A new decoupling structure in H-plane for Vivaldi antenna is proposed in this article. The ultra-wideband (UWB) decoupling structure is shaped like an short-ended E, placed between two adjacent Vivaldi antennas along the H-plane. At the same time, a series of bowtie patches with overlapping edges is introduced, and a row of grid-shaped metal strips with equal height is loaded on the edge of the patches. All structure improvements lead to working bandwidth increasing, the total length reducing, and a better decoupling effect in the lower frequency. The simulation results show that the decoupling structure works well in 1.9-4.6 GHz, and the S21 between the two adjacent antennas in H-plane can be reduced at most 17 dB, with less influence on the radiation performance compared to the original antenna array.
For the pursuit of solving the electrically large scattering problems quickly and efficiently, an algorithm combining multilevel characteristic basis function method (MLCBFM) and compressive sensing (CS) is proposed for calculating the radar cross section of electrically large objects. According to MLCBFM, the surface of the object is first divided into large subdomains. Similarly, more levels of small subdomains are obtained by multilevel division, and the characteristic basis functions (CBFs) of each large subdomain can be obtained by those of small subdomains. Then, the obtained CBFs are taken as the sparse basis functions of compressive sensing theory, and the total impedance matrix which includes all large subdomains is evenly extracted as a measurement matrix. Finally, the coefficient of the unknown current can be obtained by implementing a recovery algorithm. The results of numerical simulations show that the algorithm of MLCBFM combined with CS has the same level of accuracy but higher efficiency than the traditional MLCBFM.
This paper presents a novel approach to detect failure elements in antenna arrays using a multi-branch artificial neural network (ANN) based on compressed sensing (CS) framework. The approach comprises a pre-trained decoder for quick far-field power pattern calculation and an encoder for recovering the sparse signals of the excitations of failed elements. The decoder trains on data from randomly impaired arrays to learn to calculate the sampled far-field patterns, while the encoder predicts the excitations of failed elements by inputting a small amount of phaseless measurement. Numerical examples demonstrate the effectiveness of the proposed method in terms of computational efficiency and diagnostic accuracy.
In this article, we present a balanced dual-channel filter based on dual-mode square dielectric resonator (DR) with a short end at the bottom. In this case, the resonator acquires a compact size, while it still retains a high unloaded quality factor. The electric fields of the modes are half cut like the shape of DR. According to the field distributions of the applied degenerate modes (half-TE011 and half-TE101), the pair of feeding probes are placed on both sides of the DR to realize differential excitation. Compared with the single-end feeding scheme, the higher order mode adjacent to the two used modes disappears after differential feeding. This property is beneficial to alleviate the adverse effect of higher order mode on the stopband suppression of the filter. The two orthogonal modes of half-TE011 and half-TE101 provide the internal isolation between the two channels. Based on the above analysis, a balanced dual-channel dielectric filter is designed with the center frequency of 1.52GHz and 3dB fractional bandwidth (FBW) of 2.6% for both channels. Because the structure is completely symmetric, the Chebyshev filtering responses of the two outputs are the same, and the number of size variables is significantly reduced. The results exhibit that the balanced dual-channel filter has the features of low loss, small size and wider upper stopband suppression.
This paper investigates the orthogonal subspace projection-based adaptive anti-interference technique for polarization-sensitive arrays. The mathematical model of the polarization-sensitive array adaptive anti-jamming and the orthogonal subspace projection algorithm are introduced first, followed by the derivation of the orthogonal subspace projection anti-interference algorithm. The simulation results demonstrate that the orthogonal subspace projection algorithm can be effectively applied in the polarization-sensitive array, which may and deepen the null depth by 15 to 25dB and considerably enhance the anti-jamming performance of the spatial and polarization filtering.
