Signal Processing - May 2017 - 73

(a)

(b)

Figure 1. Different activities from the 2012 Engineering Summer Camp for high school students: (a) understanding frequency with speech and music
and (b) line-following robot.

The students also develop a method for
solving the multilateration problem
with noisy measurements using the
computer. This can be accomplished
either in Excel or in a programming
language such as MATLAB, depending
on the students' background.

Understanding frequency
with speech and music
This is a series of experiments in which
students are led to an understanding of
the importance of the concept of frequency in everyday signals, mainly in
speech and music. The experiments are
performed in real time on dedicated
digital signal processing (DSP) chips,
using a visual programming environment. Audio clips and the students' own
voices are taken as inputs via microphones, and loudspeakers are used as
the main outputs. In addition, preset oscilloscopes are used to obtain a realtime visual concept of the outputs. The
experiments enable the students to create
sound effects on their own, in addition to
performing assigned tasks. Figure 1(a)
shows students in the DSP laboratory
working on various experiments.

Experiment 1-Effects of suppressing and
removing frequencies from a signal
Students first synthesize and play sinusoids of various frequencies and change
frequencies while listening to the outputs.

They then synthesize and play sums of
sinusoids, both harmonically related and
nonharmonic, and change relative amplitudes while listening to the outputs. By
inputting an audio signal (music or the
student's voice) to the DSP board and listening to the effects of preprogrammed
filters (high pass, bandpass, and low
pass) on the signal, they observe how
these effects change as the cut-off frequencies are altered. The same exercise is
repeated by inputting a recorded audio
signal corrupted by noise and using a
low-pass filter to lower the noise audibility. Finally, some initial concepts related
to Fourier manipulation of signals is
introduced, and students synthesize or
input a square-wave and listen to the output. Then they pass the square-wave
through a narrow-band, bandpass filter,
and vary the center frequency to identify
the sinusoidal components. The outputs
are observed both audibly and on the
oscilloscope.

Experiment 2-Effects of shifting
and scrambling frequencies
A preprogrammed frequency shifter is
used to shift the frequencies of voice
inputs in both directions (up and down)
by up to an octave to demonstrate the
effects of pitch changes. The frequency
shifter demonstrates the limitations of
sampling by continually raising the output frequency until aliasing converts it
IEEE Signal Processing Magazine

May 2017

into a low frequency. Also, the frequency shifter is set to half of the sampling
frequency, which results in spectral
inversion (high frequencies are converted to low and vice versa). The result is
a simple voice scrambler, which is tested on the students' voices. Finally, a
more complex voice scrambler, based
on multiband spectral shifting and
inversion, is demonstrated and again
tested with the students' voices.

Line-following robot
Students learn concepts related to a linefollowing robot, a mobile machine that
automatically follows a specified path
without the need for human steering [Figure 1(b)]. Such a machine has various
applications in areas such as industrial
automation, warehousing, and automatic
guided vehicles on roads of the future. A
line-following robot has three main
components: a sensing system, a drive
system, and a microcontroller. The sensing system is responsible for determining the position of the robot with respect
to the line it has to follow; the drive system generates the motion of the robot;
and the microcontroller runs the control
algorithm that controls the speed and
direction of the robot along the specified
line. Students build, program, and test a
line-following robot. The sensing system consists of six reflective optical sensors. These sensors have a light-emitting
73

Table of Contents for the Digital Edition of Signal Processing - May 2017