LAUSR

Dear editor, Array antennas, which are able to adaptively generate the mainlobe in the target direction and notches in the jamming directions, are usually employed by modern radars [1–3]. A notch is adaptively placed in the mainlobe direction when mainlobe jamming (MLJ) exists, which changes the shape of the mainlobe and decreases the antenna gain on the target echo. Hence, in this scenario, it makes detecting the target or estimating its direction more difficult, and the problem becomes more serious when MLJ is closer to the target. Multi-input-multi-output (MIMO) system with distributed antennas is an emerging technique for modern radar. The scattered echoes from identical target are uncorrelated [4] or partly correlated [5] on different antennas. In contrast, MLJ from identical transmitter is completely correlated on different antennas [6]. According to these correlation characteristics, the target echo can be identified even in the case where it comes from the same direction as MLJ [6]. However, when target echo is overlapped with MLJ, the correlation coefficient of the received signal could be much larger than that of the target echo. Thus, the overlapped signal could be denied as MLJ with large probability. In a practical scenario, MLJ could be generated by both airborne jammer and accompanying jammer. Moreover, target echo and MLJ could overlap in time domain. To deal with this scenario, we propose a new target detection and direction estimation method on distributed monopulse arrays (DMA). An adaptive MLJ cancelation (MLC) filter is designed with a condition that the steering vector of target echo on DMA is perturbed by the partly correlated random vector. This filter on DMA can filter out MLJ and reserve target echo even in the case that target echo and MLJ come from the same position. The monopulse ratio is maintained while simultaneously performing MLC with the identical filter on Σ, ∆e, and ∆a beams of DMA. Hence, the direction of target echo can be estimated under the monopulse principle. System model. Assume there are I mainlobe jammers and a target in the far field of DMA, as shown in Figure 1(a). The target is placed at T0, and the i-th jammer (0 < i 6 I) is placed at Ti. There are Ns monopulse arrays placed parallel to each other at positions Mn (n = 1, 2, . . . , Ns). For any two arrays with positions Mm and Mn, the system geometry relationship satisfies |MmMn| ≪ |MnTi|. Each array forms Σ beam, ∆e beam, and ∆a beam simultaneously, and the monopulse axis intersects the axes of other arrays at predicted target position P . Hence, the Σ (or ∆e, ∆a) beam gain of each array on Ti can be approximated to be identical, which is derived in Appendix A. The gain of monopulse beams are defined as gΣ(θei , θai), g∆e(θei , θai), and g∆a(θei , θai), respectively. The steering vector on DMA corresponding to position Ti can be expressed as