Let's Make Robots!

The most advanced amateur humanoid robot project

Hi LMR!

This is the MAAHR project. MAAHR stands for Most Advanced Amateur Humanoid Robot. The goal is to design, build and program a robot which is similar sophisticated as ASIMO. The project is completely open source, everybody can participate. I'll sponsor MAAHR with 15000USD for hard-, software and logistic.

MAAHR's specifications so far:

  • 2DOF head with 2 cameras and LCD mouth
  • 2DOF main body
  • 2 x 5DOF robot arms
  • 2 x robot hand, capable to lift at least 1kg
  • 3 wheel or tracked base
  • Moving objects
  • Postures and gestures recognation
  • Speech recognation (and distinguishing sounds)
  • Synthesized voice
  • Recognizes the objects and terrain of its environment
  • Facial recognition
  • Internet connectivity

A first sketchup draft I did (arms still missing):

 

Comment viewing options

Select your preferred way to display the comments and click "Save settings" to activate your changes.

I don't know much about SBCs ... the chumby looks interesting, but probably because its coming closer to a standard laptop or netbook in capability.

As far as power, there is all the ACPI and other great things you can do with laptops - programmatically turn off the screen, wake on lan, heck even turn off the hardrive - all of this has been around for years... and is pretty tried and true.  While on the chumby it looks like they were going through some loops and spins getting Java to work.  

HDD is not necessary on a laptop, but its a great option to have if you want it.  You can boot off of SD or anything else from a laptop too.

Some of the design considerations depends on the size, etc.  To do "cool" things (at least in the ASIMO venue) it will need be large.  Large enough to do human things, lift objects, play a violin, open doors, etc..  That is why ASIMO is close to the same "formfactor" as a person.  These operations will consume lots of power, a laptop with the screen shut-off, fan on only when temp gets over a limit, and hardrives only on when needed - is not going to burn nearly as much as the motors and arms

Why go through the effort of moving code from a laptop to another device, if you don't half too.  I see it as just another step in an already complex/lengthy process.

A lot depends on what you want to do.  SBC might be better if size was the most important factor.  For instance, if this robot had to slip down some mine-shaft ventilation tube.  Now that would be "cool".  

With that said, maybe it would be good to determine a list of what "cool" things it should be capable of?  I'm a believer in "Form Follows Function".  So far, we have, "The most advanced amateur humanoid robot project could be my sponsor entry for 2011". 
It's an entry for the Sugar Shot competition?  

I'm not clear as to what it's suppose to do?  Markus?

 

There are 2 types of architectures for SBCs. (Look at Mini-box website.) The PC like architecture, where there is x86 processor on an all-in-one motherboard, and there is the new ARM based architecture, like in tablets and smart phones, this is what Chumby is. Now YOU need to determine the best architecture/speed/memory specs for the vision software. All other software can work on a Chumby or even on an Android phone (speach synthesis, mapping, soem A.I., perhaps even speech recognition) or run along the vision software. The only reason I'm against a full fledged laptop is cost. Sure, Markus is sponsoring the project, but he can't do that for all participants or whoever wants to replicate even partially the build. I surely want to do it, perhaps changing some stuff to comply with my limited budget and the already owned stuff that I have. So, a powerful laptop for vision may be in the 800-1200 dollars range, compared to just a motherboard, memory, PSU, storage drive which can be in the 200-300 dollars range. But, for the project, you can choose to use a laptop, that's fine with me, I'll replace it in mine with something as powerful as I can get and as cheap as possible. Will that work for you?

I am waiting for the specs on the vision computer from you Grog :D.

Thanks for the distinction..

I'm pretty inexperienced with ARM based architecture..  I was thinking in cost, but my situation is probably different, I have business associates who have "upgraded" their laptops with the latest version of MS Office. Once they realize their computer is 20 x slower due to this software behemoth, they get a new laptop and ask me to move all their stuff.  In the process I buy the "cast off" at a very large discount...  Sadly, there is a lot of churn in American business for laptops, so I thought I'd like to give them new life..  

Your link reminded me of how much I wanted to get one of those mini/micro/pico-atx boards...    I think they would be GREAT !

The ARM architecture might be great, but my lack of experience concerns me .. also threads like this for Chumby http://forum.chumby.com/viewtopic.php?id=177 

So I would suggest a X86 platform - specifically if its new hardware for a bot, a mini/micro/pico ATX board.  That way people who are interested in robotics but not willing to buy new hardware can play or tests parts of it on a regular computer/laptop.

Now which one?  I think they are all awesome, but I have rarely "shopped" for hardware....  But for comparison this is what I use when developing.  My OS sees it as a dual core "Intel(R) Pentium(R) 4 CPU 3.20GHz"  I have 2G memory, run MRL through eclipse debugger, and have a full blown virtual box VM running another development rig on windows simultaneously in addition to around 240 other processes.  Lots of memory is always good :)  

I'd trust Ro-Bot-X or TH to pick out something really awesome.

 

 

 

Running Linux on an ARM is no different than on an X86 from an user's perpective at least. Not sure why you are concerned about the architecture since Java runs on whatever device it has been ported on just the same. Of course there could be issues with drivers and stuff but let's face it X86 is just as vulnerable to thais kind of stuff: ATI videocards on Linux anyone?

However the major point here is harware scalability which on an X86 laptop/desktop world is simple, not so simple on SBC systems, take the memory upgrade for example, on a SBC it is mounted on board you cannot add more. 

Maybe the best choice would be some Intel Atom board or based on the newer AMD fusion, both X86 MiniITX platforms.

MiniITX platforms are SBCs too. SBC is Single Board Computer, and there are so many form factors and architectures out there. Any one of them will run Linux, but we should focus on one that will make things easy. I have an old Via EPIA 800 MiniITX board I could use for starters then move to something newer when I get more funds. I am still building 2 versions of the same robot, one big (using a MiniITX board) and one small (using the Chumby). I will have a lot to learn about Linux and the toolkits in the near future. As I said, building the hardware part is not impossible for me, the software is.

... Bill's awful creation they call an "OS" the software should run on everithing. I can create an Linux image optimized for our needs.

All we need is to settle on which "brain" we use, I would go for some ARM board, China has plenty of such things cheap too. ARM is better suited due to efficiency vs X86, that's what I would use.

I've updated the spec's on my blog, but just google ASIMO specification :)

It's no entry for Sugar Shot competition. This is MAAHR. I'll sponsor it with 15000USD.How simple is that? :)

I am not a complete believer of "design follows function". Most products would be so ugly, nobody would buy it even they work very good. MAAHR's outlooking should be human like (ok, human's have no wheels).

Beauty/Aesthetics can be a function... it's just one of the trickier ones, because it's so subjective

To me, making the hands dextrous and with many DOF, powerful enough to lift a beer is a challenge, but not impossible. The hard part to me, that seems impossible, is corelating stereo vision with IK to control the arms to pour beer in a glass! Or, to be more helpful around the house, to get some fruit from the refrigerator, put it on a plate and bring it to an elderly person. This is where I got stuck with my butler robot and then with MiniEric. I simply do not have the knowledge to do it myself and not enough time to learn how to to it. I hope this project will help me get closer to my goal...

Hands, seen at robotshop:

How cool would it be if robotshop be at our hand and donate two!

Total awesome but fu**ing $49900 a piece:

also seen at robotshop: http://www.robotshop.com/world/kinova-technology-jaco-robot-manipulator-academic-edition.html