中国科技核心期刊
中国指挥与控制学会会刊
军事装备类重点期刊
中国指挥与控制学会会刊
军事装备类重点期刊
中国指挥与控制学会会刊
军事装备类重点期刊
The cooperative region search of UAV swarm can obtain ground information of the mission region and reduce the uncertainty of environmental information effectively. The traditional collaborative region search methods based on the balanced allocation of divided region and the heuristic algorithms depend on the pre-designed rules and heavy computation, and have no ability to generate new rules of the cooperative search. These algorithms belong to the algorithms that can not evolve new rules. Due to the complexity of the mission environment, the algorithms must contain fast,intelligent and robust characteristics, the cooperative searching algorithms of UAV swarm based on emerging theory with strong information fusion ability, self-learning ability have been widely concerned. Evolutionary and reinforcement learning algorithms are the important parts of the emerging theory and both of them can generate some new cooperative searching rules according to the different environment and task. The paper would systematically analyze and summarize the current research status and progress of cooperative search methods. Finally, the shortcomings of the existing research and the further development are put forward.
Aiming at the target search task, considering the existing problems of unreasonable initial distribution of UAVs and poor consistency of start and end time of swarm tasks in the existing coverage path planning algorithms, this paper designs an equal-time tendency region segmentation algorithm based on the initial scenario of realistic UAV cluster centralized deployment. The optimization goal of the algorithm is to minimize the maximum waiting time of individuals in the UAV swarm, and the task time of the UAV is affected by changing the size of the UAV search area. The algorithm has a secondary structure. The first level is the initial coarse segmentation, which solves the problem of too many iterations of boundary points. In the second level, the task time deviation value is used as the adjustment value to ensure the unity of task time. The simulation results show that the algorithm designed in this paper is more suitable for the scene of concentrated UAV delivery. The algorithm successfully reduces the task waiting time between UAV individuals. It facilitates the secondary scheduling of resources and the synchronous execution of multi-order tasks.
A dynamic obstacle avoidance algorithm is proposed for the obstacle avoidance problem in the cruise task of unmanned boat cluster. Firstly, the square grid trajectory cell (SGTC) situation matrix of the waters around the unmanned surface vessle is obtained. Then, based on the principle of bacterial walking foraging, the optimal obstacle avoidance path is dynamically matched in the SGTC set. Finally, on the basis of dynamic obstacle avoidance, the autonomous cruise process of unmanned surface vessle is divided into stages of moving towards the target, dynamic obstacle avoidance, and return replenishment, so as to realize the normal cruise of unmanned surface vessle to the designated waters. The simulation results show that the proposed method has efficient avoidance ability for static point, surface obstacle and moving obstacle. Based on the algorithm, it can realize the normalized unmanned cruise of unmanned boat cluster. The algorithm is not only suitable for obstacle avoidance of unmanned surface vessle, but also has broad application prospects in other heterogeneous unmanned equipment.
The Manned-Unmanned Teaming technology can fully leverage the low cost and expendable advantages of unmanned platforms to expand the operational space of manned platforms, thereby enhancing their situational awareness, penetration capability, and survivability. This paper provides a research overview on the development of MUM-T in the United States, Britain, France, Australia, South Korea, and other countries. It provides several application examples of MUM-T in foreign militaries on marine, land, and air space. It also analyzes some technical challenges and key technologies of MUM-T in wireless communication and networking, remote measurement and control, autonomous decision-making, and human-machine interaction. Finally, the future development trend of MUM-T is prospected, which can serve as a reference for the study of collaborative combat styles and related technological research in future joint operational systems.
Multi-UAV cooperative regional search is widely used in military and civil, such as search, rescue, patrol, monitoring, and environmental survey. It is an open topic to enhance the efficiency of target search. The paper proposes a probability calculation method based on UAV track elimination strategy and policy iteration on a target region without obstacle to improve the efficiency of UAV cooperative search based on prior information. The track elimination strategy is measured by the track of the UAV and the prior information, which include detectivity of sensor as well as heat power of target region. Then heat power on the track of a single UAV is refreshed, and the track is dynamically planed by policy iteration algorithm. On the basis of the method, the tracks of all UAVs are calculated one by one with appropriate order. Finally, the effectiveness of the algorithm is verified by simulation.
In view of the catastrophic forgetting of previous knowledge in class incremental learning for image classification, existing replay-based methods focus on memory updating and sampling, while overlooking the feature relationships between old and new samples. To this end, the paper proposes a method called contrast metric enhancement based on memory extraction(cME2) for Online Class-incremental Learning in Image Classification, which designs two new types of positive and negative sample-pairs, enhances the reuse of old sample feature information, and strengthens the ability of model to express redundant features and common features. It improves the distribution of samples in embedding space based on the nearest class mean classifier. Finally, the effectiveness and efficiency of the proposed method are verified by comparison experiment and ablation experiment.
To solve the problem of low track prediction accuracy caused by the limitations of deep learning and the cumulative error generated by recursive prediction strategies, a short-term prediction algorithm for air target tracks based on residual correction CNN-BiLSTM was proposed. Firstly, a convolution module was introduced to extract potentially associated spatial location features from the track data, and a bidirectional long and short time memory network was used to extract temporal features from the track data, achieving real-time one-step prediction and multi-step advance prediction of air targets. Then, the integrated moving average autoregression was introduced as a residual correction model to model the residual generated by real-time one-step prediction, and the residual value of the hybrid neural network model for multi-step advanced prediction is calculated. Finally, the output results of the hybrid neural network model and the residual correction model are fused to obtain the final trajectory prediction value. Experiment results proved that the algorithm can significantly improve the accuracy of short-term prediction of airborne target tracks.
In the ship information big data system, aiming at the problems such as long time consuming, distortion of thermal map effect and poor tracking or lag in the roaming of map software when the client requests a large amount of historical data from the server and draws heat map in real time, this paper proposes an improved real-time generation strategy of map map of ship history activity. Based on the thermal map generation method with the idea of pixel clustering, this strategy uses the image enhancement algorithm based on nonlinear transformation and pixel complement method to correct and optimize the display distortion problems existing in the head map. On this basis, the dynamic loading and drawing strategy and multi-thread real-time mapping strategy are proposed to generate the heat map of the ship’s historical activities. Experiments show that the strategy proposed in this paper has been improved in the efficiency of heat map drawing, mapping effect and interactive experience of map software.
To estimate the parameters of Linear Frequency Modulated (LFM) signal based on Fractional Fourier transform (FRFT), the key issue is to determine the optimal order of FRFT. A new parameter estimation algorithm is proposed based on the idea of error iteration. The algorithm utilizes the conversion relationship between normalized bandwidth and rotation angle. Calculating the angle difference from the estimation error effectively reduces the amount of computation and does not require the prior information of positive and negative frequency modulation slope. The improved logarithmic search algorithm can further improve the stability and reliability of parameter estimation results. The simulation results show that the proposed method still has good parameter estimation performance under the premise of high efficiency when the signal-to-noise ratio is above -8 dB. The average estimation error is less than 1%, and the estimated results are close to the Cramer-Rao Lower Bound, meeting the requirements of engineering real-time processing.
This paper proposes an improved adaptive noise-aided complete ensemble empirical mode decomposition (ICEEMDAN) method to address the issue of low signal-to-noise ratio in distributed optical fiber acoustic sensing systems. The proposed approach utilizes sample entropy and wavelet threshold denoising algorithm to extract valuable components from high noise components. The ICEEMDAN is applied to decompose the acquired signals, and sample entropy is calculated to identify the noisy components, which are then subjected to wavelet threshold denoising. Finally, the denoised components are reconstructed with the untreated intrinsic mode functions. Experimental results demonstrate that the denoising treatment significantly enhances the signal-to-noise ratio by 5.34 dB, reduces the mean square error by 0.014 8, and improves waveform similarity by 5.7%. Compared to other commonly used denoising methods, the proposed approach not only exhibits superior performance in terms of signal-to-noise ratio but also demonstrates better performance in mean square error and waveform similarity, thereby preserving useful signals more effectively.
For the lack of command and control modeling principles and mechanisms in the current end-to-end campaign tactical level simulation for unmanned air combat, this paper systematically combs and studies the recent research in unmanned air campaign tactical simulation. Combined with the relevant research, this paper explored the principles and methods of different C2 modeling at different simulation levels and established a C2 simulation system suitable for the characteristics of CISE in unmanned air combat. Further, theoretical and practical guidance is provided for innovative demonstration and evaluation methods of CISE through the effectiveness evaluation support of unmanned air combat tasks.
Intelligent training is the process of using machine learning algorithms to train and optimize neural network agent models. The agent model achieves intelligent improvement through continuous trial and error training. Large scale training data is a necessary condition for intelligent learning training, which is usually difficult to obtain directly from the real world. How to generate a large amount of effective training data through simulation is an important research direction for intelligent agent training. This article proposes an intelligent parallel training method based on simulation experiments. By utilizing simulation experiment management, batch simulation experiment scenarios can be quickly generated, and nodes can be automatically deployed and run. Intelligent parallel training can be achieved through reasonable training architecture design and effective training process design. The simulation experiment management process of intelligent training is demonstrated through practical cases, and combined with training results, it is proven that the method proposed in this article improves the efficiency of intelligent training and the generalization of intelligent agents.
The penetration ability on warship target by naval gun semi-armor-piercing ammunition is analyzed. LS-DYNA finite element software is used to simulate the ricochet, attacking main deck and ship board by Oto Melara 127 mm and 76 mm naval gun ammunition. It is concluded that the critical ricochet angle of 127 mm and 76 mm ammunition is 9.5°and 29°when striking on 15 mm E36 plate in 300 m/s. Both ammunitions are able to penetrate several layers attacking from main deck and ship board. When attacking warships, it is considerable to delay initiation time properly, so the ammunition can enter central cabin for improving damage effect.
In this paper, aiming at the problem that the explosion belt is offset by wind during the operation of rocket explosive device, the dynamic model affected by wind and launching angle in different environments are established. The relationship between wind speed, wind direction and launching angle of rocket explosive device is analyzed and studied, and the model is simulated and verified by combining the relevant data. The model can provide method support for rational use of rocket explosive device and improvement of operational efficiency under actual combat conditions.
In order to assess military requirement scientifically and effectively, through steps division of stages, determination of assessment object and construction of assessment indicators are systematically described the military requirement assessment problem. Adopt formal language, focus on description of requirement value in the requirements demonstration stage, capability valve in the implementation stage and assessment satisfaction value in the inspection stage assessment. Study their assessments trends. Based on Kalman filter ideas, conduct experimental analysis on the proposed formal model. The results show that assessment trend can gradually approach the ideal value, and errors can converge. Factors such as assessment cycle, initial assessment value, expected error, measurement error will have an impact on both the assessment trend and the assessment results. Relationship between the comprehensive assessment results and the sub assessment indicators are non-linear. The results indicate that the basic laws of assessment can be characterized by established formal model. Research results can provide a theoretical basis of military requirement assessment.
In this study, a joint power and time allocation algorithm in multi-radar system for cooperative target detection is proposed in multi-target search and tracking scenarios. The evaluation metrics for radar search and tracking performance are derived separately, based on signal detection theory and the Cramér-Rao lower bound. A joint optimization allocation model is developed for multi-aircraft radar systems geared towards cooperative detection. This model seeks to maximize radar operational performance as the optimization objective, which will be achieved under the constraint of predefined system resource. Parameters including radar selection, radiated power, and time in radar search and tracking missions are jointly optimized. By incorporating the interior point method and particle swarm algorithm, a three-step decomposition approach is employed to solving the optimization problem. The results reveal that the proposed algorithm compared to existing algorithms improves radar system search performance and tracking accuracy effectively while adhering to the predefined system resource constraint.
Aiming at the comprehensive application of marine relay equipment, combined with the maximum communication distance and target threat distribution and other conditions, this paper proposes an programming model to solve the optimal deployment of communication relay platform array. This model is based on the principle of the shortest distance between the route starting point and the array postion point of the relay platform. It models the array deployment problem of the relay platform under the constraints of communication distance and target threat, and uses the convex optimization method to obtain the analytical form of the optimal solution. The optimal solution has strong interpretability, and the solution process avoids using traditional iterative or heuristic methods, which has the characteristics of small calculation. The example analysis shows that the optimal solution of the array deployment can be obtained quickly for the different position distribution of the mission platform and the threat target.
Tactical Internet traffic has extremely dynamic spatio-temporal features and is closely related to external features such as weather and elevation, existing network traffic prediction models can not extract its complex features well, a tactical Internet traffic prediction model that fuses multi-source dynamic spatio-temporal features is proposed. Firstly, external features are fused with traffic features as multi-source features; then the spatial features of network traffic at the current moment are extracted and the convolution weights over time are iteratively updated to obtain the spatial feature information under different time slices; finally, the spatial information of the current and historical moments are aggregated by the temporal convolution layer to predict the multi-source dynamic spatio-temporal traffic at the next moment. Compared with the single base model, the proposed method is better in all three evaluation metrics, namely, mean absolute error (MAE), root mean square error (RMSE) and coefficient of determination (R2).
This paper discusses the reliability deployment of Service Function Chain (SFC) on the hybrid satellite network based on SDN architecture. Firstly, this paper describes the reliability protection problem of SFC, establishes the underlying network and SFC request model, and then establishes the reliability requirement model of network service function and the reliability requirement model of LEO satellite link, and clarifies the optimization goals and constraints. Then, a reliability-based satellite service function chain protection method is proposed, including a reliability protection algorithm based on deep reinforcement learning and a reliability backup algorithm based on LEO satellite node and link. Experiments show that the proposed reliability-based satellite service function chain protection method can improve the SFC request acceptance rate, reduce the average delay, and maintain high request acceptance on hybrid satellite networks based on SDN architecture.
The operational application of Starlink system is becoming more and more mature. Considering that Starlink system is still under rapid construction, it is necessary to conduct a systematic analysis of its future operational application. In order to explore the future operation mode of the Starlink system of the US military, this paper, based on the analysis of the fundamental state of Starlink, focuses on high-end wars of great powers, explores the impact of Starlink on the operations of the US military from four levels of operational theory, and shows the system supporting role of Starlink. Finally, an operational simulation experiment of Starlink is given to analyze the long-term capability. By analyzing the operational mode and capability of the mid-stage Starlink operation, it lays a foundation for the follow-up countermeasures.
