Ideas for an Independent Research Project in Signals

[WolfOfTheSteps]

I'm looking ahead to winter quarter 2008 when I will be starting an independent research project. I wanted to get an early start in thinking about what I might be able to do. I am thinking that I want to do something in signal processing, but don't have enough knowledge to make a choice. Here are the EE classes I will be taking this school year which I would like to be relavent to my project:

Fall Quarter Signals and Systems, Digital Signal Processing, Communications Systems

Winter Quarter Intro to Feedback Systems, Digital Communications, Digital Filtering

Spring Quarter Wireless Communications, ? (My adviser suggested that I perhaps take some graduate courses this quarter... Don't know what they would be yet.)

So I want my research project to be a 2 quarter project: Winter and Spring 2008. Also, for at least one quarter of my project I'd like to focus on some design aspect to get rid of my design requirement. Any ideas?

The only thing I can think of off the top of my head is doing something with image processing and perhaps designing some sort of program or something... Don't really know.

I don't have to choose until late in this Fall's quarter, so I basically just want some suggestions for the direction I might want to head in. I'd also like to get my feet wet now, so I'm not overwhelmed when the time to start comes.

Thanks!

PS. This is pretty tangential, but if anyone reading this is a systems expert, could you take a look at this question? It's been driving me mad for days! :)

 

[berkeman]

I think it would be fun to do a project involving CDMA. The concept of code division has always fascinated me. There are also some new spread spectrum techniques that are very computationally intensive -- I forget the name of the channel. I'll see if I can scare it up.

 

[WolfOfTheSteps]

Thanks, berkeman. I'll look into CDMA.

Also, are there any especially interesting signals fields that rely heavily on probability?

I've taken the EE departments probabilistic systems course, but I'm minoring in math so I'm going to take the 3-quarter probability and stochastic processes sequence this year. The more I can link all my courses together in my research project, the merrier. :)

 

[edmondng]

yea lots of topics. try classes like statistical signal processing, information theory and coding, detection and estimation (more control)

most projects in wireless are moving towards more complex algorithm and other modulation scheme. uwb, mimo, ofdm, mc-cdma, ds-cdma, application of wireless theory/modulation like sensor, detection, estimationo etc.

if you have a chance, get an internship and not only that work on research project with a professor for like a year. it'll put you ahead of everyone else.

 

[WolfOfTheSteps]

Thanks edmondng.

I looked into some of your suggestions, and I'm really fascinated by information theory. I just ordered a copy of Shannon's The Mathematical Theory of Communication out of pure curiousity.

But I'm just wondering... Where would the line between computer science and EE be drawn when it comes to information theory? I don't want to go off on a tangent that has nothing to do with my major as an EE! (Although I am somewhat interested in computer science)

Thanks again.

 

[berkeman]

But I'm just wondering... Where would the line between computer science and EE be drawn when it comes to information theory? I don't want to go off on a tangent that has nothing to do with my major as an EE! (Although I am somewhat interested in computer science)

Just keep exploring practical applications of information theory, and you'll be mostly in the EE domain. Like, how cell phone networks work at all levels (from the phones themselves to the cell phone towers, to the networking and handoffs and codes and multipath and channel noise issues... (great stuff). And like how satellite communication works in different ways, and how other practical channels work and why they are optimized the way they are...

 

[edmondng]

well with ee they will always be some programming, whether c, c++, java, matlab, assembly or using control software like spice for example (notice the difference between learning how to use software and writing one). unless in ee you went into a hardware control field like power, controls then you'll be dealing more with circuit and power design, FPGA, VLSI etc

If you have programming background, you'll end up somewhere with some hardware exp (not expert but enough to keep you knowledgeable when it comes to design) and writing software control/interface etc.

CS is more for like software development (eg: creating windows software) but then it is not necessary true sometimes.

And information theory is a lot to do with probability. In signal system, and wireless communication, a lot of random variables, probability, linear algebra etc involve. I've seen people work on thesis with calculating cost between service providers like how when they are roaming, and how to reduce it using different modulation scheme and method such as diversity making use of multipath fading and increasing SNR)

With image processing, you can go into designing vision system, imaging, sensors, but seems a lot more software oriented.