EE/Math electives for signal processing

[walk_w/o_aim]

Hey,

I'm an EE undergrad looking to tentatively specialize in signal processing. My relevant background includes a course in signals/systems, introductory feedback control (using Nise), analog communication systems (using Haykin), introductory probability, the usual Calc I-IV sequence, linear algebra, ODEs, complex variables and two terms of discrete math (using Grimaldi). I'm also tentatively considering taking an intro to PDEs course if possible.

For scheduling and early graduation purposes, I'm trying to decide my future courses in advance. I have my pick of 5 fourth-year EE electives. I've decided on the first three slots: intro to DSP, analog signal processing circuits, and multimedia (image/audio) processing. I'm undecided on the remaining two slots and have narrowed it down to the following:

High Frequency Electronics

Transmission lines and waveguides, microwave devices, traveling wave devices. An introduction to the theory of radiation, antennae and wave propagation, and microwave scattering theory. The design of complete communication systems incorporating microwave, optical and satellite channels.

Digital Communications

This course will cover the physical-layer design issues in digital communication systems. The major topics covered are: information measures and the notion of channel capacity; link budgets; digital modulation techniques, including the signal space concept and optimal detectors, error performance in noise, suboptimal detectors, pulse shaping, synchronization, and equalization; error control techniques such as block and conventional codes, as well as comparisons between FEC and ARQ.

Modern Control Systems

Analytical representation of the finite dimensional linear systems, analysis and design of linear feedback control systems based on the state space model, and state/output feedback. Topics include: review of the linear spaces and operators, mathematical modelling, state space representation and canonical forms, controllability, observability, realization of transfer function, and solution of the state equation. Applications include: stability concepts and definitions. Lyapunov's Direct Method, design of the state and output feedback control systems, eigenspectrum assignment, and state estimator design.

There's also a course in Biomedical Signal/Image Processing that I'm not really considering (I want to stay away from biomed as much as I can). All three of the above courses look fairly interesting to me, and so my question is, assuming I can only take two of the above courses, which combination would be the most useful for someone interested in signal processing?

Also, would any math courses beyond what I've listed above be useful at this point?

Thank you.


p.s. the list of upper-year electives is much longer, but most of them concentrate on digital design, VLSI design, microfabrication, CAD/CAM, robotics and the like.

 

[Poopsilon]

Make sure you know applied Fourier analysis, that is the theoretical core of signal processing, hopefully your signals and systems class covered it, if not you need to take a course that does.

 

[walk_w/o_aim]

Thanks, Poopsilon. My signals/systems and analog communications classes gave me quite a bit of practice with applied Fourier analysis of CT signals. I'll see if the DSP class covers Fourier analysis of DT signals; if not, I'll go and get a book to practice with.

 

[sweetpotato]

As a EE grad who specialized in signal processing, I think the first 2 choices are good, but I'm skeptical about the control systems class. This is because I took 2 semesters of control systems and now work in controls engineering. Almost everything they taught me in control systems class was useless (basically we learned how to draw block diagrams of open and closed-loop systems and look up Laplace transforms and integrals in the back of our textbooks). However, that is just my personal experience so take it for what it is.

Also, Poopsilon is right: you need to know Fourier and Laplace like the back of hand.

 

[walk_w/o_aim]

Thanks, sweepotato: I was mainly considering the control systems class mainly because a lot of the math used seems similar. I'll keep what you said in mind.

 

[Poopsilon]

Hopefully this information will be useful to you too walk_w/o_aim as I'm certainly not trying to hijack your thread =].

@sweetpotato: I'm approaching signal processing from the applied math side rather than the EE side and you say you specialized in signal processing as an EE grad, I'm wondering if you are now working in industry as a signal processing engineer or something similar, and if so maybe you could let us know what it's like, how much math you use, if it's more hands on or a lot of programming, if the pay is good etc. And anything else about it you would like to impart would be great, thanks.

 

[sweetpotato]

I did specialize in signal processing as an EE major, but am now working in a manufacturing environment as an electrical engineer/controls engineer (my title is Electrical Engineer but the work I do is mainly in controls). So I really have no first hand knowledge about working as a signal processing engineer.

Maybe I can add something useful about my experience with control systems. I probably failed to clarify enough why I'm somewhat skeptical about the control systems class being considered. My colleagues like to remind me that the "controls engineering" classes I took have almost nothing to do with real-life controls, which requires a lot of programming, PLCs, and mechanical aptitude/knowledge (remember, you are "controlling" mechanical parts, not just making cute block diagrams on paper!). So this experience has led me to the (perhaps biased) view that if a Control Systems class doesn't mention any of the above (PLCs, programming, mechatronics stuff, etc) it is of doubtful value.

 

[walk_w/o_aim]

That was definitely useful. It gives me more reason to avoid controls as a career given my non-existent mechanical aptitude :)