I am teaching a course of signals and systems. We start from complex numbers and the end the course on Laplace transforms. In the course, we focus on continuous signals covering topics like LTI, Fourier (discrete and continuous), C.T. Fourier transforms and Laplace. The discrete portion is covered in the next course.
I am quite concerned about the lab work of the course. I have to use Matlab as a tool for the Lab. Can anyone give me some good suggestions so that my students understand the course very well?
Rabya
As you are using MATLAB for lab purpose of DSP, I would advise that don't allow your students to use the inbuilt MATLAB functions. This will defer their understanding. For eg: DFT algorithm, FFT (Radix 2 and 4) these things can easily be done using fft1 and ff2 functions in MATLAB. But students benefit nothing from these. So you need to suggest them that they make their own functions. Secondly the usual lab topics which we follow during teaching are:
1. Convolution in MATLAB, implementation and realization.
2. Correlation in MATLAB.
3. Using correlation to find convolution and vice versa.
4. FFT by radix 2 and 4.
5. Filter design (FIR and IIR).
In our lab course we use these topics using MATLAB.
We also use few practical hardware to realize DSP applications like FPGA and DSP starter boards.
Hope this helps.
As far as making students learn the concepts effectively - is concerned - labview software is good. I think its important to start with fundamentals such as different types of sampling, frequency and issues related to that, apart from the typical filtering applications
Oh that is strange. I have been using both for research and teaching for several years, I did find it to be very different in its purpose and use. If computation is what you are chasing go for matlab, but for signal accusation, processing and ploting labview does a better job. Anyway, since the question was more about how to make things better for students in terms of understanding given you have to use matlab, use the simulink part of it. More importantly, work on creating experiments that make them understand the fundamentals, the more time you can put into designing each of these question it would work better.
i was thinking to initially introduce basic signals and systems thru m files, thn move towards simulink and final simulink based TI's DSK C6713...
First, teach the fundamentals of MATLAB tools on Signal Processing.
I am a little puzzled, since to me MATLAB/OCTAVE (or LabView for that matter) is not working on continuos signals.... So, wouldn't it make sense - before using MATLAB - to cover the whole signal processing in the continuous/analog cases by just pencil and paper? Even z-transform for filter-design can be done on paper, can't it? And then, as soon as you introduced sampling, go for the digital stuff?
In that case, I can recommend the BeagleBoard system to do yo....
Here are a few ideas for lab sessions:
1. Comparison of DFT and FFT algorithms w.r.t number of operations and execution time. Compare results to analytically.
2. Effect of location of poles and zeros of Laplace and Z transfer functions on magnitude of frequency response. Deal with different kinds of filters.
3. Let them record their voice and import in Matlab. They can perform various operations such as convolution and correlation and hear the results back.
@Ulrich Hofmann
Yes i think you are very right. Actually, we are not just conducting lab rather its a Signals and Systems course with Lab. So first the students and introduced to the basic concepts and thn we use Matlab to help them in understanding topics that we have taught
Here is what i designed for Signals and Systems lab
Your feedback will be very beneficial to enhance the Lab work
P.S. pls. note the lab starts from basic Matlab Introduction and ends on Fourier Transforms, I have to add Laplace transforms to it...
Also from Z Transform to Filter designing is not included in this part, it will be part of the second course i.e. Discrete Signal Processing
i have conducted Signals Lab in last semester. Recording voice and importing it into Matlab is very interesting especially perform various operations such as convolution and correlation and hearing the results back. Analying shifted and scalled version in both time domain and frequency domian (using FFT ----------------------
)
- Badges
- Science topic
- Similar topics
- Engineering
- Electrical Engineering
- Signal Processing