Approximations for Performance Analysis in Wireless Communications and Applications to Reconfigurable Intelligent Surfaces
Tanash, Islam (2022)
Tanash, Islam
Tampere University
2022
Tieto- ja sähkötekniikan tohtoriohjelma - Doctoral Programme in Computing and Electrical Engineering
Informaatioteknologian ja viestinnän tiedekunta - Faculty of Information Technology and Communication Sciences
This publication is copyrighted. You may download, display and print it for Your own personal use. Commercial use is prohibited.
Väitöspäivä
2022-11-25
Julkaisun pysyvä osoite on
https://urn.fi/URN:ISBN:978-952-03-2614-2
https://urn.fi/URN:ISBN:978-952-03-2614-2
Tiivistelmä
In the last few decades, the field of wireless communications has witnessed significant technological advancements to meet the needs of today’s modern world. The rapidly emerging technologies, however, are becoming increasingly sophisticated, and the process of investigating their performance and assessing their applicability in the real world is becoming more challenging. That has aroused a relatively wide range of solutions in the literature to study the performance of the different communication systems or even draw new results that were difficult to obtain. These solutions include field measurements, computer simulations, and theoretical solutions such as alternative representations, approximations, or bounds of classic functions that commonly appear in performance analyses. Field measurements and computer simulations have significantly improved performance evaluation in communication theory. However, more advanced theoretical solutions can be further developed in order to avoid using the ex- pensive and time-consuming wireless communications measurements, replace the numerical simulations, which can sometimes be unreliable and suffer from failures in numerical evaluation, and achieve analytically simpler results with much higher accuracy levels than the existing theoretical ones.
To this end, this thesis firstly focuses on developing new approximations and bounds using unified approaches and algorithms that can efficiently and accurately guide researchers through the design of their adopted wireless systems and facilitate the conducted performance analyses in the various communication systems. Two performance measures are of primary interest in this study, namely the average error probability and the ergodic capacity, due to their valuable role in conducting a better understanding of the systems’ behavior and thus enabling systems engineers to quickly detect and resolve design issues that might arise. In particular, several parametric expressions of different analytical forms are developed to approximate or bound the Gaussian Q-function, which occurs in the error probability analysis. Additionally, any generic function of the Q-function is approximated or bounded using a tractable exponential expression. Moreover, a unified logarithmic expression is proposed to approximate or bound the capacity integrals that occur in the capacity analysis. A novel systematic methodology and a modified version of the classical Remez algorithm are developed to acquire optimal coefficients for the accompanying parametric approximation or bound in the minimax sense. Furthermore, the quasi-Newton algorithm is implemented to acquire optimal coefficients in terms of the total error. The average symbol error probability and ergodic capacity are evaluated for various applications using the developed tools.
Secondly, this thesis analyzes a couple of communication systems assisted with reconfigurable intelligent surfaces (RISs). RIS has been gaining significant attention lately due to its ability to control propagation environments. In particular, two communication systems are considered; one with a single RIS and correlated Rayleigh fading channels, and the other with multiple RISs and non-identical generic fading channels. Both systems are analyzed in terms of outage probability, average symbol error probability, and ergodic capacity, which are derived using the proposed tools. These performance measures reveal that better performance is achieved when assisting the communication system with RISs, increasing the number of reflecting elements equipped on the RISs, or locating the RISs nearer to either communication node.
In conclusion, the developed approximations and bounds, together with the optimized coefficients, provide more efficient tools than those available in the literature, with richer capabilities reflected by the more robust closed-form performance analysis, significant increase in accuracy levels, and considerable reduction in analytical complexity which in turns can offer more understanding into the systems’ behavior and the effect of the different parameters on their performance. Therefore, they are expected to lay the groundwork for the investigation of the latest communication technologies, such as RIS technology, whose performance has been studied for some system models in this thesis using the developed tools.
To this end, this thesis firstly focuses on developing new approximations and bounds using unified approaches and algorithms that can efficiently and accurately guide researchers through the design of their adopted wireless systems and facilitate the conducted performance analyses in the various communication systems. Two performance measures are of primary interest in this study, namely the average error probability and the ergodic capacity, due to their valuable role in conducting a better understanding of the systems’ behavior and thus enabling systems engineers to quickly detect and resolve design issues that might arise. In particular, several parametric expressions of different analytical forms are developed to approximate or bound the Gaussian Q-function, which occurs in the error probability analysis. Additionally, any generic function of the Q-function is approximated or bounded using a tractable exponential expression. Moreover, a unified logarithmic expression is proposed to approximate or bound the capacity integrals that occur in the capacity analysis. A novel systematic methodology and a modified version of the classical Remez algorithm are developed to acquire optimal coefficients for the accompanying parametric approximation or bound in the minimax sense. Furthermore, the quasi-Newton algorithm is implemented to acquire optimal coefficients in terms of the total error. The average symbol error probability and ergodic capacity are evaluated for various applications using the developed tools.
Secondly, this thesis analyzes a couple of communication systems assisted with reconfigurable intelligent surfaces (RISs). RIS has been gaining significant attention lately due to its ability to control propagation environments. In particular, two communication systems are considered; one with a single RIS and correlated Rayleigh fading channels, and the other with multiple RISs and non-identical generic fading channels. Both systems are analyzed in terms of outage probability, average symbol error probability, and ergodic capacity, which are derived using the proposed tools. These performance measures reveal that better performance is achieved when assisting the communication system with RISs, increasing the number of reflecting elements equipped on the RISs, or locating the RISs nearer to either communication node.
In conclusion, the developed approximations and bounds, together with the optimized coefficients, provide more efficient tools than those available in the literature, with richer capabilities reflected by the more robust closed-form performance analysis, significant increase in accuracy levels, and considerable reduction in analytical complexity which in turns can offer more understanding into the systems’ behavior and the effect of the different parameters on their performance. Therefore, they are expected to lay the groundwork for the investigation of the latest communication technologies, such as RIS technology, whose performance has been studied for some system models in this thesis using the developed tools.
Kokoelmat
- Väitöskirjat [4908]