Lab Exercises

Main content

Fachpraktika (Lab exercises)

Please use the online registration

Room

ETF B104

Date and Time

Monday, Wednesday and Friday: 13:15 - 16:15

Assistants

Federico Wadehn

Autumn Semester 2016

SV 1: Aktive Analoge Filter mit Operationsverstärkern (PDF)

Lineare analoge Filter können mit Widerständen, Kondensatoren und Operationsverstärkern realisiert werden. Im Versuch wird ein solches Filter berechnet und praktisch ausgemessen.

SV 2: Digitale Filter (PDF)

Lineare digitale Filter werden auf dem Computer entworfen und auf Audiosignale angewendet. Die Abtastrate wird konvertiert und das dabei entstehende Aliasing mit einem Filter verhindert.

SV 4: Egalisationsverfahren für Datenkommunikation (PDF)

Bei der drahtlosen und drahtgebundenen Datenübertragung wird ein Signal durch einen Kanal geschickt. Der Kanal verzerrt und verrauscht das Signal. Die Entzerrung der empfangenen Signale ist eine Aufgabe, für die es viele Lösungsmöglichkeiten gibt.

Im Versuch werden mehrere bekannte Egalisationsverfahren experimentell erprobt.

SV 7: K-means and Spectral Clustering (PDF)

When confronted with a huge amount of data, we are interested in quickly learning the structure of this data and clustering it by similarity. The goal is to find a smaller representation which still explains quite well our data. Clustering or the art of grouping similar objects together has a plethora of applications in various fields such as vector quantization, grouping proteins or density estimation. In this experiment, we focus on two different clustering algorithms: K-means, for clustering vectors and Spectral Clustering for grouping objects described by a graph. These two algorithms will be implemented in Python and applied to simple examples: clustering data points generated by Gaussian distributions, quantizing colored pixels of an image and clustering research areas. No prior knowledge on Python is required.  

SV 8: Continuous Phase Modulation (PDF)

Many digital modulation techniques result in a transmitted signal which - depending on the data to be transmitted - might change abruptly. This leads to detrimental spectral characteristics (e.g., poor spectral efficiency). In contrast, continuous phase modulation (CPM) modulates the data bits in a continuous manner and has therefore a high spectral efficiency. This is particularly interesting in wireless communication where bandwidth is expensive. In fact, CPM is most notably used in GSM.

In this experiment you will investigate a communication system that uses CPM modulation. With the aid of SIMULINK, you will simulate the modulation and decoding processes of CPM systems, and you will analyze key figures such as spectral characteristics of the modulated signal or probability of a decoding error.

Spring Semester 2017

SV 3: Switched-Capacitor Filter (PDF)

Zeitdiskrete analoge Filter können mit Kondensatoren und Schaltern (Transistoren) realisiert werden. Im Versuch wird ein solches Filter berechnet und praktisch ausgemessen.

SV 4: Egalisationsverfahren für Datenkommunikation (PDF)

Bei der drahtlosen und drahtgebundenen Datenübertragung wird ein Signal durch einen Kanal geschickt. Der Kanal verzerrt und verrauscht das Signal. Die Entzerrung der empfangenen Signale ist eine Aufgabe, für die es viele Lösungsmöglichkeiten gibt.

Im Versuch werden mehrere bekannte Egalisationsverfahren experimentell erprobt.

SV 5: Error Correcting Codes (PDF)

In this exercise, error correction codes are introduced. A few codes are demonstrated and their error correction capabilities are compared. A few simple examples describe the basic structure of a code. This is followed by more complex codes, and shows how a picture is completely recovered after being corrupted by errors.

SV 6: Polynomial Regression and Neural Networks (PDF)

In a test case, polynomial fitting is compared with regression using a neural network. The issue of overfitting is addressed for both methods.

SV 7: K-means and Spectral Clustering (PDF)

When confronted with a huge amount of data, we are interested in quickly learning the structure of this data and clustering it by similarity. The goal is to find a smaller representation which still explains quite well our data. Clustering or the art of grouping similar objects together has a plethora of applications in various fields such as vector quantization, grouping proteins or density estimation. In this experiment, we focus on two different clustering algorithms: K-means, for clustering vectors and Spectral Clustering for grouping objects described by a graph. These two algorithms will be implemented in Python and applied to simple examples: clustering data points generated by Gaussian distributions, quantizing colored pixels of an image and clustering research areas. No prior knowledge on Python is required.

 
 
Page URL: http://www.isi.ee.ethz.ch/teaching/lab.html
Sat Mar 25 00:40:34 CET 2017
© 2017 Eidgenössische Technische Hochschule Zürich