Research Of The Technologies In OFDMA Based Cognitive System

Cognitive radio technology is a milestone in the history of the development of radio technology. It has been the key technology in solving the paradox between spectrum scarcity and spectrum underutilization. Cognitive radio enables the secondary users (SUs) to dynamically alter its power, frequency, modulation and other parameters to efficiently utilize the vacant spectrum without disrupting the operations of primary users (PUs), which effectively improves the spectrum efficiency. Orthogonal frequency division multiplexing (OFDM) technology is widely recognized as relatively easy to achieve efficient control and use of spectrum resource, which can make full utilization of spectrum resource through flexible spectrum combination or cutting. As an extension of OFDM technology, Orthogonal Frequency Division Modulation (OFDMA) technology provides a very flexible air interface mode for cognitive radio scenario. Therefore, cognitive OFDMA technology has been a hotspot in wireless communication research. For now. the research of OFDMA technology is relatively mature, but its application in cognitive radio still has some issues to be solved or improved. This paper is focused on the study of some key technologies of OFDMA based cognitive system, and proposed some innovative solutions to solve the problems in cognitive OFDMA systems.Firstly, the problem of carrier detection with inter-carrier energy emissions that inevitably exist in OFDMA-based coexisting system is studied. First we analyze the detection model considering the inter-carrier emissions in cognitive OFDMA system, and investigate the characteristic of inter-carrier energy emissions distribution. Then a novel carrier detection algorithm with self-cancellation of inter-carrier energy emissions is proposed based on a 3-order interference model. In our carrier detection algorithm, the carriers with high degree of confidence will be detected first by conservative thresholds, then the energy emissions of the detected adjacent carriers got by conservative pre-decision are gradually removed from the carrier to be detected. Our proposed algorithm effectively overcomes defect of tradition energy detection method that cannot distinguish the signal energy of PUs and interference, thus significantly reduces the false-alarm probability and provides more access opportunities while maintaining high detection probability.Next, the problem of interference cancellation in cognitive OFDMA system is studied for mitigating the interference introduced by SUs to PUs. After analyzing the model of interference to the PUs caused by the SUs'carrier frequency offset and the characteristics of interference coefficient, we propose an interference cancellation method based on the splitting of constellation chart. This method can effectively increase the Euclidean distance between the symbols on the adjacent carriers, so that the interference caused by the adjacent carriers would cancel each other in some way. Furthermore, the proposed method can also increase the Carrier-Interference Radio (CIR) of the SUs, which compensates the loss of Signal-to-Noise Radio (SNR) caused by the extension of the constellation chart, and improves the performance of the SUs.Finally, the problem of power allocation with fairness constraints for cognitive OFDMA system is studied. The consideration of fairness constraints in our model is to prevent the "starve phenomenon" of some SUs with poor channel condition. Nevertheless, the fairness constraints make the power allocation for SUs more complicated. Our goal is to optimize the SUs'capacity with the constraints of total power, interference to the PUs, and the fairness requirement among the SUs. This is a typical model optimization problem with multi-constraints. The traditional mathematical methods, such as game theory, or Lagrangian dual decomposition methods, can be used to solve such problem, which, however, may have quite high complexity and so aren't conductive to engineering implementation. The highlights of our study in this section is that, instead of using some complicated mathematical method, we fully exploit the characteristic of the model itself, and get a effective method to solve this problem with appropriate approximate and simplified treatment. The proposed algorithm can effectively solve the problem of power allocation with fairness constraints in cognitive OFDMA system with lower complexity.