Esearch On Spectrum Sensing Scheme In UAV Networks

With the widespread use of UAVs in military,agriculture,aerial photography and disaster relief,and the rapid development of their industries,the low-altitude airspace band of UAV communication networks is becoming more and more crowded,spectrum resource is becoming more precious,and the shortage of spectrum resource is becoming increasingly prominent.Therefore,this thesis mainly studies the spectrum sensing technologies in UAV networks,hoping to find spectrum holes for UAV networks through efficient sensing and alleviate the shortage of resources.This thesis considers two scenarios of homogeneous and heterogeneous networks,and a cluster-based distributed cooperative spectrum sensing scheme and a spectrum prediction-based spectrum sensing scheme are proposed to maximize the accuracy of spectrum sensing in UAV networks,discover potential spectrum reuse opportunities,and improve the spectrum utilization efficiency.First,the differences between UAV networks and traditional ground cognitive networks are considered,and a cluster-based distributed cooperative spectrum sensing scheme is proposed for homogeneous UAV networks:first,the max-min distance clustering algorithm is used to divide the UAV network into a few clusters,and then a two-step fusion scheme is proposed based on the clustering information to fuse the sensing data in the UAV network.The proposed scheme obtains globally convergent consensus information,reduces the fusion delay caused by traditional distributed data fusion,and improves the spectrum sensing accuracy of the UAV network.Secondly,for the problems of unequal computing capability of different nodes in the UAV network and different spectrum sensing methods adopted,this thesis studies the spectrum sensing scheme in the heterogeneous UAV network,which focuses on solving the problem of efficient fusion of data from multiple sensing methods.In this thesis,energy detection and cyclic stationary feature detection are used in the UAV network,and this chapter analyzes and deduces the performance of the three fusion methods of the two data within the cluster;then,a heterogeneous sensing data fusion method based on hidden Markov model is proposed:the cluster head adopts the cyclic stationary feature detection which has high detection accuracy,and the hidden Markov model is used for spectrum prediction,then,the spectrum prediction information is used as the primary user’s prior knowledge for fusion spectrum sensing,which improves the detection performance under the low signal-to-noise ratio condition.Simulation results show that,for the homogeneous and the heterogeneous UAV networks,the proposed spectrum sensing scheme can obtain a higher detection probability.Compared with the traditional method,the detection performance of the proposed approaches has significantly improved.