Tips/recommendations for continuing my learning in DSP?

[tomizzo]

Hi there, this is a shameless repost of a thread I made over on the DSP subreddit, however, I'm very interested to get some feedback from the folks of PF. Here is my question:
--
Hi there,

I recently graduated with an undergraduate degree in electrical engineering. During my schooling, I took various courses on linear systems, control theory, and a single course dedicated to digital signal processing. I very much enjoyed the course, however, I took this course my final semester of my undergrad and did not have the capability of take more classes.

It had recently dawned on me that after the years of advancing my mathematics skills and learning about linear systems and controls, I've feel like I'm now at a point where I can truly appreciate the material. I'm at a point where I realize how little I know regarding the subject of DSP but have the capability to learn more on my own. As an example, I've recently taken up an interest in wavelet theory to analyze the time-frequency characteristics of signals.

However, I'm bummed now that I'm out of school. Besides the obvious answer of grad school, I'm wishing there was a next step in guiding how I learn more about the subject. As of currently, my further pursuit of learning the material is just a mess of jumping from one subject to the next.

My question: what's the best method for learning more about the subject of DSP from an intermediate to advanced level? It seems like the majority of the books I come across start at the basics of DSP and conclude at the discrete Fourier transform. And when I do find a book on a more advanced topic, the entire book is dedicated to the subject.

So what I'm looking for is some tips for further pursing the subject of DSP from a non-beginner perspective. I'm looking for book recommendations, compiled lists of topics that you consider 'advanced fundamentals'. If it's any help, my main area of interest is audio and sensor data (i.e. not image processing). Furthermore, I'm incredibly intrigued with the concept of time-frequency analysis hence my previous mentioning of wavelet transforms. The idea of signal decomposition, statistical analysis, adaptive filters, and signal coding are all very interesting also.

 

[jim hardy]

I'd look for employment someplace that specializes in DSP

TI or a radar outfit or military

 

[tomizzo]

I'd look for employment someplace that specializes in DSP

TI or a radar outfit or military

Ahhh that's the problem, I already have a job as a controls engineer. While I do quite a bit of work analyzing FFTs in the context of motion control and tracking, there are probably some areas of advanced DSP that I will never touch outside of being a hobbyist. This question hopes to fill that void.

 

[jim hardy]

i see .

Well DSP is one of those things i know exists and i do appreciate its power

i'd suggest join TI's E2E DSP community
and since it's a hobby, build fun gizmos
spectrum analyzer monitor for car engine would sell to amateur gearheads like myself.

a couple amateur radio links to peruse

http://members.inode.at/576265/rainradar.htm

http://www.rtl-sdr.com/ads-b-virtual-radar-rtl-sdr-tutorial-arrl/

the parts are inexpensive enough,
http://www.microchip.com/ParamChartSearch/Chart.aspx?branchID=8097
good luck !

 

[donpacino]

Ahhh that's the problem, I already have a job as a controls engineer. While I do quite a bit of work analyzing FFTs in the context of motion control and tracking, there are probably some areas of advanced DSP that I will never touch outside of being a hobbyist. This question hopes to fill that void.

I've been out of school working full time for about three years and I had a similar problem to yours.
I tackled it by getting my masters degree (I just finished). I contacted a few professors at various schools in the fields I was interested in and spoke to them about the research they are doing. I ended up doing advanced research in the field I was interested in.
Will your work pay for you to get an advanced degree? most will

Also when you learn to learn, school becomes almost trivial. It is really one person with knowledge trying to impart that knowledge onto other people. Once you get deep enough into a subject really the only way to learn is to do. If you want to go from intermediate to advanced in DSP then do it. Like Jim said pick a few projects and go for it.
Use your background in controls. Find projects that merge the two and build on your knowledge. Bonus points if you can make it applicable to work (convince your boss to let you work on a side project every friday for 4 hours).

For example.
Now that my degree is done, I'm going to start a side project. I have a background in controls and power. I've been interested in optimization lately.
I've been thinking about writing a better waypoint navigator and path-finding algorithm for my UAV, optimizing battery usage, time to target, and risk of collision. Its not a "nessisary" project, but I am doing it because its interesting and will push my level of understanding.

 

[donpacino]

you said wavelet transformation. are you interested in image processing. Buy yourself a camera and write your own processing software. Then use your controls background to have the camera track various targets. Add a motion detector and make the camera record at high resolution. Then have it take awesome nature videos in your back yard, automatically zoom into view the birds flying (or if you live in a city plastics bags flying).

again necessary, no. Fun, yes

 

[DrSteve]

Buy a development kit like the Blackfin and go crazy developing your own applications. That's what I did