Detection Algorithm Design And Performance Analysis For Spectrum Sharing Wireless Communication Systems

The past decades witness a rapid development of wireless communication technolo-gies. In order to satisfy the demands of higher data rate and better user experience,researchers propose many novel services, which requires more spectrum resources. How-ever, according to the existing spectrum allocation schemes, the available spectrum left for these new services is limited, which restricts the development of wireless communi-cations. As a result, spectrum sharing is suggested as a promising technology to solve the problem of spectrum resources scarcity.Traditional spectrum sharing mainly focus on cognitive radio communication. How-ever, with the development of 5th generation mobile communication and Internet of Things, spectrum sharing is applied in many new fields, such as ambient backscatter communication and mobile communication in high speed railway.The key idea of spectrum sharing is that two or more wireless communication sys-tems share the same band to complete their own information transmission in the identical time and location to improve the spectrum utilization efficiency. In the same space with more than one users, it is difficult to avoid all the interference and collision between them.Therefore it is important to make the users share the space harmoniously and effective-ly. One of the key factors is the sensing for the behavior of other users. Similarly, in a spectrum sharing system, the users also need to sense the spectrum and avoid causing in-terference to other users. Accurate and efficient spectrum sensing algorithms can improve the throughput of the whole system.This dissertation studies the detection algorithms and performance analysis of spectrum-sharing-based wireless communication systems. In terms of the spectrum shar-ing network with cognitive communication, the spectrum sensing algorithms are investi-gated and the corresponding throughput performance is provided. For the users in high speed trains, the ergodic capacity of time-selective and frequency-selective relay channel is studied. For the spectrum sharing network utilizing ambient backscatter technology,we develop a coding strategy and propose corresponding detection algorithms. The main contributions of this thesis are listed as follows:(1) Existing studies on spectrum sensing usually assume that the primary user (PU) has two states, i.e., engaged and idle. In practice, the PU may adopt power control to transmit with multiple transmit power levels. In such a scenario, we propose a soft-combination-based cooperative spectrum sensing algorithm to detect the power level used by the PU. Utilizing the PU's power level, we improve the the throughput per-formance of secondary users (SUs). We also investigate the connection between the detection ability of the SUs and the throughput of both the PU and the SU, and find that the throughput of the system can reach a balance without detecting all the power levels of the PU.(2) Spectrum sensing based on cyclostationary detection is considered for cognitive radio networks. We propose a cooperative spectrum sensing method for a cognitive radio networks with multiple terminals and one fusion center. It is shown that our proposed method can provide reliable performance even in low signal-to-noise ratio (SNR)region. We also utilize the signals received by the cooperative SUs so as to detect successfully without the knowledge of the PU's cyclic frequency.(3) We study the problem of ergodic capacity for the time-selective and frequency-selective wireless channels in the high-speed railway environment. The users in a high speed train connect to the access point (AP) through spectrum sharing, and then communicate with the base station via the antenna above the carriage. Firstly,we build up the mathematical model for this situation. Next both the upper bound and the lower bound for the ergodic channel capacity are derived on the basis of this model. Furthermore, the closed-form expression for the ergodic channel capacity is obtained at high SNR. These results can be utilized to evaluate the capacity per-formance of the spectrum sharing systems and provide a reference to the design of spectrum sharing networks in the high speed train.(4) The spectrum sharing network based on ambient backscatter technology is investi-gated. When the tag has two states to transmit symbols, we design the detector for the reader and evaluate the bit error rate (BER) performance. Specifically, we derive the optimal closed-form detection threshold and the corresponding BER expression.When the tag has three backscatter states, we propose a coding scheme which can minimize the BER, and design the corresponding detector. The BER performance of the proposed detector is also analyzed and an error floor is found in high SNR region.We also show that the tag with three states outperforms the one with two states in throughput.