B.S. level project on Intelligent Systems and Robotics

Hello!

I was wondering if anybody here could help me find a topic for a Thesis in my course Computer Science. I've only taken up Intro to AI, Assembly, and Logic Circuits in my previous subjects. Is it possible for me to actually make an Intelligent System in the span of about 6 months? What should I learn first?

Thanks in advance

Please Register or Log in to view the hidden image!

 

Log in or Sign up to hide all adverts.

I'm just looking up my "secrets of the game programming gurus" book abd there are several AI topics listed. The book has source code too, but let's face it, publishers never give away code that *actually* works, so I'll skip that.

Topics:

Random motion
Tracking algorithms
Anti-tracking: Evasion algorithms
Patterns with conditional logic processing
Elementary behaviour state machines
Personality
Memory and learning with software
Plans
Decision trees
Trial & error pathfinding
Contour tracing
Collision avoidance tracks
Waypoint pathfinding
Racing
Artificial neural networks
Genetic algorithms
Fuzzy logic (normal set theory & fuzzy set theory)
Fuzzy linguistic variables
Fuzzy manifolds and membership
Fuzzy associative matrices
Fuzzified inputs
Warm and fuzzy

Please Register or Log in to view the hidden image!

 

Log in or Sign up to hide all adverts.

Thanks for the info ^^

I searched for the book over the internet but the only one i found was 'Tricks of the 3D Game Programming Gurus" by Andre Lamothe. Is it the same book?

Is it safe to assume that game programming is in part AI programming?

 

Log in or Sign up to hide all adverts.

AI is a large part of game programming. Games wouldn't be fun with out some sort of antagonist working against you. Imagine tic-tac-toe witout an opponent.

 

That's not the same book. There are a few game AI books out there. Gamedev.net has a few listed.
Is this an undergraduate thesis?

 

Yes, this is for an undergraduate thesis.

Would it be possible for me to learn necessary concepts in computing and be able to make an intelligent system in the span of 6 months?

 

Correction, the title is "Tricks of the Windows Game Programming Gurus" by Andre Lamothe. It's all MS Visual C++ 6.0, and for all my efforts, I haven't been able to duplicate anything in the book. It's not that the code doesn't work, it's that the graphics engine is provided with the book and used as a "black box". The reason is that most games are console programs, and the console application that you can make with MS VC++ 6.0 doesn't support graphics, only text. So theoretically the author did us all a favor and provided a graphics console, ready made.

The code is exhaustive. Games are composed of nested loops, similar to multithreading, which allow many processes to be executed simultaneously (not simultaneous, but as far as the human eye can detect, it is simultaneous). Did I mention the code is exhausting? The book is over 1000 pages... suggest a handheld scanner and character recognition software if you have any hope of ever actually getting these programs typed into your computer. BTW, character recognition is a form of pattern recognition!

You don't do this whole book for a thesis. The author states that you should set aside two years for this subject. So I'd pick something that you can do for sure, or at least get a jump on, and make that your thesis. Don't do it because the terminology "sounds cool". For example "genetic algorithms".... whatever that means. Or some abstract method that claims to solve an np-complete problem. Guaranteed bullshit. Or it might not be, but still turn out to be a complete waste of time (polynomial-time?).