Applications Of Filter Banks Based Multicarrier Techniques In Cognitive Radio System

Cognitive radio (CR) is the key technology which can solve the problems of low utilization of the spectrum band and shortage of spectrum resource. In a licensed band, the main issue is the protection of the primary user, this means that the primary task in any CR network is to dynamically explore the spectrum holes and determine spectrum band that can be used without causing interference to other users. Due to the flexibility of the spectrum usage in CR, a multicarrier transmission technology in physical layer is needed. OFDM is usually the first choice when talking about transmission technology in CR because of the anti-multi-path interference, simple channel equalization, anti-ISI/ICI by cyclic prefix (CP) and so on. However, the weaknesses of the OFDM technique, i.e., decreased bandwidth efficiency due to CP, high out-of-band emission, have been pointed out and the introduction of filter banks based multicarrier (FBMC) system has been advocated by a number of researchers. This paper studied the applications of FBMC in CR, including the spectrum sensing based on analysis filter banks, the FFT Pruning in FBMC transmitter and so on.Firstly, my paper briefly introduced the FBMC, including the filter banks, polyphase structure and the two basic components of FBMC:analysis filter banks and synthesis filter banks.Then we studied the spectrum sensing based on analysis filter banks. In order to reduce the computational complexity while improve the detection precision, we proposed multi-stage filter banks. Meanwhile, an adaptive threshold scheme in the power estimator was proposed, the threshold can be kept very close to the noise power, which can increase the detection probability.The secondly user usually can only use a small part of the spectrum band. This will cause the input points in the transmitter contain many zero points, which can involve many unnecessary computations. In order to solve this problem, this paper studied the applications of FFT Pruning in FBMC. We choose the Transform Decomposition (TD) as the FFT Pruning method in FBMC due to its efficiency and easy implementation. Based on TD, we proposed a Decimation TD algorithm (DTD) by dividing the input points into even and odd sets and carried out TD on each set. Simulation results show that DTD is more efficient than TD algorithm when the non-zero input points are almost in one set (even or odd).