TSKS04 Digital Communication Continuation Course
You sign up for the labs in groups of about five students. This is done by contacting <Okänt användarnamn: saif> . In any case, I will discuss about the laborations in more detail during Lecture 5 on Jan 31. We will decide on the groups in Lecture 5.
Introduction to Simulink
This is not a part of the laboration. Most students should have done the laborations in the prerequisite course TSKS01 Digital Communication, where we used Simulink. If for some reason you have not done that, make sure that you are in a group where the others have.
The Lab Task
The lab task is the following. You get a partly unknown channel, and your task is to construct a communication system that communicates over that channel.
Information about the channel
- It is saturated. Apart from that, it is linear.
- It is time-invariant.
- It is bandlimited, but not ideally.
- It adds white Gaussian noise.
- It takes a time-discrete input, and the sampling frequency has to be 100 kHz. If not, you will be flooded with error messages.
Demands on the System
Your communication system must fulfill the following demands:
- The system must be non-trivial, in the sense that it is not simply a modulator-channel-demodulator system.
- The bit-error rate must not exceed 0.5% when communicating independent equally probable bits.
- You should strive to maximize the data-rate. 250 b/s should not be impossible.
Demands to pass the laborations:
- Your system fulfills the demands above.
- Your lab report (one per group) is handed in no later than 2013-03-15 together with a Simulink-model of your system that demonstrates that the bit-error rate demand is fulfilled.
- Your lab report is written in English and includes a description of your system. All relevand measurements should be accounted for, as well as the reasoning that resulted in your design.
Your report should be sent by email to Mikael Olofsson as a PDF.
- The noise is substantial and will cause trouble when measuring the properties of the channel. You will not be able to see the signal in the noise with your bare eyes. A good idea is to correlate in a suitable manner over long time.
- A Bernoulli source can be used to generate independent bits.
- There are no formal demands on the signal power. Note, however, that the saturation of the channel results in a power constraint.
- Look around in Simulink. Most of what you need might be already available in the Simulink Communication Blockset.
- First construct a working system. Then use whatever time you have left to improve it.
The following Simulink models are needed or useful. Open them in Matlab version 7.13 or higher. Note that the default version of matlab in your lab might be set to 7.7. Matlab 7.13 is also available, you just need to change your setting, so as to use matlab 7.13.
- TSKS04.mdl. Open this first. The files below refer to it.
- ProjektLib.mdl. Then open this. The file below refers to it.
- A trivial, but working system using your channel. In the interactive Matlab window, you need to define the variables Rb=100 and Spb=1000. Rb is data-rate, i.e. the sample-rate of the bernoulli source that generates the input. Spb is samples per bit, which tells us the number of samples that represents a signal interval. Note that Spb has to be an integer, while Rb can be fractional. The product of those have to be 100000. If you include error correction in your system, that must be adjusted based on the rate of the code.
The channel that you are supposed to use is the one that is used in the trivial example above. In case you encounter any difficulty, feel free to contact Mikael Olofsson .
Senast uppdaterad: 2019 07 30 10:17