Nice! Can’t wait to see
Nice! Can’t wait to see it!
Unfortunately I got in a black hole lately and I had bad luck after bad luck. Last thing that happened was losing the new $140 Android phone I got from China. The guy that found it phoned my wife, but since she didn’t answer, he got time to study the phone and decided it was too good to return it. Oh well… That was my new to be Android robot controller… And this was just the icing on the cake…
Here a first teaser of my
Here a first teaser of my robot hand prototype…
Well done Markus…Looks very
Well done Markus…
Looks very clean and well crafted as per your regular workmanship.
Thanks GroG,What do you
Thanks GroG,
What do you think about the hand gesture thingy? Difficult to archive? Do you have requirements on the cameras for the head, maybe a link? Hope the chickens are doing well 
Chickens Are A’ Layin…
To archive the positions once they are created ? That would be trivial.
To recognize the gestures from a human… not so trivial… however, it begs of cvMatchTemplate . The templates would need to be loaded (trivial) … but the video stream would need to be processed, and the hand isolated (not trivial at all), then the results matched with cvMatchTemplate…
For the camera I still believe my previous comment, however, I have looked at some of the fire-wire cameras. I feel a bit nervous of their support on Linux systems… cutting edge can be a frustrating place to be without good driver support.
At this point I would suggest 2 X Logitech C600 webcams - It’s what I currently use. The only difficult part to deal with is the 600-800 ms lag. One nice thing about them, is they are cheap $42 USD AND they have a microphone. So you could have stereo hearing - and therefore use the mics for speech recognition AND - sound origin triangulation, not bad for $84.
I have 2 Mics on my computer - a desk mic and the one off the camera. The desk mic works better for voice recognition - I have not analyzed the reasons, but my guess is the one on the desk is more uni-directional and therefore filters out ambient noise.
OK, so use the desk mic for
OK, so use the desk mic for speech recognition and the 2 camera mics plug them in a comparator to determine the direction of the sound.
BTW, how did you make the fingers, Markus? What materials did you use?
I am using ABS sheet, 2mm
I am using ABS sheet, 2mm thick, to make the fingers:
Easy to cut with a carpet knife, easy to grind, easy to bend with a hot air gun, easy to glue with super glue, easy to make it more sturdy by using multiple layers, glued together:
Arms?
This sounds like an interesting project, JAX suggested that my TROBOT’s would make a good set of arms for this project (I just discovered this site). Here is the link to a couple of robotic arms I recently completed: https://www.robotshop.com/letsmakerobots/node/27200. I’m going build a new (stronger/more robust) version of it soon, which I think would be nice for this project.
hey marcus, I would like to
hey marcus,
I would like to particiapate but I have limited experience and knowledge and I do not exactly know how to help but if there is anything I could do, I will try and get it done
Brain?
I’m glad to see someone just defined a serious robotic project with clear aims and a project plan. Have you decided about the brain for the robot? To process stereo pair vision planed as the robot’s eyes as well as performing all the assigned tasks, you would require a powerful processing unit.
Yes, we have defined a
Yes, we have defined a brain: Intel-Dual-Core-Mini-ITX.
Interesting. Could be a
Interesting. Could be a possibility. Please make sure to post pictures
About Kinect and template matching
Definitelly Kinect is a good addition because it will offload a lot of the processing work for obstacle detection and avoidance from the processor.
About template matching, if scalling is not working with template matching maybe you could try to do a feature extraction and detection algorithm for wall sockets, something like it is done for face recognition (not detection).
A Intel?
Pretty good, Im suprised by that. Isnt that only part of it.
And how do you plan on programming it?
Maybe some one should update
Maybe some one should update some of the info that is displayed as “not decided”
Thanks Pat…
Updated.
Regarding your update on
Regarding your update on 2011.07.08, I like the idea of a Twitter feed. You can have limit the feed to those you approve, who would have access to log in and associate the object with the appropriate word for your Haar data.
Arms?
Is there any plans on arms yet?
Hi Joe
No, there isn’t. Markus might have some, but we do NEED ARMS ! I’m wrapping up finishing the brain’s software install, eyes and power supply unit. If you have ideas, designs, or materials for Arms… let us know ! Actually, the same thing goes for the base.
( Wow, look how quickly you were volunteered ! ) 
