Research On Spectrum Sensing And Signal Reception Methods For CR Nodes Under Impulsive Noise

With the introduction and development of the concept of smart grid,vehicle networking,and the Internet of Things,the human society is advancing toward informatization and digitalization.The wireless sensor networks that can collect and transmit data has become an indispensable part of the intelligent networks with their widespread applications and increasing scales.As a result,the demand for quality of service(QoS)and bandwidth has increased dramatically.This situation has led to a growing shortage of spectrum resources in wireless communications,which has greatly limited the practical implementation of wireless sensor networks.Therefore,researchers have been gradually focusing on new communication network models in which spectrum resources can be used efficiently and intelligently.The cognitive radio(CR)technology,which has broad application prospects,aims to provide unauthorized devices with the ability to accurately perceive the spectrum usage of authorized users,and to access the spectrum of the authorized users to transmit data by opportunities.Consequently,the establishment of cognitive sensor networks in combination with cognitive radio technology has naturally become an effective solution to the problem of insufficient spectrum of wireless sensor networks.The birth of cognitive sensor networks undoubtedly guarantees the construction of digital cities,but there are usually human activities around the application of the networks.Frequent vehicle ignition,power line transmission,switching on and off,and other artificial electromagnetic pulses often make the environmental noise of cognitive sensor networks have pulse characteristics.As a result,the performance of communication equipment,which is studied and established based on the Additive White Gaussian Noise(AWGN)channel hypothesis,will be seriously degraded in this environment.Therefore,in this research,the traditional AWGN channel noise model is replaced by the more realistic Middleton Class A noise model to simulate the real environment with impulse interference.And the spectrum sensing and signal reception of cognitive sensor nodes under this impulse noise model is thoroughly examined from both perspectives of transmission and reception,aiming at finding reasonable spectrum sensing and signal reception methods and strategies to resist the influence of impulse noise,thus improving the efficiency and reliability of data transmission between nodes.In view of the fact that the spectrum needs to be opportunistically occupied for establishing the communication link in the cognitive sensor networks,the spectrum sensing problem under impulsive noise is studied in the present research.Firstly,the performance of traditional energy detection based spectrum sensing method under Middleton Class A Noise is analyzed,and the analytical expression of detection probability and false alarm probability under this model are derived.Noise parameter uncertainty model is established to analyze the influence of noise parameter uncertainty on the performance of energy detector,and it is subsequently proved that the main factor affecting the detection performance is the fluctuation of the noise power;secondly,a spectrum sensing method based on Fractional Lower Order Moments(FLOM)is proposed.Compared with the energy detector,the proposed method has a significant improvement in sensing performance.Then,to solve the problem that the noise power fluctuation still exists in the FLOM detector affecting the sensing performance,an NPE-FLOM sensing algorithm with noise power estimation is proposed.The analytic expressions of its detection probability and false alarm probability are given.Finally,the superior performance of the proposed methods is verified by simulation analysis.Besides,the purpose of the spectrum sensing by the nodes is to find the spectrum to transmit data,and the greater the throughput,the better it is.Allocating time slots in the frame for spectrum sensing must aim at maximizing node throughput.However,in order to ensure the adaptability to the environment,the sensor nodes are often powered by batteries.For the consideration of the lifetime of the nodes,most studies assume that the nodes have the ability of RF energy harvesting.The process of energy harvesting provides energy compensation for nodes but occupies time slots of the frame at the same time,which undoubtedly makes the time slot allocation problem more complicated.Therefore,for the problem that there is a tradeoff between energy harvesting duration,spectrum sensing duration and node throughput,the frame structure of cognitive sensor nodes is optimized based on the previously studied spectrum sensing methods to maximize node throughput in this research.Firstly,the time-switching frame structure model is established;secondly,according to whether the node has a fixed power supply,the energy supply of energy harvesting is divided into two modes: compensation supply and full supply.Then,the node average throughput models are established.The capacity of Middleton A noise channel is derived,and the average throughput expression of nodes in each mode are also derived;Thirdly,for how to allocate each functional time period,a joint optimization model is established to optimally allocate the functional time periods for maximizing the average throughput of nodes under the premise of fully protecting the communication of the authorized users.Finally,the rationality of the optimization method in this research is verified by simulation,and the effects of parameters are also analyzed,such as energy harvesting rate and signal-to-noise ratio of primary user signal on the allocation results and the maximum average throughput of nodes.Finally,from the perspective of receiving,the problem of signal reception of cognitive network under Middleton A noise is studied.Firstly,the performance of the traditional optimal receiver under impulse interference is analyzed and its BER expression is derived.It is concluded that the performance degradation of the traditional optimal receiver under Middleton A noise is serious;secondly,because the optimal receiver under Middleton A noise cannot be implemented,a signal detection method based on fractional lower order moments is proposed.The BER expression of the receiver is derived;thirdly,for the problem that the noise parameter uncertainty will cause the deterioration of the receiver performance,this research also proposes a signal detection method based on the fractional power,and gives its approximate expression of the BER.Finally,simulations verify the superior performance of the above two methods.The present research provides a reference to overcome the difficulties existing in the application process of cognitive radio from the perspective of specific implementation.