Research On Joint Channel State Estimation And Spectrum Sensing Technique

With the rapid development of wireless communication, as a non-renewable resources, spectrum has become extremely scarce, which has brought great challenges to the future development of 5G communication. Cognitive Radio (CR) relieves this problem effectively on the basis of reusing and sharing spectrum resources. Spectrum sensing is a prerequisite of cognitive radio technology, due to that cognitive user can not interfere with the primary user (PU). Spectrum sensing technology means that the secondary user (SU) obtain occupancy information for a specific band through the signal detection and processing. With the continuous development of spectrum sensing technology, traditional sensing technology only to detect the presence or absence of PU is far from meeting the increasing various requirements.In the spectrum sensing process, some state information of the channel will seriously affect the spectrum sensing performance (such as the asynchronous work between PU and SU, the channel time-varying fading factor, the situation of multiple main users competing to access in some spectrum, etc.). Therefore, it is important to jointly estimate this channel state information in the process of spectrum sensing. In this paper, we mainly study the following two aspects of the channel state information in the spectrum sensing process.Firstly, a new asynchronous sensing algorithm is proposed to solve the problem that PU and SU can't achieved cooperative timing in the future heterogeneous wireless network. Different from the existing spectrum sensing scheme, this paper firstly considers the time difference between PU's transmitter and SU's receiver. In this part, firstly, a new state dynamic space model is proposed to describe the relationship among the observable energy, the state of the dynamically authorized users and the unknown time difference. Then, an iterative estimation scheme is designed by using the stochastic finite set and the maximum a posteriori probability criterion. Finally, the estimation results are obtained by particle filter through numerical approximation. The simulation results show that the proposed method can effectively eliminate the uncertainty of the received signal and improve the performance of spectrum sensing by obtaining the perceptual time difference accurately.Secondly, to solve the problem of competing access among multiple PUs in future cognitive networks, a multi-user spectrum sensing algorithm based on Underlay spectrum sharing is designed.Compared to Overlay sharing, underlay sharing mode can limit the interference degree of SU to PU. In this way SU can access the target channel at any time, which can greatly improve the utilization of frequency band. However, there are at least the following problems to achieve Underlay spectrum sharing strategy: First of all, there is no effective way to adjust the SU's transmit power dynamically to meet specific interference limits. Secondly, the existing spectrum sensing algorithm can only detect the possibility that a single PU is occupying channel frequency band. However in time-division multiplexing system,a number of primary users may exist in the same time slot. Different from the existing spectrum sensing schemes, the proposed multi-PU spectrum sensing algorithm can not only systematically perceive the possibility of using the channel frequency band by a plurality of main users in a communication system, but more importantly, it is possible to perform multiple PU-SU link states. Then the interference temperature of SU can be calculated from the accurate information of the channel,which can solve the above mentioned problems. In this part, firstly, a dynamic state space model is proposed for two main users based on TDMA. The main user state and channel fading factor are regarded as two hidden system states. Secondly, a recursive algorithm is proposed to obtain real-time fading channel and spectrum occupancy state based on Bernoulli. Finally, a competitive enhancement mechanism is proposed to correct the inaccurate estimation result by dynamically changing the prior density of estimation. Simulation results show that the proposed method can accurately estimate the presence of two authorized users and corresponding channel information, making it possible for Underlay spectrum sharing to be used in real world.This paper not only gives a spectrum sensing algorithm considering the above two different channel state information, but also compares the new algorithm with the existing legacy algorithms. The simulation results show the superiority of the new algorithm. In addition, the proposed algorithm has strong expansibility and can be easily extended to other application scenarios, e.g. the time offset information obtained in asynchronous sensing algorithm is important for the design and improvement of cognitive communication system. The proposed two new algorithms are very promising in the future dynamic spectrum sharing network.