Izvorni znanstveni članak
https://doi.org/https://doi.org/10.24138/jcomss-2024-0059
Leveraging Outage Probability of a Weibull-faded Gamma-shadowed Channel with co-channel Interference for ChatGPT-Driven QoS Prediction
Miloš Popović
; Belgrade Business and Arts Academy of Applied Studies, Belgrade, Serbia
Nikola Samardžić
; Belgrade Business and Arts Academy of Applied Studies, Belgrade, Serbia
Suad Suljović
; Academy of technical vocational studies Belgrade, Serbia
Nenad Petrović
; University of Nis, Faculty of Faculty of Electronic Engineering, Nis, Serbia
*
Selena Vasić
; Metropolitan University Belgrade, Belgrade, Serbia
Vuk Vujović
; MB University in Belgrade, Belgrade, Serbia
* Dopisni autor.
Sažetak
In this paper, a model of a wireless receiver based on
macro diversity (MD) technique is considered. The system consists
of a macro diversity SC receiver (MD SC) that combines the
output signals from two micro diversity (mD SC) receivers with L
input branches. This approach is used to simultaneously reduce
the effects of Weibull short-term fading, Gamma long-term
fading, and co-channel interference (CCI) on system performance.
In this paper, closed-form expressions for the outage probability
(Pout) of the system for the ratio of the signal-to-interference (SIR)
at the outputs of the mD SC and the MD SC receiver, were
derived. The system Pout of the ratio the signal-to-interference at
the output of the MD SC receiver can further be used to calculate
the average fading duration (AFD) of the proposed system
configuration. Further, we explore the potential of using Large
Language Model (LLM)-based trending ChatGPT service to
estimate the degree of the Quality of Service (QoS), considering
Pout as one of the input variables. Finally, we have compared the
proposed method to traditional Weka-based machine learning
algorithm in predicting the QoS.
Ključne riječi
Outage Probability (Pout); Selection Combining (SC); Weibull short-term fading; Gamma long-term fading; Gamma co-channel interference (CCI); ChatGPT; Weka
Hrčak ID:
321424
URI
Datum izdavanja:
23.8.2024.
Posjeta: 0 *