This study proposes a model-driven deep network based on the linear alternating direction method of multipliers (L-ADMM), to solve the problem whereby the inverse synthetic aperture radar (ISAR) generates defocused images of targets exhibiting micro-motion with sparse aperture. The network unfolds the operation process of L-ADMM into a model-driven deep network, and automatically optimizes the parameters of the network through learning instead of manually adjusting the parameters, which can better obtain images. Analyses of data acquired through simulations and experimental measurements were used to compare the results of imaging obtained by L-ADMM-net with those of the range Doppler (R-D) algorithm, chirplet algorithm, and L-ADMM. The entropy of images obtained by L-ADMM-net was the lowest, and their image contrast and resolution were the highest. Moreover, L-ADMM-net can generate high-resolution images of targets exhibiting micro-motion with sparse aperture at a low signal-to-noise ratio (SNR), which verifies its robustness. It can also automatically update and adjust parameters more stably than L-ADMM. The proposed method significantly improves the resolution, robustness, and stability of images of targets exhibiting micro-motion in different situations compared with traditional methods, and can provide technical support for target recognition in the future.

Atmospheric refraction is a special meteorological phenomenon mainly caused by gas molecules and aerosol particles in the atmosphere, which can change the propagation direction of electromagnetic waves in the atmospheric environment. Atmospheric refractive index, an index to measure atmospheric refraction, is an important parameter for electromagnetic wave. Given that it is difficult to obtain the atmospheric refractive index of 100 meters (m)–3000°m over the ocean, this paper proposes an improved extreme gradient boosting (XGBoost) algorithm based on comprehensive learning particle swarm optimization (CLPSO) operator to obtain them. Finally, the mean absolute percentage error (MAPE) and root mean-squared error (RMSE) are used as evaluation criteria to compare the prediction results of improved XGBoost algorithm with backpropagation (BP) neural network and traditional XGBoost algorithm. The results show that the MAPE and RMSE of the improved XGBoost algorithm are 39% less than those of BP neural network and 32% less than those of the traditional XGBoost. Besides, the improved XGBoost algorithm has the strongest learning and generalization capability to calculate missing values of atmospheric refractive index among the three algorithms. The results of this paper provide a new method to obtain atmospheric refractive index, which will be of great reference significance to further study the atmospheric refraction.

Inthis paper, global position system high-resolution sounding data from 1998 to 2008 were used to statistically analyze the spatiotemporal distribution and determine the probability, thickness, and intensity of atmospheric ducts at 12 stations in Alaska. In addition, the singular value decomposition (SVD) was used to examine the relationship between the Arctic vortex and atmospheric ducts. The annual average probability of atmospheric ducts, primarily surface and elevation ducts, was approximately 30% in Alaska. The probability of elevation ducts was greater than that of surface ducts. The Arctic vortex area and intensity index of each subarea were significantly negatively correlated with the occurrence of atmospheric ducts. Thus, when the area of the Arctic vortex increased and the intensity index of each subarea strengthened, the probability of atmospheric ducts decreased and their characteristics weakened.

An evaporation duct is a type of atmospheric stratification that affects radio systems. Atmospheric duct prediction is helpful for radar detection. In this paper, we used the deep forest, which is different from a deep learning framework, to predict the atmospheric duct height. At the same time, the long short-term memory neural network and other machine learning algorithms, such as the logistic regression, random forest, Bayes, and support vector regression algorithms, were adopted to predict the evaporation duct height. The predicted results with filled and unfilled missing data show that an accurate prediction of the evaporation duct height can be achieved using the deep forest.

Evaporation duct is a kind of chaotic phenomenon over the ocean. In this paper, a new nonlinear prediction algorithm, the Darwinian evolutionary algorithm (DEA), is introduced to obtain the specific nonlinear formula P(·) of the chaotic phenomenon. Based on Darwinian natural selection and survival theory, the method first selects a suitable training set of samples, and then produces an initial population before going through an evolutionary process of selection, reproduction and mutation until the optimal individual is found. Finally, a specific expression for a nonlinear chaotic time series is obtained, which can realize the short-term prediction of evaporation duct height (EDH) quickly and accurately. After that, the DEA, the support vector regression (SVR), and the back propagation (BP) neural network were applied to predict the EDH which were formed over the ocean by using sounding data. After interpolation and smoothing of the original data, we selected the first 250 data as training samples and the last 115 data as test samples to test the effect of the EDA algorithm. The results showed that the root mean squared error (RMSE) for the DEA was about 7% less than that of the SVR and 10% less than that of BP neural network; the mean absolute percent error (MAPE) for the DEA was about 9% less than that of the SVR and 15% less than that of BP neural network. In addition, the DEA obtained, for the first time, a nonlinear expression for EDH, which provides an important reference for future research on the evaporation ducts.

The inverse synthetic aperture radar (ISAR) can obtain high-resolution ISAR image of moving aerial targets. However, the micro-Doppler (m-D) interference generated by target exhibiting micro-motion components can lead to locally defocused ISAR image. Moreover, the echoes received by the receiver of the radar may be missing due to the environment for the propagation of electromagnetic waves and the switching of the frequency band of the radar, where this leads to a sparse aperture. A sparse aperture can result in severe interference from sidelobe or grating lobe in the ISAR image. To solve these problems, this paper proposes a structured prior model based on the joint constraints of the nuclear norm, l ₁ norm, and l _2,1 norm. The nuclear norm and l ₁ norm are used to constrain the low rank and sparsity of the high-resolution range profile sequence of the target, while the l _2,1 norm is used to constrain the structured features of the ISAR image of the main body. Furthermore, we use the linear alternating direction method of multipliers to solve the model and obtain the final ISAR image of the main body. The proposed model based on the above constraints can not only eliminate interference from the sidelobe and grating lobe as well as m-D interference, but also retain the structured features and connectivity of the ISAR image of the main body with a sparse aperture. The performance of the proposed method is verified based on data from simulations and experimental measurements.