Algoritmi cu cooperare pentru sisteme radio cognitiv / Cooperative algorithms for cognitive radio systems

This thesis addresses first the spectrum sensing methods used in relation to the technology known as cognitive radio (CR) to solve the issues of spectrum scarcity, which affects today's wireless services. The licensed spectrum, which is temporarily left unused by the licensed or primary users (PU), is made available to smart and adaptable secondary users (SUs) through this framework. By means of CR technology, these unlicensed SUs can take advantage of any spectrum utilization opportunity by dynamically adjusting their transmission characteristics. In the first part of the thesis, the conventional spectrum sensing methods are presented and compared with each other in terms of sensing sensitivity, complexity of implementation, and computational demands. In order to improve the sensing performance, the complexity of the spectrum sensing algorithm must be moderately increased either by optimizing the decision threshold value with respect to a performance meter, in a single SU scenario, or by gathering signal detection information from more SUs in a cooperative manner. Cooperative spectrum sensing for CR systems involves more opportunistic SUs that help each other detect the presence of the signal transmitted by a PU. The non-centralized Amplify and Forward (AF) technique assumes that SUs relay the received signals without additional processing and the final SU will take the decision. Another cooperative spectrum sensing solution consists in a centralized configuration where SUs send the detection/decision information to a common CR node, named fusion center (FC), which processes the signals and take the decision using a simple voting rule. Non-cooperative single SU solutions must gather additional sensing information either from different adjacent sensing time slots or to operate using one adapted decision threshold or even more decision thresholds with the purpose of increasing the detection performance. Some adaptive single decision threshold and double- or even triple-threshold sensing algorithms are investigated, with the threshold values set to minimize the overall decision error probability (DEP). An additional function is proposed to adjust the test statistics by time averaging the received signal energy in several adjacent sensing slots. Also, combining cooperation with adaptive threshold solutions will provide even better sensing performance. Hence, a double-threshold spectrum sensing algorithm based on mean energy detection (MED) is modified by adding the third threshold, which adaptively performs the MED test. Finally, the FC node takes the individual detection information from the SUs and uses it to take the final decision. To find an efficient way to use the limited radio spectrum resources, it became necessary to design equipment with flexible RF front ends. The software defined radio (SDR) approach offered such characteristics, and several producers developed such platforms in recent years. Another spectrum utilization optimization method considers the interference level control for more users accessing the same spectrum bandwidth at the same time, which is known as the spread spectrum non-orthogonal multiple access (NOMA) problem. In this context, one key issue addressed in this thesis is the design of the set of spreading codes, which impacts the level of the multiple access interference. Many companies have proposed various NOMA architectures for 5G and 6G networks.