Research On Techniques Of Multi-Target Detection And Parameter Estimation In The Same Radar Beam

Target detection and estimation of the range,velocity,angle,and other parameters of the targets are the basic tasks of radar.Radar beams are directional.Therefore,using a single beam to observe a single target is one of the common work modes of a radar system.However,the increasing occurrence of multi-target scenarios poses a great challenge to the resource scheduling capability of radar systems.Since the time resource of radar systems is limited,changing the work mode from the serial mode that each beam observes a single target to the parallel mode that each beam observes multiple targets can increase the dwell time allocated to each target.On the one hand,the increase of the dwell time leads to an increase of the integration gain of the individual target,on the other hand,the increase of the dwell time implies a reduction of the observation interval of the individual targets,i.e.,an improvement of the real-timing of measurement.However,radar systems encounter new problems in target detection and range/velocity/angle estimation when using a single beam to observe multiple targets.Therefore,in this thesis,the following research works are carried out to address the problems of detecting and estimating parameters of multiple targets within the same beam.1.To address the problem of detection performance degradation caused by the range migration phenomenon,we propose a method for multiple high-speed targets detection,i.e.,the RVDT method.The proposed method utilizes the proposed range-velocity decoupling transform to eliminate the RM phenomenon of multiple targets simultaneously,therefore the energy of each target can be accumulated into the corresponding rangevelocity cell which improves the detection performance.Since the proposed method retains the resolution ability in range-velocity domain,it can detect multiple targets with different ranges or velocities.In addition,the proposed method involves only one complex multiplication and one 2-dimensional fast Fourier transform,therefore it is highly efficient.2.To address the problem that the amplitude comparison monopulse method cannot handle the cases that multiple targets locates at the same range cell,we propose a method based on RELAX and maximum likelihood estimation(MLE)for the estimation of multiple target angles,i.e.,the ACM-ML method.Benefiting from the statistical properties of MLE such as asymptotic unbiased,asymptotic validity,and asymptotic consistency,the proposed method has high accuracy in angle estimation.With the help of RELAX,the proposed method has a high-resolution performance and is able to estimate the angles of two targets located in the same range-velocity cell under some conditions.3.To address the problem of high computational complexity caused by the coupling between targets and the higher-order motion of maneuvering targets,we propose a method based on the slow-time-reversing process and MLE for the estimation of the range and motion parameters of multiple maneuvering targets,i.e.,the STR-MLE method.By using the slow-time-reversing process,the proposed method can reduce the number of motion parameters that need to be jointly estimated by half,which reduces the computational complexity significantly.By utilizing RELAX,the proposed method decouples a joint estimation problem of the range and motion parameters of multiple targets into a series of estimation problems of the range and motion parameters of a single target,which further reduces the computational complexity.4.To address the problem of low resolution and limited performance in dealing with dense multi-target cases of the conventional range-Doppler processing,we proposed a range-velocity map estimation method based on sparse Bayesian learning(SBL),i.e.,the SBL-RV method.Under the framework of SBL,the proposed method improves the resolution by adopting the sparsity of radar signal as a priori.Thus,the proposed method can resolve multiple targets with close range and velocity.Moreover,by integrating the range-velocity coupling terms into the redundant dictionary,the proposed extends the measurable range of velocity and eliminates the range bias in the range-Doppler processing.