Course program autumn 2011
Specification of course aims
8.0 Contents
 8.1 General
 8.2 The introductory task
 8.3 The question part
 8.4 The problem part
 8.5 Grading
 8.6 Possibility to improve (if you fail the introductory task).
 8.7 Laborations
8.1 General
One demand on examiners that is related to the Bologna process is that the course syllabus in the study guide (studiehandboken) must contain a few aims, and it must be clear how those aims are tested in the exam. All those aims must be tested at every examination opportunity. A consequence of this is that those aims are given a fairly vague formulation, and that those aims rather are section titles in a more traditional aim description.
For that reason, this page contains a specification of the aims in the study guide and how they are examined, grouped according to in what part of the exam they are treated.
8.2 The introductory task
This task tests basic knowledge and abilities, i.e. the following aim.
 The student should be able to reliably perform standard calculations regarding digital modulation, binary (linear) codes for error control and source coding.
One demand for passing the exam is that this task is treated completely correct. This task is actually two tasks, one from each of two out of the following three areas.

Digital modulation.
Simple error probability calculations for binary modulation forms, not necessarily standard choices, and only based on vector representations. So, this is primarily about distances between vectors. There will be no need for time domain descriptions. 
Binary codes for error control.
Primarily simple parameter calculations. For example: Given length, dimension and minimum distance of a code, determine its error correction or error detection capabilities Or: Given a matrix description of a linear code (generator or parity check matrices), determine the length and dimension of the code. You may be asked to determine the minimum distance of a code, but in that case only for very small codes. 
Source coding.
Determining a Huffman code for a small alphabet, or perform runlength encoding of an explicit binary sequence. You will not be asked to determine any quality measures of source codes here.
See also:
 Example of an introductory task.
 Possibility to improve (if you fail the introductory task, further down on this page).
8.3 The question part
Tasks number two and three examine the following aims, and they are tested by sample questions.

The student should
be generally aquainted with modern communication,
especially digital communication,
i.e. be able to briefly describe several communication techniques.
This deals with more or less everything that is treated in the compendium, but on a general level. This is not about detail, but about principle descriptions. E.g. being able to describe PSK. 
The student should
be able to briefly account for some common channel models,
especially for cables, radio channels and optical channels.
This is about accounting for things like thermal noise, linear time (in)variant filtering, multipath transmission, fading, primarily in general terms. 
The student should
be able to describe problems that arise in telecommunication
situations,
using own words, and be able to describe, and in a relevant way,
compare methods to counteract those problems.
The problem at hand is primarily noise. 
The student should
be able to account for the connection between different concepts
in the course in a structured way using adequate terminology.
This is about connections between noise and various quality measures, the geometrical interpretation of digital modulation and connections between parameters for codes for error control.
See previous exams for examples on what those tasks can be like.
These two tasks are given at most five points per task, i.e. totally at most ten points are available in this part. One demand for passing the exam is that you have gotten at least three points from this part.
8.4 The problem part
Tasks number four through seven examine the following aim, and it is tested by sample problems.

The student should
be able to, with some precision,
analyze and compare various choices of digital modulation methods
and coding methods in terms of error probabilities,
minimum distances and related concepts.
Those are traditional exam tasks. You should be able to analytically solve given problems from the parts of the course that we treat in problem classes, i.e. the following course parts, and the examination is based on samples from two or three of those parts.
Baseband representation of passband signals
Transformations between different representations of signals, and handling filtering and modulation in those representations. 
Digital modulation:
There can be both binary and nonbinary modulation schemes here, both standard choices and arbitrary modulation schemes. Typically you need to perform error probability calculations, average energy calculations of maximum energy calculations. It is possible that you are asked to compare two modulation schemes in these terms or to construct a modulation scheme that fulfills some demands. 
Codes for error control.
Primarily, we treat linear block codes here, but nonlinear codes can also be treated. Those codes are fairly small, and you typically need to analyze them in terms of their parameters, to determine the minimum distance, the weight distribution, the distance distribution, the error correction capability or the error detection capability, of the code. Most often, the starting point is a generator matrix or a parity check matrix, and it may happen that a code is created by modifying a code. Some concepts that you may be able to handle here ar cyclic codes, product codes, duality, syndrome and decoding. 
Source coding.
You should be able to create a tree code from given source statistics, using the Huffman algorithm. You should be able to analyze tree codes in terms of compression ratios and redundancy. You should be able to use Krafts inequality to determine if a tree code exists for a given collection of codeword lengths. You should also know and be able to use the concept source extension. Finally, there may be runlength encoding here.

Baseband representation of passband signals
See previous exams for examples on what those tasks can be like. Notice that problems where you are asked to calculate PSDs (Power Spectral Densities) or bandwidths are no longer relevant.
These four tasks are given at most five points per task, i.e. totally at most twenty points are available in this part. One demand for passing the exam is that you have gotten at least six points from this part.
8.5 Grading
To pass the exam, you first and foremost need to fulfil the basic demands on the three parts, i.e. correctly treated introductory task, at least three points from the question part and at least six points from the problem part. If all this is fulfilled, then the exam is graded according to the following grading limits based on the sum of the points obtained from the question part and the problem part.
 Grade 3 (ECTS C): 14 points.
 Grade 4 (ECTS B): 19 points.
 Grade 5 (ECTS A): 24 points.
Totally, you can get at most 30 points on the exam.
8.6 Possibility to improve
The rules regarding the introductory task, that it has to be completely correctly treated, as a partial demand for passing the exam, may seem a bit hard. For that reason we practice the following. If you miss one of the two parts of the introductory task, and if you have fulfilled all other demands for passing the exam, then you are given the opportunity to solve a new task that could have been part of the introductory task.
This is done at the examiners office at a time that you and the examiner agree upon, but it has to be done before the next exam opportunity.
8.7 Laborations
The laborations test the following aim.
 The student should be able to implement such communication systems that are treated in the course in block form and empirically evaluate them.
Sidansvarig:
Mikael Olofsson
Senast uppdaterad: 2019 07 30 10:17