Research On Distributed Passive MIMO Detection

With the application of sound absorbing materials and the reduction of ship noise radiation level,the detection performance of traditional passive sonar is limited in shallow ocean environ-ments,and active sonar will expose its own existence? what’s more,reverberation interference also brings challenges to target detection.Due to utilizing spatial diversity gain,Multiple Input Multiple Output(MIMO)develops dramatically and is widely applied,which shows great potential applica-tion prospects.This paper focuses on the use of cooperative sound sources to conduct distributed passive MIMO sonar processing research,and introduces Bernoulli filtering method to improve the detection capability of the system.Firstly,this paper analyses the MIMO channel capacity,and gives the constraint condition for achieving spatial diversity gain.Then,considering the multipath propagation resulting in the creasing of side lobe levels in conventional matched filtering outputs,adaptive Sparse Learning via Iterative Minimization is utilized to improve the estimation precision of the time-delay and Doppler parameters.Furthermore,based on the cooperative sound source,the zero trap method is used to suppress the direct wave interference.The time-delay Doppler factor is considered in the MIMO signal model,and the generalized likelihood ratio test is used to get the maximum likelihood estimation of the target position.Numerical simulation results have verified the effectiveness of the passive distributed MIMO processing method.Secondly,in order to improve the detection performance,distributed detection fusion method is studied when the target echo is regarded as the secondary sound source.With the local node’s detection performance,the detector is designed with a constant false alarm rate.Meanwhile,subop-timal detector is realized under the Bayesian framework.At-lake experimental results have shown that the detection capability of distributed fusion system has 3d B gain than that of single node detection under a constant false alarm probability.Finally,considering the false targets due to multiple transmission of MIMO detection,the Bernoulli filter is derived based on random finite sets modelling the target state and measurements.For multiple and maneuvering targets,the nonlinear tracking method is combined with the cardi-nality balanced multi-target multi-Bernoulli filter to avoid the computational complexity due to the data association in conventional multi-target tracking algorithms and reduce false alarm in MIMO detection.The numerical simulation results have verified the effectiveness of the Bernoulli filtering algorithm for tracking single target or multiple targets.The at-sea experimental results have shown that the cardinality balanced multi-Bernoulli filtering algorithm can improve the performance of MIMO detection.