Handcraft features are commonly used for signal classification, which is a time-consuming feature engineering. In order to develop a general and robust feature learning method for radio signals, a novel One-dimensional Deep Attention Convolution Network (ODACN) is proposed to automatically extract discriminative features and classify various kinds of signals. First, one-dimensional (1-D) sparse filters are designed to learn hierarchical features of raw signals. Second, an attention layer is constructed to weight and assemble feature maps, to derive more context-relevant representation. By using simple 1-D filtering, ODACN is characteristic of less parameters and lower computation complexity than traditional Convolutional Neural Networks (CNNs). Moreover, feature attention can mimic a succession of partial glimpses of humans and focus on context parts of signals, thus helps in recognizing signals even at low Signal-to-Noise Ratio (SNR). Some experiments are taken to classify 31 kinds of signals with different modulation and channel coding types, and the results show that ODACN can achieve accurate classification of very similar signals, without any prior knowledge and manual operation.

A finite-time extended, state-observer-based, fixed-time backstepping control algorithm was designed for hypersonic flight vehicles. To enhance the robustness of the controller, two novel finite-time extended state observers were introduced to compensate for the negative effects of lumped disturbances such as uncertainties and external disturbances. Two hyperbolic sine tracking differentiators were used to approximate the derivatives of the virtual control signals and guidance commands, thereby alleviating the computational burden associated with traditional backstepping control. Furthermore, a fixed-time backstepping attitude controller was used to guarantee that the tracking errors converged to a small neighbor of the origin in fixed time. According to the simulation results, the proposed controller outperformed a fixed-time sliding mode disturbance, observer-based, finite-time backstepping controller in terms of the tracking precision and convergence rate. Moreover, the proposed controller was noted to be robust in simulations involving lumped disturbances.

Due to the domain shift, it is challenging to achieve ideal experimental results for cross-domain few-shot learning (FSL) in hyperspectral image classification. Most existing FSL algorithms are impacted by the limited samples, and they do not effectively leverage the representations from different layers of the network. Therefore, this paper proposes a few-shot learning based on embedded self-distillation and adaptive Wasserstein distance (ESAW-FSL) for hyperspectral image classification. Firstly, the embedding self-distillation network is proposed in the feature extraction process of the source and target domains. The embedding self-distillation network utilizes the self-distillation from different perspectives to get discriminative features. In the source domain, the mask evaluation of embedded features is employed to guarantee the learning of guiding features. Secondly , a domain adaptation based on adaptive Wasserstein distance is designed to alleviate the domain shift problem between the domains. A lightweight feature correlation network learns the comprehensive cost matrix in theWasserstein distance adaptively, and the obtained cost matrix helps achieving domain adaptation by an iterative algorithm. Finally, a focal loss based on double softening is adopted in the process of few-shot learning. The probability is double softened to improve the ratio of correctly classifying hard samples. Experiments are conducted on three widely used hyperspectral datasets and compared with six state-of-the-art algorithms. The overall accuracy and average accuracy are achieved in multiple experiments, demonstrating the effectiveness of ESAW-FSL.

Solar sails have many advantages over traditional chemical propulsion spacecraft, such as needlessness of fuel, high payload ratio, long service life, and great application potential in deep space exploration, interstellar voyage, and other aerospace application fields. However, the period of solar sail transfer from the initial orbit to the desired orbit is relatively long. Thus, it is necessary to optimize the transfer trajectory of solar sail according to specific tasks. The hp-adaptive pseudospectral method combines the global pseudospectral method with the finite element method, which adopts double layer optimization strategy to solve the optimal control problem and has higher computational efficiency and accuracy than the traditional global pseudospectral method. In this paper, the pressure of solar light acting on the solar sail is analyzed, and the kinematic equation of the solar sail is established in polar coordinate system first; then the basic principle of the hp-adaptive pseudospectral method is introduced, and the steps of solving the transfer trajectory optimization problem by hp-adaptive pseudospectral method are proposed; finally, the trajectory optimization of the solar sail from Earth orbit to Sun-centered Mars orbit is simulated as an example to demonstrate the effectiveness of hp-adaptive pseudospectral method in the orbit transfer optimization problem of solar sail. The simulation results show that the adopted method is not sensitive to the initial values and has more reasonable distribution and less computational cost than the Gauss pseudospectral method.

With the expansion of space exploration and activities, the space environment becomes increasingly complex, and thus the importance of non-cooperative spacecraft detection (NCSD) becomes more significant for enhancing space situational awareness capabilities. However, the implementation of NCSD in space faces two challenges: the resource constraints of spaceborne embedded systems and the object scale variations caused by complex imaging environments in space. Accordingly, an improved lightweight saliency model is proposed in this paper for NCSD in the space environment. In this model, a lightweight feature extractor that aims to extract multi-level, multi-scale features is designed as the backbone network using in-layer multi-scale block and intra-frequency multi-scale block. Integrity channel attention feature fusion modules are then applied to enhance the channel discrimination of multi-level features. Furthermore, a part-whole verification module is introduced to strengthen the learned integrity feature by measuring the consistency between the target parts and the complete target region. Extensive contrast experiments were conducted on the SwissCube dataset and the results clearly demonstrate that the proposed lightweight saliency model achieves relatively ideal segmentation effect of non-cooperative object spacecraft and competitive performance in terms of a wide range of metrics with only 1.99M parameters.