Introducing Spectral Analysis to Undergraduate Engineering Students

Currently, engineering students are only exposed to the theory of Fourier analysis in one of their math classes. They are not taught the relation between this transform and the frequency spectrum of the time domain data, how to find and plot its spectrum, or how to filter the data to remove unwanted noise and disturbance. Since a significant range of engineering applications require analysis of the measured data in the frequency domain, students will need to fill this gap between theory and practice without proper guidance. While MATLAB makes implementing these processes simple, only electrical engineering students who have taken a Digital Signal Processing course can understand and implement these processes. This paper presents a module for teaching spectral analysis to second-year engineering students using an engaging and hands-on approach without the intense level of math found in Digital Signal Processing (DSP) books. The module was applied in a core engineering course at Stevens Institute of Technology, which 400 students took from nine different engineering programs. The module consisted of three steps: research in which students were asked to report an application or process that uses spectral analysis. This started with a class discussion of the shared demo examples from each engineering discipline. In the second step, the students learned to use MATLAB to analyze music signals. The authors found music to be an invaluable illustrative tool for spectral analysis. It appeals to a wide range of students and is easy to generate. The analysis included understanding the concepts of the sampling frequency, single and multi-tone signals, and finding and plotting the frequency spectrum using the MATLAB FFT command. In the third step, students learned how to use MATLAB to design low pass, high pass, band pass, and band stop filters; then filtered the music signals to remove certain frequency bands. Finally, the students observed the effect of increasing the filter order on its performance.