Joint Detection and Delay-Doppler Estimation Algorithms for MIMO Radars

In this paper, we consider the joint detection and localization (in the delay-Doppler domain) of targets by using Multiple-Input Multiple-Output radars. Specifically, inspired by the fact that most of existing contributions do not encompass target energy spillover between adjacent matched filter samples, we firstly analyze all the energy configurations associated with the target position within the delay-Doppler cell under test. Then, at the design stage, we formulate the detection problem by considering all data samples that might contain target components. The resulting problem is solved by applying the generalized likelihood ratio test (along with the two-step modification) when possible or ad hoc approximations of it dictated by the mathematical intractability of specific energy configurations. In these last cases, we devise two estimation procedures grounded on either a cyclic optimization of the likelihood function or the minimization of a suitable upper bound. Remarkably, the resulting architectures are capable of estimating the target position within the cell under test in the delay-Doppler domain. Finally, the behavior of the proposed architectures in terms of detection and localization performance is assessed over synthetic data and in comparison with conventional detectors that do not consider the spillover of target energy. The examples clearly show the superiority of these new detectors over the considered competitors as well as their capability of returning reliable estimates of target delay-Doppler position.