls. So i know basic control theory, embedded systems, and microcontroller stuff (made a few projects on my own with PICs and C).

I'm interested in starting some sort of robot project with the following goals:
1. fun/hobby/interest
2. gain meaningful experience for a future job.
3. take over the world

I'd love to work for the dozens of awesome robotics companies in Boston (i've applied, crickets) but i don't have shit on those M.I.T boys with PhDs, etc. So maybe some real experience would help, but if not this is fun, too.

So i'm trying to figure out what to do.. anyone build a kit like these: http://www.trossenrobotics.com/robot-kits.aspx ? Some are simple and use easy GUI building block "programming" some get more complex. Some use Microsoft Robotics Dev Studio, some is just you and a C compiler.

So the first step is to decide on a platform. I don't want a simple build-your-own-toy thing. Yet i'd be lost and probably get frustrated trying to use bare C. There are C-based platforms like Arduino (http://www.arduino.cc) that might work...

Then i need a goal for the robot itself. I'm interested in biokinetics so maybe a quadropod or hexapod would be cool. But i think a lot of the kits like http://www.trossenrobotics.com/lynxmotion-phoenix-hexapod.aspx have built in inverse-kinetic algorithms to simplify programming.. i'm still not totally sure how these kits are actually programed.

There's the bioloid (http://www.trossenrobotics.com/bioloid-robot-kits.aspx) which is very flexible (bioped, quadraped, anything) and easily programmed. But it is sort of plugnplay in that it uses proprietary parts, servos, etc. but maybe i'm trying to run before i even crawl.

any thoughts?

3/14/2009 11:29:56 PM

i love robotics but don't have the background to pursue them myself, so i mostly follow a couple robotics blogs to keep up with what's going on in the research world. a coworker of mine designed the software for the http://www.usfirst.org/ competitions, which i made some minor contributions to recently. i'd love to at some point get involved in some control systems software development, but i'm not actively pursuing it

stuff like this makes me jizz in my pants


http://www.damienfrancois.be/blog/
http://neurodudes.com/
http://www.robots-dreams.com/
http://sanssapiens.blogspot.com/
http://www.sciencedaily.com/news/computers_math/robotics/

[Edited on March 14, 2009 at 11:43 PM. Reason : add to my topics]

3/14/2009 11:35:59 PM

damn thats cool.

good links, thanks

3/14/2009 11:57:41 PM

Dan, did you see iRobot located in Durham was hiring for some positions recently? I thought about you, but apparently not enough to let you know in case you didn't see it. One of them was a test job but it seemed from the description you'd be a little more involved in troubleshooting and design changing than just outright functional/fault testing.

[Edited on March 14, 2009 at 11:59 PM. Reason : This is Chance, btw]

Funny thing, this thread...I've been getting the itch to do a project of some sort myself, just can't figure out what I'd like to do. I've wavered on either an embedded something or a really bizarre dual rail power supply for a chip amplifier (like an LM3886, etc) for a car amplifier. Instead of the typical smps with the big torroid I'm thinking of 2 seperate dc/dc converters for each rail. Trying to squeeze a ton of amplifier power into a teeny package.




[Edited on March 15, 2009 at 12:03 AM. Reason : .]

3/14/2009 11:59:43 PM

heres a nice robotics blog by a state grad student

http://www.gorobotics.net/


http://www.ncsurobotics.org

3/15/2009 12:54:20 AM

^^yeh, I applied for that and the location in Boston. :/ Yeh that'd be cool. I need a project as well.

3/15/2009 1:29:27 AM

this thread is relevant to my interests

3/15/2009 4:25:22 AM

oh how I miss kilroy

3/15/2009 8:21:01 AM

http://www.triangleamateurrobotics.org/
https://techshopdurham.com/

[Edited on March 15, 2009 at 9:30 AM. Reason : ]

3/15/2009 9:29:14 AM

The TAR group linked ^ is a good group.

And HOLY SHIT. I wish I had one of those Tech Shop places down here in Charleston. That is so badass.

3/15/2009 10:15:32 AM

^^^ reminds me of
http://www.damonkohler.com/2007/09/olpc-telepresence-instructable.html
http://www.damonkohler.com/2007/11/internet-famous.html

3/15/2009 10:46:45 AM

is that that wavecom irobot thingie that txt messages you when it detects movement or something?

3/15/2009 7:39:08 PM

It's kind of funny..the guys who have installed all the robots where I work were recruited from a community/technical college..Not from a big name university. So, not sure why you got rejected. Maybe you're applying for the wrong thing?

Or maybe that's just not the level of job you're looking for. They hated a lot of things in how the robots were programmed but had no say in things..

3/15/2009 8:08:21 PM

wait.

do you mean.... that INSTALLERS are from TRADE SCHOOLS?

say it isn't true!



i want my installers to be PhD Candidates from MIT. i refuse to accept anything less.

3/15/2009 8:42:11 PM

i sorry joe

but it is true

i mean, heaven forbid the guys who have to actually deal with an explain the epic programming fuck ups aren't from MIT

[Edited on March 15, 2009 at 8:57 PM. Reason : a]

3/15/2009 8:57:04 PM

yeah, i was so stunned when i found out the guy who installed my muffler doesnt have a mechanical engineering degree. he actually went to a trade school. and the guy who installed the electrical wiring in my house... once again, just a trade school. i was really hoping for at least an MSEE.



wait.

what was your point? i think i missed it.

3/15/2009 9:12:12 PM

i was mainly just fucking with you

but there you go making it all intense and shit

[Edited on March 15, 2009 at 9:13 PM. Reason : WHOOPS SPELLING OMG]

3/15/2009 9:13:14 PM

yeah HOLY SHIT i can't believe the installers went to a community school

i too would have expected master degrees at the least

3/15/2009 9:17:18 PM

I think I won this round, guys.

3/15/2009 9:20:54 PM

Quote :

"It's kind of funny..the guys who have installed all the robots where I work were recruited from a community/technical college.."

oh that's definitely funny. i mean, not "ha ha" funny... but, you know, funny funny.

3/15/2009 9:29:13 PM

RAWR RAWR RAWR

3/15/2009 9:33:49 PM

RAWR RAWR RAWR KING DOUCHEBAG HERE

and my originals point was a. that there should be more communication between the installers and the programmers and b. that he could get a job doing installation to get experience and go up from there. both points are valid.

[Edited on March 15, 2009 at 9:36 PM. Reason : a]

3/15/2009 9:35:19 PM

VERY GOOD!!!

now, tell me, what was so hard about that?

see, you could have just made this point directly instead of pussing around with some passive indirect bullshittery assuming we can all can read your mind.

3/15/2009 9:45:09 PM

3/15/2009 9:49:12 PM

Quote :

"my originals point was a. that there should be more communication between the installers and the programmers and b. that he could get a job doing installation to get experience and go up from there. both points are valid. "

naw to me is sounded like you were calling him a loser since the neighborhood stoners can get jobs installing robots but he can't even get a callback.

3/15/2009 9:52:09 PM

which is what i thought he was saying, too.

which would be a retarded thing to say, because obviously an engineer isnt going to be applying for an installation/tech job, and would probably be regarded as suspiciously overqualified if he tried to do so.

I'd never recommend for hire any degreed engineer interviewing for a tech/assembler/operator position at my employer.

3/15/2009 10:07:06 PM

it was in fact, retarded

3/15/2009 10:53:11 PM

its like the kids who go into ME because they want to learn how to fix cars or the CPE's that want to learn how to put together PC's better

3/15/2009 11:35:59 PM

oh you fuckers.

I'm thinking about using a PIC dsPIC33p in this board:
http://www.hexapodrobot.com/store/product_info.php?cPath=25&products_id=40
http://www.hexapodrobot.com/store/product_info.php?cPath=25&products_id=41

These power this guy: http://www.hexapodrobot.com/store/product_info.php?products_id=30

So i'd get just the microcontroller and SMB (basically breakout sort of board for the uC), a programmer, and a couple servos and just dick around. theyre programmed in C so it should be the most universal and PICs are pretty standard and i've used the Microchip MPLAB IDE before. There is are also pretty good support forums out there for this particular hardware and the robots that use it.

[Edited on March 16, 2009 at 12:19 AM. Reason : cxc]

3/16/2009 12:10:06 AM

Are you a current student? It was linked above, but this group

http://www.ncsurobotics.org

has done some really cool shit. You work with a team for a robotics competition and the guys I talked to had gotten corporate sponsorship for their robot.



Dunno if you're profile on here is accurate.

3/16/2009 1:14:45 AM

i really want to play around with the Arduino board and see what it's like, i've heard good things

also, remember you can get free PIC microcontrollers for free from http://www.microchip.com samples

3/16/2009 9:31:51 AM

http://www.youtube.com/watch?v=4oMTKuzc__M
pretty damn cool

3/16/2009 9:36:04 AM

very cool

3/16/2009 9:39:59 AM

yeh, BigDog is the coolest thing i've ever seen in my life. I was at the M.I.T Museum a couple weeks ago and BostonDynamics had some displays of their old quadrapeds and how they evolved up to that. really cool.

BigMan157, i was looking hard at the Arduino stuff, too, but i think the Atmel ATmega168 in them is too slow (8bit 20mhz, 20 MIPs, low memory) for robotics. It'd be the perfect setup for a general purpose microcontroller, PID controller, etc.

The PIC dsPIC33s are 16bit, 40MIPs, 160mhz, have a shit ton of memory and have a built in DSP to handle PWM stuff (24 channels) so you don't need an external servo controller. This is whats necessary to do inverse kinetics for multiple servos, legs, etc.

But there is probably more community support for the Arduino, so it might be a great way to start. Its all open source and there is lots of good example code out there.

3/16/2009 11:26:55 AM

You might also check out the propeller.

I've enjoyed playing with it, its a neat platform with a lot of potential. What ^ listed has more power, but the propeller is fun to play with for sure.

http://www.parallax.com

3/16/2009 8:31:53 PM

Well i've been working on this a bit. Long term, i'll be making a hexapod and writing the code 100% myself. Short term, i have the servo control finished and im working on the IK engine. Next up is trajectory planning and then gait generation.

A found a great book on the subject, Introduction to Robotics, Mechanics, and Control.
http://www.amazon.com/Introduction-Robotics-Mechanics-Control-3rd/dp/0201543613/ref=pd_bbs_sr_1?ie=UTF8&s=books&qid=1240247280&sr=8-1

I ordered the 2nd edition used off Amazon for only $4.82. I also have an electronic copy in DJVU format if anyone is interested, shoot me a PM with your email.

It is an engineering textbook and a little mathy, but not too bad. It starts with defining your bot in 3D and setting up the frames (3D subspaces) for all links, how to translate frames using matrix algebra, goes over FK and IK, and trajectory planning. It is weak on code implementation examples though, its almost all theory.

[Edited on April 20, 2009 at 1:13 PM. Reason : a]

4/20/2009 1:12:11 PM

that's what ECE555 with Dr. Grant is all about, btw

4/20/2009 3:11:55 PM

Thats really interesting. I wish it was available through EOL. :/ Maybe i can use my project as course credit somehow.

4/20/2009 10:57:02 PM

it's surprising that it's not

when i did it i know for a fact it was being filmed/transmitted to asheville at the very least

4/21/2009 3:00:43 PM

^It was probably for the unc-a mechatronics degree people... I know my controls class (mae 469) was broadcasted to them.

On a side note I work with (program, troubleshoot, fix) a Fanuc M-6iB at my job.


random image of that model

4/21/2009 3:16:03 PM

I get to play with some ABB 580's

although everybody here who has used Fanuc robots hates ABB

4/21/2009 3:20:25 PM

awesome.

Here's my base to wrist transform matrix for a leg on this robot. I just calculated it.



Now i just need to make the tool frame and multiply it in and i'll have my forward kinematics all figured out.

4/21/2009 3:38:25 PM

I just discoverd Sage (http://www.sagemath.org) and am now using it to do my maths and 3D simulating.

I have my forward kinematics all done and working. sweeeet.

I'll start IK next week.

4/23/2009 1:12:56 PM

FK, IK, and trajectory all done and working. My single leg can now smoothly move to X,Y,Z coords. pretty sweet. Maybe ill make a vid of it later.

Next i can work on the serial communication and some command protocols while i wait for a job so i can afford the servos and parts to build the rest of the robot.

5/12/2009 3:57:30 PM

i hate serial communications, but it's a necessary evil

parallel 4 life

5/12/2009 5:38:18 PM

I want to create something that moves the shift linkage on a car.

needs to be two axis i think. (in , out) and swivel. how hard do you think it would be? 2 torquey stepper motors?



[Edited on May 12, 2009 at 7:31 PM. Reason : .]

5/12/2009 7:31:27 PM

Quote :

"is that that wavecom irobot thingie that txt messages you when it detects movement or something?"

oh, missed that at first. but to answer your question, yes!

that's the only sensor we really got integrated that was testable, although we had several others working in isolation.http://imanomnivore.com/robot/history.php

[Edited on May 12, 2009 at 8:51 PM. Reason : ,]

5/12/2009 8:50:53 PM

Quote :

"i hate serial communications, but it's a necessary evil parallel 4 life"

Spoken like a true software engineer

5/12/2009 10:29:10 PM

5/13/2009 1:58:27 PM

set em up

5/17/2009 11:04:59 PM