Discrete wavelet packet transform based efficient energy detector for cognitive radio systems

Abstract

Current spectrum management model, command-and-control, is creating a fundamental cause not to exploit the spectral efficiency of the radio spectrums in times and spaces. The FCC (Federal Communications Commission) announced a very interesting report that relates to this fact. It pointed out that more than 70% of radio spectrums are underutilized in certain times or geographic locations. This means that spectrum scarcity is not due to a fundamental lack of spectrum but wasteful static spectrum allocations. Hence, various spectrum sharing schemes which can raise spectrum utilization are actively investigated by many researchers. In this context, an emerging technology, Cognitive Radio (CR), has been come out to increase utilization of spectrum in the wireless communication field. CR recognizes its environment and learns from the environment to make a decision. In addition, it also adapts operating parameters such as transmit power and modulation scheme according to the variation in real time. Among its fundamental functions, the most important function is a spectrum sensing which requires precise accuracy and fast speed. There have been several possible candidates of spectrum sensing for CR such as energy detector, fine/feature detector and two-stage sensing architecture. Although the two-stage sensing architecture is the best compromise for CR, the conventional energy detector which acts as a coarse detector in two-stage sensing architecture has redundancy during its sensing processing. Therefore, we consider how to improve the performance of the conventional energy detector in two-stage sensing architecture. Discrete wavelet packet transform (DWPT) can separate a given frequency band (or scanning range) into uniform channels step by step. Using this property, this thesis devises the modified DWPT based energy detector that applies double threshold at each operation of DWPT. First of all, the proposed detector divides the interested frequency band into...