The inverse synthetic aperture radar (ISAR) image is a kind of target feature data acquired by radar for moving targets, which can reflect the shape, structure, and motion information of the target, and has attracted a great deal of attention from the radar automatic target recognition (RATR) community. The identification of ISAR image components in radar satellite identification missions has not been carried out in related research, and the relevant segmentation methods of optical images applied to the research of semantic segmentation of ISAR images do not achieve ideal segmentation results. To address this problem, this paper proposes an ISAR image part recognition method based on semantic segmentation and mask matching. Furthermore, a reliable automatic ISAR image component labeling method is designed, and the satellite target component labeling ISAR image samples are obtained accurately and efficiently, and the satellite target component labeling ISAR image data set is obtained. On this basis, an ISAR image component recognition method based on semantic segmentation and mask matching is proposed in this paper. U-Net and Siamese Network are designed to complete the ISAR image binary semantic segmentation and binary mask matching, respectively. The component label of the ISAR image is predicted by the mask matching results. Experiments based on satellite component labeling ISAR image datasets confirm that the proposed method is feasible and effective, and it has greater comparative advantages compared to other classical semantic segmentation networks.

A novel satellite target recognition method based on radar data partition and deep learning techniques is proposed in this paper. For the radar satellite recognition task, orbital altitude is introduced as a distinct and accessible feature to divide radar data. On this basis, we design a new distance metric for HRRPs called normalized angular distance divided by correlation coefficient (NADDCC), and a hierarchical clustering method based on this distance metric is applied to segment the radar observation angular domain. Using the above technology, the radar data partition is completed and multiple HRRP data clusters are obtained. To further mine the essential features in HRRPs, a GRU-SVM model is designed and firstly applied for radar HRRP target recognition. It consists of a multi-layer GRU neural network as a deep feature extractor and linear SVM as a classifier. By training, GRU neural network successfully extracts effective and highly distinguishable features of HRRPs, and feature visualization technology shows its advantages. Furthermore, the performance testing and comparison experiments also demonstrate that GRU neural network possesses better comprehensive performance for HRRP target recognition than LSTM neural network and conventional RNN, and the recognition performance of our method is almost better than that of other several common feature extraction methods or no data partition.

The addition of telescopic joints will enhance the manipulator’s motion dexterity but increase the motion complexity, thereby potentially reducing the efficiency of motion planning. To address the issue, a manipulability priority principle is used to reduce the computational dimension, aiming to enhance computational efficiency. Firstly, an improved Monte Carlo algorithm combined with Weighted Ensemble Learning is proposed to rapidly obtain a well-distributed and information-complete reachable workspace under small-sample conditions. Then, all dexterous points are computed by the manipulability formula and the least squares algorithm is applied to obtain the boundary of the reachable workspace. Based on the boundary, a spatial quadratic surface-based segmentation algorithm is proposed to obtain the dexterous workspace corresponding to all dexterous points. Subsequently, the Telescopic Motion Computation (TMC) method is designed to transform the kinematics problem of a 9-DOF manipulator into a 7-DOF problem by precomputing the telescopic joint lengths, effectively reducing the computational time for inverse kinematics. Finally, simulations are performed to verify the effectiveness, dexterity, and efficiency of the TMC method. The results indicate that applying the TMC method to PInv, DLS, and PSO algorithms enhances the efficiency by 3.43, 3.53, and 3.64 times while maintaining the motion dexterity.

Aiming at the penetration problem of high-speed vehicles against a modified proportional guidance interceptor, a three-dimensional mathematical model of attack–defense confrontation between the high-speed vehicle and the interceptor is established in this paper. The modified proportional navigation guidance law of the interceptor is included in the model, and control constraints, pitch angle velocity constraints, and dynamic delay are introduced. Then, the performance index of the optimal penetration of high-speed vehicles is established. Under the condition of considering the 180-degree BTT control, the analytical solutions of the optimal speed roll angle and the optimal overload of high-speed vehicles are obtained according to symmetric Hamilton principle. The simulation results show that the overload switching times of high-speed vehicles to achieve optimal penetration are N − 1, where N is the modified proportional guidance coefficient of the interceptor. When the maximum speed roll angle velocity is [60, 90] degrees per second, the penetration effect of high-speed vehicles is good. Finally, the optimal penetration guidance law proposed in this paper can achieve a miss distance of more than 5 m when the overload capacity ratio is 0.33.