Research And Performance Analysis On The Spectrum Sensing Technologies In Cognitive Radio System

In face of rarer and rarer wireless spectrum resources, Cognitive Radio is proposed, in order to improve spectrum usage efficiency through reusing licensed channels. The key point of cognitive radio is that without disturbing primary user, the secondary user could find idle channels for their own data transmission through spectrum sensing. So spectrum sensing is one of key technologies in cognitive radio, which can prepare for subsequent spectrum access and data transmission. In traditional spectrum sensing, there is only one result to describe its condition: busy or idle. However, they cannot work well for gray space, which has been partially occupied by primary users. In evaluation of sensing method, with energy consumption increasing, the concept of green communication has drawn more and more attention. Correct sensing probability is always one of key targets. According to the definition of green communication, it is not worth costing too much time and energy to seek higher correct sensing probability.In this paper, we research a two-level spectrum sensing scheme, make analysis of its throughput and energy consumption and give a more complete evaluation and optimization model.We first propose a two-level spectrum sensing scheme, which is based on energy detection and multi-band joint sequential sensing, and aims to maximize system throughput. This method is specially designed for gray space. At first, SU senses the whole bandwidth spectrum using energy detection; then, the second level multi-band joint sequential sensing can be dynamically launched based on results from first level detection. Only when the spectrum is partially occupied and unoccupied ratio is big enough, the second level detection can be launched. The threshold, which is used to determine whether the second level detection can be launched, is obtained in order to maximize system throughput.In energy analysis, we propose a energy optimization solution for multi-band joint sequential sensing. There are two kinds of popular network system: distributed and centralized. In distributed network, multi-subchannels are simultaneously sensed by multi-users; in centralized network, multi-subchannels are sequentially sensed by a special SU. We propose energy optimization model using sampling rate and sense time in these two networks and obtain the optimal solutions. We use AD9051 as an example to test and verify the solutions. At last, we use energy optimization solution to optimize two-level spectrum sensing, which makes it not only have the maximal throughput, but also the minimal energy consumption.