Multi-Level Joint Detection Algorithm In Distributed MIMO Radar

As the performance gain from spatial diversity and frequency diversity,distributed MIMO radar has become an important approach to solve modern warfare problems,like the expansion of battle area and the development of stealth technology.According to the format of addressing information which is transmitted to the fusion center by node radars,there are two sorts of joint detection algorithm of distributed MIMO radar currently: The joint detection algorithm on signal level,which applies to the case of fusing complete measurements with good communication,usually provides the best detection performance because of target information reserved in maximum.And the joint detection algorithm on data level,which applies to the case of fusing processed measurements with bandwidth constraint,greatly decreases network traffic at the expense of performance.However,the studies above only apply to the case that fusing the same form of information fusion with single communication conditions,which have the problems with simple local processing mode,incomplete information utilization,and great loss of system performance.Therefore,in consideration of the complexity and variability of the actual combat scene,this thesis studied the joint detection of distributed MIMO radar algorithm on multi-level for solving the problems,like spatial mismatch of multi-channel signals and difficult information fusion on the different levels.The algorithm effectively improved the detection performance of the system under system constraints.The main contents are as follows:(1)A multi-channel signal registration algorithm based on the irregular grid index is proposed.To solve the spatial mismatch problem for detection cells of multi-channel signal,an improved registration algorithm based on the traditional regular grid is proposed,which eliminating the grids storing redundant indexes.Besides,the proposed algorithm effectively reduces the number of grids and decreases the computation without affecting the registration accuracy.(2)A joint detection algorithm on data level with multi-bit quantization is proposed.Firstly,a GLRT detector with multi-bit quantization is designed for the scenario under limited communication bandwidth.Then,to overcome the computational problem caused by numerical analysis,the Rao detector,whose detection statistic has a closed form,is derived as an alternative without MLE in low SNR.Finally,the simulation shows that,the performance with 3-bit quantization approaches the cases without quantification,which demonstrating the superiority of quantization detector.(3)A joint detection algorithm on fusion level is proposed.As the equivalence of GLRT detection and Rao detector,a detection algorithm based on the information fusion of signal level and data level is derived for inhomogeneous communication background of nodes.And the theoretical asymptotic distribution is also provided to generalize actual detection performance.Then the design of quantization threshold is optimized based on PSO algorithm to further improve the detector design.Finally,the numerical results corroborate our theoretical analysis and demonstrate that,the proposed algorithm which is widely applicable to the complex and changing international environment,effectively improves the detection performance by fusing the target information of different levels adequately.