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 general fusion model only considers the single attribute (type) of the difference feature and the set-valued mapping between the fusion algorithm to determine the fusion algorithm, which makes it impossible to dynamically adjust the algorithm according to the changes of multiple attributes of the difference feature, which leads to the problem of poor fusion effect. A dual model is proposed. Deep set-valued mapping between multi-attributes and fusion algorithms of different infrared images. By studying the characteristics of multiple attributes (type, amplitude, and frequency) of the difference feature and the effective distribution of the algorithm for the fusion of the different attributes of the difference feature, the method of the set-valued mapping between each attribute of the difference feature and the algorithm is pointed out. Possibility distribution synthesis is a means to give the method of the deep set-valued mapping of the difference feature and the fusion algorithm. the two key technical problems of the algorithm for the fusion effectiveness distribution of multiple attributes of the difference feature and the multiple iterations of the mapping are analyzed. This research provides a theoretical basis for the fusion model to select a targeted fusion algorithm based on multiple attribute changes of different features, and to improve the fusion effect of the two types of images.
Sensor scheduling refers to a method under the certain criteria and combined with the practical application scenarios that make a series of adjustments and controls for working mode, working time, maneuver path and others of the sensor platform or the sensor itself according to the observation information, thus to reduce the use cost of the sensor system and improve the tracking efficiency and prolong the system life, it's also an important content of sensor management. Based on the partial observable Markov decision process, this paper briefly describes the basic structure of sensor scheduling, analyzes the scheduling model and basic process, summarizes the relevant rules and constraints in the scheduling, and discusses the scheduling technology and method from the decision indicators and optimization methods, and analyzes the application of various optimization algorithms in solving the scheduling scheme. Finally, it puts forward some personal view about the hot topics and development direction of sensor scheduling.
In the face of various rampant illegal activities, advanced illegal means, and changeable law enforcement tasks, law enforcement departments have diversified equipment types, single machine and single control, and lack integrated cutting-edge law enforcement command means, which cannot meet the requirements of all-round and multi-angle real-time command. To solve this problem, this paper applied systems engineering method to design a set of police cluster command system based on multi-sensor integration, multi-network integration scheduling application technology, multi-sensor information integration technology, based on the evidence of deep learning feature extraction technique, based on the map image command auxiliary decision system technology, to research the key technologies such as suspect complete evidence information quickly and to provide powerful guarantee for law enforcement command action effectively, the system must be innovative and available.
For the control of multi-radar co-detection mode in airplane formation, a co-detection mode control strategy is proposed based on neural network. Firstly, the fuzzy logic theory is introduced to perform the fuzzy characterizing of battlefield situation. Then, the mathematical model of the mapping relationship between the battlefield situation and co-detection mode based on neural network method. Finally, the choose advise of co-detection mode is obtained with the real time battlefield situation inputting to the model. The simulation results show the neural network can control the co-detection mode switch effectively and have a favorable precision.
In the practical application of the correlation filter in long-term tracking scene, there are some problems such as illumination change, scale change, motion blur, occlusion. On the basis of fast discriminant scale space tracking algorithm (fDSST), this paper proposes a long-term target tracking algorithm, which uses confidence level for updating and occlusion determination, and adds a re-detection module. The historical peak mean value and historical APCE confidence mean value are used as the benchmark to update the template adaptively and determine the occlusion. An online cur filter is added for re detection. If occlusion is identified, the re-detection module is started to generate candidate areas and determine the target position again. With the public dataset OTB50, the results of verification experiments have shown that the proposed approach can effectively handle with occlusion and scale-changing. Compared with the representative algorithm, tracking-accuracy and successful rate are remarkably improved.
This paper establishes a dynamics model in the free segment, aiming at the situation of the spacecraft movement in the free segment, and discusses the related performance of centralized fusion and distributed fusion based on UKF filtering under multi-base radar observation. In the centralized fusion algorithm, parallel filtering and sequential filtering algorithms are used, while in the distributed fusion method, simple convex combination fusion, Bar-Shalom-Campo fusion and federated filtering are used, and the applicability of five filtering algorithms to the fusion of state estimation in the free segment of ballistic missile is compared. The simulation results show that the federated filter algorithm based on UKF filtering should be preferred in the free segment target state fusion.
In order to solve the problems of poor robustness and easy divergence of traditional unscented Kalman filter algorithm in maneuvering target tracking, this paper combines square root unscented Kalman filter (SRUKF) and interacting multiple model algorithm (IMM) in the application background of bistatic angle only passive location. Firstly, the principle of bistatic direction finding cross location is analyzed. The results of geometric positioning are taken as the initial filtering value. Then, the spherical traceless transformation is used to replace the traditional symmetric sigma point sampling, and the square root of covariance is used to replace the covariance for recursive operation, which ensures the non negative qualitative of the covariance matrix, improves the robustness of the algorithm and reduces the amount of calculation. Simulation results show that IMM-SRUKF algorithm can accurately predict the moving state of the target at different times. When the observation error is large, compared with other algorithms, IMM-SRUKF has higher stability, less divergence and higher filtering accuracy.
According the next-generation command post, to improve situational awareness and decision-making efficiency, this paper gives the concept of Smart Command Space (SCS) from the view of ambient intelligence, and provides five traits for SCS which are task-centered design, active service, spontaneous interaction, display fusion, and elaborate interconnection. It also gives the concept of display resources. At last, it introduces the framework for SCS and designs some latent scenes. All above can provide available thought for the building of next-generation command post.
In order to meet the requirement of operation systems capacity assessment,this paper using Analytic Network Process (ANP)and empirical method,designs an index framework that takes into account requirements of operation assessment and systems assessment,and presents the application process for assessment. The application case shows that the designed index systems framework can describe the operation process reasonably and effectively, reflect winning mechanism of the operation systems, and meet the assessment conditions.
This paper discusses a method of constructing land command decision simulation model. The model is a hybrid artificial intelligence system that brings strategy into the modeling of doctrine through the use of fuzzy rule sets for the doctrine and an algorithm similar to the one used to play chess for the strategy. At the same time, it builds the dispositional model of commander's character to bias those decisions with a commander's leadership style.
As a key weapon of underwater military battle, the control technology of torpedo has been the hot issue all over the world. A data-driven control based method is proposed for the research of lateral dynamics control of torpedo. When the hydrodynamic parameters of torpedo are unknown, the input-output data is used to design the control scheme of the vertical rudder angel. Because the data-driven based control method needs little calculation, it is more practical to realize in the real systems. In the designing process of data-driven based control algorithm of vertical rudder angel, the unknown dynamics and the error of the discretization are estimated by special parameter-identified updated law through the input-output data in this paper. Combined with the sliding mode control technology, the data-driven based algorithm of the vertical rudder angel comes out. It is proved to be stable using Lyapunov stability method. The simulation results also indicate the effectiveness of the proposed algorithm.
Combat behavior evaluation of the enemy based on reinforcement learning can enhance the intelligent ability of simulation and deduction. The training speed of reinforcement learning has become the key of its military application. Firstly the combat behavior evaluation of the enemy is modeled as multi-task reinforcement learning and the environment knowledge and experience are introduced into the learning algorithm. The combat behavior evaluation method of enemy based on hierarchical episodic meta-deep reinforcement learning is proposed with the idea of cooperation. The hierarchical episodic meta-deep reinforcement learning agent structure is designed and its concretely flow is discussed. The episodic memory system is designed based on the Differentiable Neural Dictionary (DND) to solve the incremental parameter adjustment problem. The architecture melds the standard Long-Short Term Memory (LSTM) working memory with an episodic memory. The hidden activations from the previous encounter of LSTM are reinstated. The OpenAI Gym and aircraft intelligent game simulation platforms are chosen for the method validation. The results show that the prediction method has better performance on the task of CartPole-V0, episodic “two-step task” and combat behavior evaluation and attains the goal of accelerating the reinforcement learning speed with cooperation of hierarchical episodic deep reinforcement learning and meta-learning.
The doppler frequency difference, power difference and the angle difference between the target and the mirror caused by the low altitude mirror seriously affect the stable tracking effect of the semi-active ground seeking guided weapon. In this paper, the echoes of missile-target and missile-target images in land (sea) environment are obtained through digital simulation, and the problems of lobe splitting are improved by means of high projectile trajectory, constant elevation angle of irradiator, etc., so as to improve the combat influence of mirror targets on semi-active guidance system weapons in ultra-low altitude environment.
On the basis of the previous research of the evaluation indexes of attacking targets, according to the special mission features of antitank element, this paper proposes the evaluation index system of attacking targets, which includes four aspects and fifteen indexes; uses BWM to calculate the weight of the evaluation of index; adopts a fuzzy optimization method, calculates the results of indexes through normalization, and calculates the value of attacking targets according to index weight. At last, based on the experiment scenario,this paper analyzes the problem of attacking target compositor,which verifies the validity of the above-mentioned method and provides decision-making evidence for commander of antitank element to sort the attacking targets in the process of fire strike.
In view of the unsatisfactory problem of operational scheme optimal selection and quantitative evaluation, an evaluation method of operational plan based on improved entropy weight matter-element model is proposed. By systematically demonstrating the evaluation index system and scoring method of operational scheme, the framework of operational scheme evaluation based on classical matter-element model is constructed, integrating the elementary correlation function and entropy weight method, the matter-element evaluation model of operational scheme is established. The anti entropy weight and Euclidean distance are used to improve the weight and distance calculation of the index respectively, four inspection modes of combined quantitative evaluation are formed,the scheme level is evaluated and the operational scheme is optimized. The effectiveness and practicability of the model method are verified by each other, providing a strong support for the commander's scientific decision-making.
In the context of increasing arctic temperature and melting ice, submarine activities in the Arctic Ocean and their ability to break the ice and rise to the surface have become a symbol of the military power competition among the countries around the Arctic. Based on the arctic meteorological and hydrological conditions, this paper constructs a natural environment risk index system to evaluate the submarine ice-breaking and floating, and uses TOPSIS method to carry out risk assessment and experimental division for the months in the Arctic Ocean. The assessment and analysis based on historical data show that the Bering Strait and Kara Sea area has a lower risk coefficient of ice-breaking and floating in winter, while the Beaufort Sea and Baffin Bay have a higher risk of floating in summer and autumn. Assessment results under different GHG emission scenarios based on CMIP6 model products show that the risk of ice breaking and floating in marginal sea areas is on a downward trend, and the risk in the Western Arctic hemisphere is generally higher than that in the Eastern Hemisphere. The research results are intended to provide scientific basis or decision-making reference for marine environmental protection of submarine activities or anti-submarine operations in the Arctic Ocean.
Aiming at the attitude control problem of underactuated quadrotor aircraft under modeling uncertainty and unknown external disturbance, an improved ADRC method is proposed. In this method, the new nonlinear function obtained by interpolation fitting are applied to ESO and NLSEF, and then an improved ADRC is obtained. Simulation results show that the control system of the quadrotor aircraft based on improved ADRC has good rapidity, strong anti-jamming and robustness.
The scientific and reasonable design of force scale structure is the basic guarantee for realizing the modernization of military organization form and improving combat effectiveness. This paper mainly aims at the demand of software tools and models to design and develop the auxiliary solution to the problem of force planning. By using the basic theory and mature methods in the field of military operations research, this paper details the analysis and demonstration process of the demonstration force scale structure, and then designs the function and structure of the demonstration support system, and advances the main functions and implementation methods such as target scheme design and implementation scheme preparation line description. The demonstration support system can provide standardized demonstration process, quantitative calculation and summary display services for multi scheme groups to work together, and improve the scientificity, standardization and work efficiency of argumentation.
In order to evaluate the calculation accuracy of various elements of setting height method, the working principle of setting height method is introduced, the influence of target height, firing distance and setting error on the calculation accuracy of various elements are analyzed by simulation method, and the calculation accuracy of various elements are compared with that of direct aiming method. The simulation results show that the various element errors of the setting height method increase with the increase of the setting error, decrease with the increase of the firing distance, and first increase and then decrease with the increase of the target height. When the shipborne radar is affected by sea clutter and other factors, the pitch angle measurement error is too large, the setting height method can significantly improve the calculation accuracy of the various elements.
Aiming at the problem of small ship course detection based on optical remote sensing images of fleet sailing, the Hough transformation is adopted and improved based on the ship size to identify and detect the small ship course characteristic effectively while the interference from sea noise wave is reduced obviously. Moreover with the combination of Hough transformation and CNN, the neural network is trained and tested by the reasonable design of the classification and the construction of dataset for ship course detection. The result shows that with the method it gives, the small ship target course can be detected effectively while the accuracy is improved that gives an important supporting for ship movement and situation estimating and fleet form recognition.
Realizable k-ε two equation model and SST k-ω two equations model are widely used in aerodynamic simulation. In this paper, the adaptability of these two models in solving this kind of complex shape projectile is analyzed through the aerodynamic simulation of supersonic cannon launched modified tail projectile. Firstly, the mathematical analysis of the two models is carried out, and then the simulation results are analyzed. The results show that the turbulent kinetic energy and turbulent energy dissipation rate of SST k-ω two equation model are more stable than that of realizable k-ε two equation model, and it is more suitable for calculating the aerodynamic characteristics of modified tail projectile.
Aiming at the decision-making problem of aviation intercepting enemy aircraft's take-off timing, the main process of aviation interception operations when defending important targets is analyzed, and the timeline of intercepting enemy aircraft is clarified. The calculation model of take-off window is built, which considers interception line characteristics, attack process of ammunition, and enemy aircraft's attack direction and so on. The influence of interception distance and incoming direction of enemy aircraft on the take-off window of the intercepting aircraft is simulated. The results show that incoming direction of enemy aircraft has a greater impact, and the take-off window is the smallest when enemy aircraft coming from other side of the airport to protecting objectives; the take-off window will shrink as the interception distance increases, and which will be close when the enemy aircraft is in a certain sector and the interception distance exceeds a certain value.
The traditional air defense operations take the command center as the information center, which has poor flexibility and easy to destroy, and low efficiency of information transmission between nodes. Therefore, it is not suitable for large-scale information interaction in the future network operations. In order to realize the decentralized communication network combining with the development of military equipment and the actual requirements of future wars, an air defense weapon information interaction system based on Data-Distribution-Service (DDS) is proposed. Will be centered on the platform of the air defense weapon system is decoupled into a distributed node function, system of destroyed node to evaluate the consequences of loss of function, distribution of share current operational communication domain with the same or similar function of the subsystem of resource, through the Data Centric Publish Subscribe (DCPS) communication mechanism to realize the combat information on-demand matching distribution and dynamic join nodes, through the service quality management to reduce the transmission bandwidth consumption, improve the transmission efficiency effectively. The test meets the requirements of technical indicators, and the signal processing time is controlled at the level of microseconds, which provides a reference for aviation, navigation, military and other real-time systems with high reliability requirements.
As a regional international organization, in order to effectively solve the semantic interoperability conflicts in the process of public service implementation, the European Commission has implemented the interoperability solutions for European Public Administrations,Businesses and Citizens(ISA) and formulated the core vocabulary specification, aiming to ensure the interoperability of e-government services between different member states and regions by building a unified data semantic specification. According to the abstract degree of objective things, EU e-government data model divided into core data model, domain data model and information exchange data model. The core data model consists of core business, core location, core person, core criterion and core evidence vocabulary, core public organization vocabulary, core public service vocabulary and application profile. The core vocabulary is extensible and has five mapping relationships with other data models, including exact match, close match, related match, broad match and narrow match. It can be applied to EU e-government data publishing, information exchange, data integration and data modeling. The basic application process includes requirement collection, information modeling, business rule definition, syntax binding (or creation building new syntax), data model mapping and generation.
