RAG-Aided Approach to Network Experimentation: Case of Wireless Environment Channel Capacity Analysis under Fading and Co-Channel Interference Conditions with Beaulieu-Xie Distribution

In order to successfully model the wireless channel in the presence of fading, in this work we have taken into account a more recent distribution, called the Beaulieu-Xie (BX) distribution. The BX distribution is able to include several line-of-sight (LOS) and non-LOS (NLOS) components of useful signal, which fact surpasses the successes of well-known distributions such as Rayleigh, Rice, Nakagami-m, etc. BX distribution is described by three physical parameters. Because of this, it is characterized with features of some other fading distributions, including those already listed, which can be obtained from it by setting special values for these three parameters. Beside multipath fading, the co-channel interference (CCI) also disturbs the wireless channels. Using BX distribution, the influence of both, fading and CCI is described. To mitigate these impacts, the selection diversity combining is used. In the situation of an increasing number of users of wireless resources, the channel capacity is one of the most important performance of a wireless system. Therefore, for the described scenario, the channel capacity is determined and graphically displayed. After, the parameters influence is observed for the sake of a successful selection of parameters and better signal propagation conditions. In the second part of the paper, we propose a retrieval augmented generation (RAG) based leveraging large language model (LLM) approach to network design taking free from text (such as network specification, requirements and compliance documents) as input. The proposed approach is evaluated in scenario involving the case of the BX fading and CCI environment.