Error Models
Background
A digital channel comprises the transmission chain from the transmitter to the receiver in the complex baseband. The output sequence of a digital channel is a sequence of a bit error data stream, a symbol error data stream or a packet error data stream. Models for digital channels, which are different from channel simulators emulating the fading behaviour of the received signals, are often called generative channel models or simply error models. These models are required for the development, parameter optimization, and test of protocols. The aim of this project is the development of error models for present and future mobile communication systems such as UMTS.
Planned Research
The basic idea of this project is the development of error models by using deterministic processes, which are based on the principle of Rice’s sum of sinusoids. The procedure runs as follows. In the first step, the error statistics of a complete transmission system (comprising the transmitter, the mobile radio channel, and the receiver), which is considered as the reference model in this project, has to be evaluated preferably by computer simulations. Then, the statistics with respect to the distribution of the length of error bursts and non-error bursts, the rates of error bursts and the correlation between error and non-error bursts can be obtained. In the second step, the transmission system with the transmitter, the mobile radio channel, and the receiver has to be replaced by a simple deterministic process followed by a threshold, which is the equivalent simulation model. During the simulation, the level of the deterministic process will be below and above the threshold from time to time depending on the value of the threshold as well as on the chosen parameters of the deterministic process. Now, if the level of the deterministic process is below (above) the threshold, then an error (error-free) sequence occurs at the output of the digital channel model. Thus, the error model is simply a deterministic process followed a threshold. Due to the deterministic nature of deterministic processes, the resulting error model is likewise a completely deterministic model. The error statistics of the proposed digital channel has to be fitted to the desired statistics by finding proper values for the threshold and the parameters of the deterministic process. The block diagram of error models is further illustrated in Fig. 1.
Collaboration – Within the frame of this project, we are co-operating with:
Ericsson, Grimstad, Norway
Ministry of Communications, Bucharest, Romania



