Research On Signal Acquisition And Channel Parameter Estimation Of Cluster Unmanned Aerial Vehicles Based On Tensor

Cluster drones,as a product of high-tech in the information age,can improve the signal transmission rate,reliability,and communication transmission range of wireless communication.They have been applied in post disaster reconstruction,regional monitoring,and military communication.In cluster unmanned aerial vehicle communication systems,knowing accurate Channel State Information(CSI)is an important prerequisite for achieving power allocation,resource allocation,beamforming,and channel equalization.Therefore,accurately obtaining CSI plays an important role in ensuring communication quality and improving system reliability.Traditional estimation algorithms for time-varying channels face the challenges of high pilot overhead and difficulty in channel tracking.Tensor is a high-dimensional data representation method that can fully utilize the spatial,temporal,and other dimensional information of signals,giving it unique advantages in processing high-dimensional data.Therefore,tensor based channel estimation and symbol detection algorithms can achieve effective channel estimation while avoiding or reducing the use of pilot sequences in advance.Based on the above background,the main research content of this article is as follows:(1)This article considers a two hop cluster unmanned aerial vehicle time-varying system and proposes the Quick Start Nested Parallel Least Square(QSNP-LS)receiver algorithm.Firstly,the destination node constructs the received signal in a form that conforms to the nested parallel factor(PARAFAC)model,and obtains the iterative initial value in conjunction with singular value decomposition(SVD);then estimate the symbol matrix and channel matrix.The simulation results show that compared with the Two Stage Training(TST)method,the proposed algorithm has a more accurate estimation of the first hop channel.Compared with the Nested PARAFAC Alternating Least Squares(NPALS)receiver algorithm,the algorithm proposed in this paper has a lower Bit Error Rate(BER)and better performance.Considering the enormous advantages of millimeter wave communication,this article further analyzes the receiving algorithm of a two hop cluster unmanned aerial vehicle communication system in millimeter wave channels and proposes the Reprojection Nested Parallel Least Square(RPNP-LS)receiver algorithm.Firstly,the destination node constructs the received signal as a nested PARAFAC model and obtains the signal matrix through cyclic iteration;then,singular value projection(SVP)algorithm and iteration are used to obtain the channel matrix;finally,one-dimensional search is used to further obtain parameters such as the arrival angle and departure angle of the channel.The simulation results show that compared with the TST receiver,the algorithm proposed in this paper has more accurate channel estimation and smaller Normalized Mean Square Error(NMSE).In addition,this paper also analyzes the impact of the number of relay antennas on receiver performance.(2)To reduce the dimensionality of matrix operations,this paper considers a timevarying cluster unmanned aerial vehicle system with two hop millimeter waves.Based on compression sensing technology,a time-varying compression sensing nested Tucker(TVCSNT)receiver algorithm is proposed.Firstly,the destination node constructs the received signal in a form that conforms to the Tucker model;then,using compressed sensing technology to process sparse signals and reduce their matrix dimensions;finally,an iterative algorithm is used to estimate the symbol matrix and channel matrix.The simulation results show that compared with the TST and NPALS algorithms,the algorithm proposed in this paper has advantages in terms of channel estimation accuracy and signal BER,but its running time is longer.To further reduce the algorithm runtime,this paper proposes the Time Varying Compression Sensing Nested PARAFAC(TVCSNP)receiver algorithm.Compared with the TVCSNT receiver algorithm,this receiver algorithm constructs the received signal into a nested PARAFAC model,avoiding the use of kernel tensors;then,an iterative algorithm is used to obtain estimates of the signal matrix and symbol matrix.Simulation shows that the algorithm performs similarly to the TVCSNT algorithm in channel estimation and symbol detection,and has a shorter running time.