| Distributed multiple input multiple output(MIMO)radar systems can observe targets in the surveillance area from different angles using the widely distributed nodes.Compared to traditional radar systems,the distributed MIMO radar may utilize the spatial diversity to mitigate performance degradation caused by target radar cross section(RCS)scintillation.Additionally,by employing the orthogonal transmitting waveforms,statistical independence between different transmit-receive channels is ensured,which may enhance performance in jamming suppression,target detection,and parameter estimation.Therefore,it has attracted increasing interest in research and its applications in modern radar signal processing.Target detection is one of the core tasks of all radar systems,which serves an important function in the practical application of most radar systems,and the target detection with high performance has always been the focus in the field of radar signal processing.As a pivotal technology in distributed MIMO radar systems,target detection aims to determine the presence of one target based on the multi-channel echoes.In order to enhance system performance,a distributed MIMO radar system often adopts the centralized signal processing approach.This not only requires the perfect alignment between targets echoes and the corresponding states in each channel,but also requires a lot of system resources to store or transmit the collected data.Consequently,the system may suffer considerable challenges in its applications.In this dissertation,the target detection problem of distributed MIMO radar system is investigated,where the theoretical analysis and methodological demonstration of the proposed methods for different application scenarios are as follows:1.To solve the resulting data puzzle that evaluates the various cells under test(CUTs)in the surveillance area resulting from the intertwined range cells across all transmitreceive channels,a multi-target detection approach for the distributed MIMO radar systems is studied.Sketchily,the approach divides the surveillance area into identically interlocking and analytically expressible grid cells and then selects the grid cells with the best fitting multichannel data to be equivalently regarded as the CUTs.Then,a generalized likelihood ratio test(GLRT)detector and its constant false alarm rate(CFAR)detection threshold are derived.Finally,a separate procedure is introduced to eliminate the “shadow targets”,which are false alarms occurring in the grid cells without a target while sharing range cells with the targets.2.To solve the detection problem for the asynchronous data reflected by the moving targets when transmitters work in a scanning mode,a spatial-temporal data match algorithm based on multiple target states hypotheses is introduced.A GLRT detector is then designed to traverse the potential target states and examine the matched data to test the corresponding states,which effectively overcomes the target energy defocusing problem caused by the interested asynchronous phenomenon.In addition,for detection on the data received by a distributed MIMO radar system suffering time synchronization errors,a joint robust detection and estimation framework is devised,solving the multi-channel target energy defocusing problem by extending the target alternative states during spatialtemporal-frequency data matching.An adaptive clustering algorithm for time delay and Doppler frequency shift,as well as the effective estimation algorithm with unknown errors,are also introduced to accurately estimate the target state.3.Addressing the challenge of target detection constrained by radar system resource,an approach adopting low-bit quantization is introduced.Detectors based on several detection criteria are developed,with their closed-form expressions for theoretical distribution being derived,which allows the design of CFAR detection thresholds to improve the system robustness.Besides,the optimal quantizer design is modeled as an optimization problem independent of target state and other time-varying parameters,which may significantly improve the detection performance.4.In order to address the scenario for detecting a moving target by a distributed MIMO radar system on the moving platforms,a low-bit quantization strategy of echo is adopted to overcome the transmission bandwidth limitation,and a GLRT detector is also designed to simultaneously determine the presence of the target and estimate its state,which achieves the comparable detection performance in a weak-signal environment.The performance,effectiveness,and robustness of the above algorithms have been demonstrated through simulated or experimental data.The results indicate that the proposed methods in this dissertation may well perform the target detection tasks of distributed MIMO radar systems in some typical scenarios. |
