views:

303

answers:

8

Hello,

I would like to get some advice about my graduation thesis. The professor has already gave me some basic guidelines and I have two weeks to decide.

My graduation thesis should be coding "something" for the NAO robot. The "something" should be coded using the Silverlight technology.

Now I just need to decide on what the "something" might be. Do you have any ideas what module or addition could be coded for the NAO robot in Silverlight? I would appreciate any ideas.

Another professor suggested that I could integrate Facebook with NAO robots in our laboratory. We have multiple NAO robots so he thought I could make them post what they are doing on Facebook every few hours, they could also become friends with other NAO robots on Facebook and comment on their statuses.

But I am not yet sure I want to do that. If you have any other ideas, please share them with me.

+7  A: 

Hmm, what about posting the video stream from his cameras on the web, and letting remote internet users direct where he should go/what he should do?

You could have remote users send him on a scavenger hunt. They tell him what to find, he wanders around the room looking for it. It would require some image-recognition code, as well as code to translate the input into an image to look for. Plus you'd have to write some auto-navigation code/search logic.

He's got speech capabilities, and can read from files, so uploading things for him to read aloud is a possibility... maybe a remote, interactive nanny? You can see what he's seeing and have him read/act out your responses.

He's a multimedia server, so you could try setting him up as an all-request DJ. Make it so he can serve smart phones, so that anyone in the room can request music.

Heck, make him chase the family dog and bark. He sounds like a lot of fun!


** Edit **

The October 2010 issue of Scientific American has an article on Ethics in robotics... the authors talk about programming the NAO for ethical decisions, such as dispensing medication to patients (how to handle the patient's refusal, etc.). Programming ethics into a robot would be VERY thesis-worthy!


** Edit **

If you don't have access to Scientific American from Slovakia, you can buy the article online here:
http://www.scientificamerican.com/article.cfm?id=robot-be-good

Another SA article on the NAO, this one's free to read:
http://www.scientificamerican.com/article.cfm?id=robocup-2010-world-cup-soccer

James B
Thanks for ideas. I will wait for other suggestions.
Richard Knop
+1 for make him chase the family dog and bark!
Adkins
+1  A: 

What would be really cool, but probably very difficult, is to reverse engineer an API for the Microsoft Kinect to work on a PC then map Kinect's full body motion tracking to the robot. You could have a mini-you that mirrors everything that you do.

But this would be really, really cool (in my opinion at least).

John Wang
+2  A: 

Why not create a thesis with a good business case as well so you can easily commercialize what you have built. Some examples might be

  1. Robot that takes food orders on command, and processes them on a queue list for the kitchen to cook. This can replace the restaurants POS as the robot can calculate and the order takers salary.
  2. You can also create a sentry robot that has facial recognition which sends a discrete alarm to call (you or police, depends on security assesment) if somebody not on his facial database gain entry to your property. This can also be a good way to stream video whats at the owners home. This can replace a lot home security systems or even augment it.
  3. You can aslo replace receptionists in big corporate offices, where it can do the same like calling extension based on the name the visitor mentioned, opening appropriate doors or activating lifts (if the doors and lifts are networked), can also get staff time in and out based on facial recognition.
Raymund
I'd really advise against facial recognition. I've had the chance to work with one serious commercial facial recognition product, and the results were quite awful. The technology has still quite a bit to mature.
Vilx-
+1 for the idea of making it commercial
James B
+2  A: 

make it dance using dance-step diagrams and an internal metronome

Steven A. Lowe
+2  A: 

The first idea that comes to mind is to make a real-life Lightbot. Or something like it. Bonus points for creating a flexible playfield with riseable tiles 'n stuff.

Vilx-
+2  A: 

Program it to walk visit a room completely, discovering all the obstacles and learning how to avoid them. That's already very hard.

Why on earth SilverLight is mandated, I can't fathom however.

smirkingman
+3  A: 

Program it to be an extension of Windows Live, that is have him respond to MSN Messenger and read the received text, and provide actions for any emoticons that may be present. He could be a visual representation of that or any chat program.

Give him a set of basic emotion behaviors (ie rubs eyes slumps shoulders for sad/crying') etc.

This could be an extension of 'accessibility'. Also note that there are already studies being done on his effects on childed with Autism (mimicking the robot)

http://www.physorg.com/news200229593.html

Sounds like a fun Thesis though.

halfevil
+1  A: 

Why not try to create an image recognition inside and let the bot travel on its own, ie. if you have a image database from your university back to the dorm then let the bot travel on its own from university to dorm.

Mondyak