Let's Make Robots!

The most advanced amateur humanoid robot project


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.

Hi Ro-Bo-X,

It should be a linear motor in my drawing not a pneumatic piston:)

The tracks in my drawing have a width of 5cm. The tracked base is 600mm long and 400mm wide. The height of the robot is 1.2m.


myrobotlab.jar myrobotlab-0001.0455 uploaded

Install instructions are here http://myrobotlab.org/download1 - including how to play with opencv

thirdparty and myrobotlab-0001.zip rarely will need to get updated 



Gareth just let me know he got the OpenCV Service running on his windows 7 laptop with a webcam !   WOOT !

The issue was copying the binaries in myrobotlab-0001/bin to system32 directory (you shouldn't have to do this, but I have not had time to sort out the issue)

Anyway, if anyone has a webcam they should be able to play with it too on windows or Linux.   I have a massive update to do, which has a lot of enhancements (and bug fixes :P)..  I'll repost with a link to download when I update it.

OK, so, I guess the hardware can be divided in 3 parts:

1. Body (driving base, torso...)

2. Arms (and hands)

3. Head.

For Grog to develop the vision software, he needs the head. So I guess we need to focus on that first, so he can start working on something. Later, when we have the arms done, we should mount them on the torso together with the head, so he can work on the IK to do something useful with the arms. He will need the height of the robot, the distance from the shoulders to the cameras, the range of the arms, etc. so they need to be determined before he can start working on the IK. Am I right Grog? Am I appointing you as the head of the software department :D as you are the vision guru on LMR (but others can help too!).

I think before we start with any hardware building we need the physical proportions. I'll start this weekend with a sketchup draft. I will start with cuboids and defining the pan/tilt points. But the first building can be the head that other(s) then can start already to work.

Good! Someone needs to determine the quality/model of cameras for the head so we can figure the phisical dimentions for the head. Some cameras with good resolution that work in Linux without too much trouble.

I suspect a good/fast mini-ATX would be the best - the smaller (micro & pico) I think you'd be paying extra for the smaller size.  A camera that works for Linux would be great (most do at this point, but I would stay away from a M$ brand).  The optics/resolution/frame rates are really awesome with modern webcams.

One thing I have had to contend with is Lag.  You can have a camera with a fast frame rate say 30 ms, but the frame grabber will take 800 ms to get it.  So the stream is fast once you get it... but the lag makes things complicated.  A USB 3 / Firewire webcam perhaps?  

mini-ATX is awesome, can be tough to power though, must look out for integrated supply (or use picoPSU), I would recomend the dual core Atom cpus with ION chipsets ( integrated CUDA capable graphics card )

higher end logitech cameras are a good start and they work with linux, lag can be a function of buffering and additional processing. maybe some tinkering with capture settings can reduce the delay to less than a second.

... require processing power, Beagle bords as brainZ?

Also there is the logistics issue, like where comes the final thing in one piece? 

Final thing mounted in China, Shanghai? Transport costs of parts I'll overtake. Programming thru internet? Pre programming? GroG suggested off the shelf PC(s) and a lots of micro controller for the lower processing stuff, main brain(s) wireless conntected to the internet.