Study Of MIMO Signal Separation And Recovery Using Blind Method Based On Antenna Array

As an important theoretical branch in the field of signal processing,blind signal separation/recovery gradually emerges its advantages in nowadays complex wireless communication system.It can bypass the complex real-time channel estimation process in modern MIMO communication systems,and can realize signal separation/recovery operation only by relying on the information contained in the received signal of the antenna array.A blind method does not need to send pilots or training sequences,which can save certain bandwidth resources,it has a good development prospect in communication systems with strict bandwidth or power constraints.This paper firstly introduces the classical conventional methods and blind methods in signal recovery.From the perspective of mathematical modeling,it introduces and analyzes methods such as matched filter(MF),zero forcing(ZF),minimum mean square error(MMSE),Sato,SG,Godard,CMA and so on.These methods provide directions for subsequent algorithm research and improvement.This paper will mainly focus on the theory and application of CMA.The main research content is listed below:(1)This paper uses the Analytical Constant Modulus Algorithm(ACMA)at the UAV receiver to achieve blind beamforming,aiming at the problem of poor beamforming performance caused by the rough MIMO channel estimation in the UAV assistance system.Furthermore,we use differential coding and decoding to cancel the phase ambiguity of ACMA.Relying on the CM property of the signals received by the antennas,the algorithm can effectively combine the target signal space with other spaces through a series of linear algebra transformations and operations such as matrixing,vectorization,diagonalization,singular value decomposition(SVD),and HT transformation.Through these operations the beamforming problem of signal recovery is transformed into a matrix decomposition problem,and the number of target users in the signal coverage area is estimated.We verify the algorithm performance for different channel,antenna,and noise environment simulations.The simulation results show that when considering the channel characteristics such as multipath effect and Doppler frequency shift,ACMA has good performance in the overall MSE and capacity of the system.The computational complexity is similar to that of traditional algorithms.(2)This paper proposes a new orthogonality-constrained analytical constant modulus algorithm(OC-ACMA)for blind signal recovery and blind equalization,aiming at the problem that the signal recovery convergence accuracy of ACMA algorithm needs to be improved and often has phase ambiguity.By exploiting the statistical quadrature condition and the CM features of the original signal,we derive an OC-ACMA receiver with a similar level of complexity as conventional ACMA.The simulation results show that the OC-ACMA algorithm has better performance than ACMA in signal recovery MSE and better robustness when SNR is low,which can achieve lower bit error rate.(3)This paper studies signal space separation for the occasions that signals partially overlap with each other.The mathematical models and properties of the subspace extraction methods such as GEVD,GSVD and SURV are studied,and the simulation verifications are carried out from the aspects of signal recovery waveform,signal recovery accuracy,running time,and subspace error.The simulation results show that all these methods can realize signal extraction for incompletely overlapping signals,and can separate the target signals from the interference signals.SURV algorithm has better performance in computational efficiency for continuously updated data packets than the other two algorithms.