With the development and application of artificial intelligence and unmanned technology, the process of large-scale confrontation in future air-control operations is accelerating.The development of future air-control combat platforms and air-control combat weapons is analyzed, the main characteristics of large-scale air combat are sorted out, including high-density coordinated confrontation on combat platforms, saturated coordinated attack of air-control weapons, and network information as the basic support, and the requirements and impact of large-scale air combat on the future development of air-to-air missiles are analyzed from the aspects of combat usage, production and manufacturing, and design and development, which can provide certain technical reference for the development of air-to-air missiles in the future.

In recent years, although deep learning has been widely used in the field of Synthetic Aperture Radar (SAR) target recognition, current insufficient sample size of annotated SAR data seriously restricts the development of deep learning in SAR target recognition.In contrast, transfer learning can overcome the data-driven limitation of deep learning by using limited SAR samples for transfer learning.In this paper, a transfer learning based target recognition algorithm with limited SAR samples is analyzed.Firstly, the basic concepts, types and common strategies of transfer learning are introduced and the feasibility of its application in the field of SAR target recognition under small sample size is analyzed.Then, according to whether the transferred data is homologous with data in target domain or not, the representative algorithms of the two types of transfer learning based methods in the field of SAR image recognition are sorted out and summarized.Finally, from the two perspectives of the insufficient sample size and the universality of the network, the shortcomings of transfer learning in SAR image recognition tasks and the next-step research directions are discussed.

To meet the application requirements of the new generation airborne avionic system for the avionics network, the structural characteristics of the airborne switching network are analyzed, and the system operation management methods of the new generation avionics network and the characteristics of the Avionics Time-Sensitive Network (ATSN) control technology are introduced.Through the detailed description of the technical scheme of ATSN control, the design idea of realizing the new generation airborne data communication switching network control oriented to future airborne system applications is expounded.

To solve the problem of stealth UAV penetration path planning in complex battlefield environment, an Improved Ant Colony Optimization (IACO) is proposed.The weight of interaction between adjacent nodes and the path selection weight are introduced, and the improved heuristic function is adopted to enhance the guidance of the target point area, so as to improve the search efficiency and maintain the diversity of solutions.The mode of pheromone dynamic adjustment is designed, and a new pheromone concentration updating strategy is proposed to enhance the global search ability.The simulation results show that, compared with the traditional Ant Colony Optimization (ACO) and the Particle Swarm Optimization (PSO), the IACO can effectively avoid the threat of networked radar and thus improve the survivability of stealth UAVs.In addition, the proposed method has better performance in computing efficiency and safety, which verifies the effectiveness and superiority of the IACO.

To solve the problem of unstable configuration of multi-UAV formation caused by model biases and external disturbances, an idea of model-compensation control is proposed.Based on this, a model-compensation backstepping controller based on Compensation Function Observer (CFO) is designed and applied to formation cooperative control.The formation cooperative control problem is transformed into a traditional control problem, so as to realize the assembly, flight and shape switching of multi-UAV formation under complex external disturbances.The proposed controller not only retains the global exponential asymptotic stability based on Lyapunov criteria, but also employs a CFO to accurately estimate model biases and disturbances and then compensate for them into the controller, effectively enhancing the anti-disturbance capability of the overall formation system.Finally, the effectiveness of the algorithm is verified by simulation and experiment.

To solve the problem of main-lobe suppression jamming suppression in Pulse Doppler (PD) radar, two algorithms are proposed.The first algorithm, which is based on the idea of Blind Source Separation (BSS) in beam space, reduces the dimension of multi-beam echo after coherent integration to construct the signal to be separated, and takes Rényi entropy as the index to identify the separated source signals.The second algorithm, which is based on the idea of correlation filtering, takes one beam echo as the matching signal, performs correlation filtering on the other beam echo, and then obtains a suppressed beam echo through inverse operation.The simulation results verify the effectiveness of the proposed two algorithms.

To solve the problems of inadequate feature information,serious feature loss and low recognition accuracy in the process of infrared UAV target recognition,an infrared UAV target detection method based on YOLOv7 is proposed.By introducing the attention mechanism,the feature representation ability of the target region is enhanced and the spatial information content of the image is improved.The improved serial connection mode is used to connect the channel attention module to the spatial attention module.While combining the channel feature information with the spatial feature information,the improved structure reduces the negative impact of the channel attention on infrared image recognition,and can better realize the feature strengthening of the infrared target.The SIoU loss function based on angle vector regression is selected as the frame loss function,which further improves the convergence and detection accuracy of the model.The experimental results show that the reasoning speed of the improved algorithm model reaches 43 frames per second,the accuracy is 95.4%, the recall rate is 87.3%, and the mAP is 96.1%.Better results are obtained in the infrared UAV detection task.

To solve the problems of small scale, possible occurrence of repeated reconnaissance and low reconnaissance efficiency without considering energy constraints, a multi-target trajectory planning algorithm based on energy constraints for UAV swarm cooperative reconnaissance is proposed.Firstly, various energies required by UAV missions are modeled, and the consumption of flight energy, hovering energy and data transmission energy is quantitatively analyzed.Then, the idea of graph theory is used to model the reconnaissance region and the reconnaissance target of the swarm, a new trajectory planning algorithm model is built under the energy constraints, and the collaborative optimization of the UAV swarms trajectory is conducted to reduce the overall energy consumption of the swarm.The simulation results show that the proposed algorithm is significantly better than the representative distance-based trajectory algorithm in terms of energy consumption and reconnaissance time.When 10 UAVs scout 100 targets, the energy consumption is reduced by 48%, and the reconnaissance efficiency is higher and the convergence speed is faster.

Anti-sorting signal design is an important direction of Radio Frequency (RF) stealth signal design.It is based on anti-sorting signal design principle, which is essentially the failure principle of radar signal sorting algorithm.Clustering pre-sorting plays a key role in radar signal sorting, which has the advantages of fast sorting, simultaneous sorting of multiple emitters, and significant reduction in the computational burden of main sorting.However, in the field of RF stealth anti-sorting signal design, the study of a unified and widely applicable clustering sorting failure principle to guide the anti-sorting signal design has not been disclosed in formal literature.In this paper, by studying the principle of K-means clustering algorithm based on data field and fuzzy C-means clustering algorithm, and as for the step of data similarity measurement which must be carried out in clustering algorithms, a clustering sorting failure principle that can be widely used is proposed.It makes both clustering algorithms fail in sorting and guides the design of RF stealth anti-sorting signal.The formula derivation and signal simulation prove the correctness of the principle, which provides theoretical support for the design of RF stealth anti-sorting signal, and can improve the efficiency of RF stealth anti-sorting signal design.

To solve the path planning problem of UAVs in three-dimensional space, the standard artificial fish swarm algorithm is improved from the perspectives of accelerating the convergence and ensuring the convergence effect.Firstly, the improvement factor α of adaptive field-of-view and adaptive step size is introduced to speed up the convergence of the algorithm.Then, the minimum field-of-view and the minimum step size are set to ensure the search ability of the algorithm.Finally, the jumping behavior is added to ensure that the algorithm can jump out of the local optimum at the later stage.The algorithm is simulated and compared with other algorithms through software.The simulation results show that the algorithm has fast search speed and good convergence effects at the initial stage of search, and can jump out of the local optimum at the later stage of search, and the effectiveness of the algorithm is verified.The comparative experiment results show that the optimal fitness of the algorithm is improved at the later stage of search, and the superiority of the algorithm is verified.

The ship has become an important monitoring target in maritime military field.Ship target detection in SAR images suffers from poor detection effects, large computation amount and weak generalization capability.To solve the problems, a lightweight ship target detection algorithm based on YOLOv5 and Mobilenetv3 is proposed.Firstly, Mobilenetv3 backbone network is introduced to reduce the computation amount and volume of the model and realize lightweight processing of the model.Then, the EIoU loss function is introduced to improve the regression accuracy and convergence speed of the prediction box.Finally, CBAM is introduced into the neck network, and attention adjustment is conducted at the stage of feature fusion to improve the detection accuracy and detection effects of the model.The experimental results on SSDD dataset show that the volume of the improved algorithm model is reduced to 18.32% of that of the original YOLOv5 model, the training time is shortened by 35.22%, the parameter quantity is reduced to 15.94% of that of the original model, the computation amount is reduced to 10.76% of that of the original model, and mAP is improved to 98.3%.The experimental results show that the improved algorithm greatly reduces parameter quantity, computation amount, model volume and training time while maintaining high-precision detection effects, which can realize real-time detection of ship targets in SAR images.

The fault-tolerant altitude and attitude control of a quadrotor UAV subjected to actuator faults and external disturbances is investigated.Firstly, the actuator faults are modeled as a constant Loss of Effectiveness (LoE) in the thrust generated by the propellers, and the dynamics UAV altitude and attitude models considering the actuator faults are obtained.Secondly, a fault-tolerant controller is designed based on the Integral Terminal Sliding Mode Control (ITSMC), and the disturbances are estimated and compensated for by using the adaptive law.Thirdly, the stability of the closed-loop system is proved by Lyapunov theory.Finally, numerical simulation results verify the effectiveness and robustness of the proposed fault-tolerant control scheme.

To solve the problem that it is difficult to measure the aircraft sortie rate under the condition of “no engine stopping”, a simulation measurement method based on Hierarchical Timed Colored Petri Net (HTCPN) is proposed to measure the aircraft sortie rate.Firstly, based on the aircraft maintenance work at the grassroots level, a continuous sortie model and the maintenance process are established.Secondly, Petri net theory and CPN tools software are used to construct the computer simulation model of the whole process of maintenance support based on HTCPN.Finally, the key factors that restrict the sortie rate of aircraft are identified through an analysis of calculation cases.The results show that the proposed method can not only find out the key factors affecting aircraft sortie, but also provide technical support for the measurement of aircraft sortie rates, which has certain value in engineering applications.

The dual-propeller propulsion compound helicopter has variable flight modes.The transition phase is a necessary stage to realize the conversion of different flight modes, and the asymmetry effect of propellers on both sides intensifies the coupling between lateral and heading attitude.In order to ensure the compound helicopter flight safety in transition flight phase, a lateral and heading attitude decoupling control system is designed.To solve the problems of control mode conversion and manipulation coupling between lateral and heading channels in the transition flight phase of compound helicopter, a linear transition strategy of control modes taking control efficiency as the constraint is formulated.Based on the explicit model tracking decoupling control theory, the attitude controller with control matrix decoupling is designed, and the simulation model of the lateral and heading attitude decoupling control system is constructed.The simulation results show that the control system can effectively ensure the smooth transition of the compound helicopter control modes and realize the lateral and heading channel decoupling control.

Faced with complex electromagnetic environment of urban low-altitude areas,GNSS signals are prone to be interfered during UAV operation,leading to inaccurate positioning.In order to improve the accuracy of UAV positioning,a data fusion method based on extended Kalman filtering is proposed.Based on the inertial navigation system,the fusion with 3D positioning based on 5G signals is conducted,and inertial navigation errors are corrected by using the fused data.The software simulation shows that the attitude error and the position error are limited within a certain range in the absence of effective GNSS signals,and the accuracy of UAV positioning is improved.The average position error under sight distance is 16.6 cm,and the positioning accuracy is improved by 49.7% compared with that before fusion.The UAV’s ability to have high positioning accuracy for a long duration is realized,and the method has certain engineering practicality.

In the traditional calibration method,the MEMS-IMU for high-speed rotating projectiles cannot fully activate the error, and the calibration process is complex and the calibration efficiency is low.To solve the problems, this paper designs an automatic calibration system and method of the MEMS-IMU for high-speed rotating projectiles.Firstly, the output characteristics of the MEMS-IMU for high-speed rotating projectiles are analyzed, an error model is constructed, a twelve-position, eighteen-speed calibration scheme is designed, and the turntable rotation scheme is optimized.Through the automatic calibration system, the automatic calibration of the MEMS-IMU for high-speed rotating projectiles is realized.The experimental results show that the proposed method can achieve the automatic calibration of the MEMS-IMU for high-speed rotating projectiles.Compared with that of the traditional method, the accuracy of MEMS gyroscope on roll, pitch and yaw axes is improved by more than 98%, 72% and 73% respectively, and the accuracy of MEMS accelerometer on the three axes is improved by 53%, 47% and 7% respectively.The duration of the whole calibration process is greatly reduced, which improves the calibration efficiency.

To solve the problem of low firing accuracy of shipborne artillery at sea, an optimal sliding mode control strategy based on RBF neural network is proposed.Firstly, the mathematic model of shipborne artillery servo system is established, and a Global Robust Optimal Sliding Mode Controller (GROSMC) is designed based on the linear quadratic optimal control theory and sliding mode control.This improves the response speed and ensures the good robustness of the system.Then, RBF neural network is used to dynamically adjust the gain of the switching control item to reduce the sliding mode chattering problem.Finally, the effectiveness of the designed controller is verified by simulation comparison.The simulation results show that this strategy has good control performance and meets the system requirements.

To solve the problem of small field-of-view of the camera in the visual servo of rotor UAVs, a visual servo control method based on PTZ camera is proposed, which realizes the fixed-point hovering control of rotor UAVs outside the field-of-view of classical visual servo.Firstly, the image feature kinematics model of PTZ camera is established, and the visual servo control method of rotor UAVs based on PTZ camera is designed.The fuzzy PID algorithm is integrated into PTZ control to improve the response speed.The visual servo fixed-point hovering simulation of the rotor UAVs under the condition of a near target point and a middle-to-far target point is conducted respectively.The simulation results show that the proposed method greatly expands the field-of-view of the visual servo control of the rotor UAVs, which widens the application scenarios of the visual servo control of the UAVs.

Real-time detection of remote sensing images is one of the key technical problems in the field of remote sensing application.As for current mainstream target detection algorithms, there are problems of a large number of model parameters, bad real-time performance, high power consumption and high costs on the image processor (GPU).To solve the problems, a real-time detection scheme of remote sensing images based on Field Programmable Gate Array (FPGA) is proposed.Firstly, in order to reduce the quantity of parameters and improve the detection speed, MobileNetv2 is taken as the feature extraction network, and the fusion with depth separable convolution is conducted, making the model lightweight and easy to deploy.Then, CA attention module is used to improve the detection accuracy.Finally, the floating point parameters of the model are quantified into 8-bit fixed point numbers, and the network model is deployed on FPGA after quantization.The experimental results show that on the remote sensing data set VisDrone 2019, the mean Average Precision (mAP) of the scheme designed in this paper reaches 14.79%, FPS reaches 46.78 frame/s, and average power consumption is 8 W.The detection speed is 375.4% higher than that of CPU, and the power consumption is 96.8% lower than that of GPU.The scheme can meet the requirements of real-time target detection, and can be deployed in power-limited satellite, UAVs and other equipment.