Eigenvalue-Based Target Detection Method For MIMO Radar In Presence Of Noise Correlation

Since 2007,the research work of Multiple-Input Multiple-Output(MIMO)radar has been rising slowly in China.It is a radar system that is developing rapidly.It has a deep connection with MIMO communication.Compared with the traditional radar system,it has the characteristics of accurate target detection and high detection ability for stealth objects.It has a good development prospect in the military and civil fields.At present,many methods have been developed for MIMO radar target detection according to the different attributes of the target,but most of them need to know or estimate the scattering matrix of the target,and assume that the number of snapshots is far greater than the number of receiving and transmitting array elements,and then carry out the detection work under the above premise.Although to some extent,the performance has been improved and improved,but When the number of snapshots is close to or less than the number of transmitting and receiving elements,the traditional detection method is no longer practical.With the development of MIMO radar technology,the trend of large array and blind detection is emerging.To solve the above problems,this paper uses the free probability theory and Stieltjes transform to explore and study a blind detection method of MIMO Radar Target Based on the eigenvalue in the background of correlated noise,and establishes a large array bistatic MIMO radar model with the same number of receiving and transmitting elements as the number of snapshots.The large array in the background of Gaussian white noise is obtained by using the traditional detection method and M-P law Then,by using the free probability theory and the Stieltjes transform,using and analyzing the correlation matrix eigenvalues of the received signal and other information,and using the maximum and minimum eigenvalues as the experimental test statistics to determine,the new threshold of large array bistatic MIMO radar target detection under the background of correlated noise is calculated In the simulation experiment,the correct decision probability is compared when two thresholds are used to make the decision respectively,and the experimental results are made.In this paper,the random matrix theory is used as a tool to solve the problem of target detection performance degradation under the background of insufficient number of snapshots and correlated noise.The simulation results show that this method is suitable for blind detection in the case of unknown noise variance and target scattering matrix,and significantly improves the robustness of target detection in correlated noise environment.