CN113038614A - Intelligent spectrum management and control framework based on spectrum knowledge graph - Google Patents
Intelligent spectrum management and control framework based on spectrum knowledge graph Download PDFInfo
- Publication number
- CN113038614A CN113038614A CN202110257241.4A CN202110257241A CN113038614A CN 113038614 A CN113038614 A CN 113038614A CN 202110257241 A CN202110257241 A CN 202110257241A CN 113038614 A CN113038614 A CN 113038614A
- Authority
- CN
- China
- Prior art keywords
- spectrum
- knowledge
- management
- intelligent
- frequency
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Granted
Links
- 238000001228 spectrum Methods 0.000 title claims abstract description 344
- 238000000034 method Methods 0.000 claims abstract description 21
- 238000007726 management method Methods 0.000 claims description 96
- 238000005516 engineering process Methods 0.000 claims description 13
- 230000009471 action Effects 0.000 claims description 11
- 230000008569 process Effects 0.000 claims description 11
- 238000012544 monitoring process Methods 0.000 claims description 8
- 238000010276 construction Methods 0.000 claims description 7
- 230000000694 effects Effects 0.000 claims description 5
- 238000013499 data model Methods 0.000 claims description 3
- 230000004927 fusion Effects 0.000 claims description 3
- 230000007246 mechanism Effects 0.000 claims description 2
- 239000011159 matrix material Substances 0.000 description 16
- 238000004891 communication Methods 0.000 description 9
- 230000001149 cognitive effect Effects 0.000 description 7
- 230000005540 biological transmission Effects 0.000 description 5
- 238000004364 calculation method Methods 0.000 description 5
- 238000004422 calculation algorithm Methods 0.000 description 4
- 238000010586 diagram Methods 0.000 description 4
- 230000003068 static effect Effects 0.000 description 4
- 210000004556 brain Anatomy 0.000 description 3
- 230000002452 interceptive effect Effects 0.000 description 3
- 238000005065 mining Methods 0.000 description 3
- 230000006399 behavior Effects 0.000 description 2
- 239000003795 chemical substances by application Substances 0.000 description 2
- 230000007547 defect Effects 0.000 description 2
- 238000009826 distribution Methods 0.000 description 2
- 230000003203 everyday effect Effects 0.000 description 2
- 239000002360 explosive Substances 0.000 description 2
- 230000006870 function Effects 0.000 description 2
- 239000005433 ionosphere Substances 0.000 description 2
- 238000013507 mapping Methods 0.000 description 2
- 238000013441 quality evaluation Methods 0.000 description 2
- 238000003860 storage Methods 0.000 description 2
- 230000009897 systematic effect Effects 0.000 description 2
- 238000012549 training Methods 0.000 description 2
- 230000007704 transition Effects 0.000 description 2
- 241000282412 Homo Species 0.000 description 1
- 206010063385 Intellectualisation Diseases 0.000 description 1
- 238000010521 absorption reaction Methods 0.000 description 1
- 238000001720 action spectrum Methods 0.000 description 1
- 230000006978 adaptation Effects 0.000 description 1
- 238000004458 analytical method Methods 0.000 description 1
- 238000013459 approach Methods 0.000 description 1
- 238000013473 artificial intelligence Methods 0.000 description 1
- 230000008859 change Effects 0.000 description 1
- 238000012512 characterization method Methods 0.000 description 1
- 230000003930 cognitive ability Effects 0.000 description 1
- 238000007405 data analysis Methods 0.000 description 1
- 238000013135 deep learning Methods 0.000 description 1
- 238000001514 detection method Methods 0.000 description 1
- 238000011161 development Methods 0.000 description 1
- 238000011156 evaluation Methods 0.000 description 1
- 238000002474 experimental method Methods 0.000 description 1
- 238000000605 extraction Methods 0.000 description 1
- 229910052500 inorganic mineral Inorganic materials 0.000 description 1
- 239000011707 mineral Substances 0.000 description 1
- 238000005457 optimization Methods 0.000 description 1
- 230000008447 perception Effects 0.000 description 1
- 238000013139 quantization Methods 0.000 description 1
- 230000005855 radiation Effects 0.000 description 1
- 238000007670 refining Methods 0.000 description 1
- 230000004044 response Effects 0.000 description 1
- 238000004088 simulation Methods 0.000 description 1
- 238000000638 solvent extraction Methods 0.000 description 1
- 230000003595 spectral effect Effects 0.000 description 1
- 230000007480 spreading Effects 0.000 description 1
- 238000003892 spreading Methods 0.000 description 1
- 238000013179 statistical model Methods 0.000 description 1
- XLYOFNOQVPJJNP-UHFFFAOYSA-N water Substances O XLYOFNOQVPJJNP-UHFFFAOYSA-N 0.000 description 1
Images
Classifications
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04W—WIRELESS COMMUNICATION NETWORKS
- H04W72/00—Local resource management
- H04W72/04—Wireless resource allocation
- H04W72/044—Wireless resource allocation based on the type of the allocated resource
- H04W72/0453—Resources in frequency domain, e.g. a carrier in FDMA
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04B—TRANSMISSION
- H04B17/00—Monitoring; Testing
- H04B17/30—Monitoring; Testing of propagation channels
- H04B17/382—Monitoring; Testing of propagation channels for resource allocation, admission control or handover
-
- Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
- Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
- Y02D—CLIMATE CHANGE MITIGATION TECHNOLOGIES IN INFORMATION AND COMMUNICATION TECHNOLOGIES [ICT], I.E. INFORMATION AND COMMUNICATION TECHNOLOGIES AIMING AT THE REDUCTION OF THEIR OWN ENERGY USE
- Y02D30/00—Reducing energy consumption in communication networks
- Y02D30/70—Reducing energy consumption in communication networks in wireless communication networks
Landscapes
- Engineering & Computer Science (AREA)
- Computer Networks & Wireless Communication (AREA)
- Signal Processing (AREA)
- Physics & Mathematics (AREA)
- Electromagnetism (AREA)
- Information Retrieval, Db Structures And Fs Structures Therefor (AREA)
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
Abstract
The invention provides an intelligent spectrum management and control framework based on a spectrum knowledge graph. The device comprises a map layer, an equipment layer and a scene layer. The map layer is a driving kernel of an intelligent spectrum management and control framework, namely a multi-domain associated spectrum knowledge map. The device layer is an execution unit of an intelligent spectrum management and control framework and mainly refers to intelligent frequency utilization devices configured with spectrum knowledge maps. The scene layer is an application presentation of the intelligent spectrum management and control framework. The intelligent frequency management center issues a frequency management task for the intelligent frequency utilization equipment in the scene, the intelligent frequency utilization equipment realizes a set frequency spectrum management and control target, and simultaneously reports information to the intelligent frequency management center. The intelligent frequency management center can also expand and perfect the spectrum knowledge graph according to diversified scenes and tasks. The method can realize modeling representation of the multivariate relation in the spectrum space, improve the intelligent level of spectrum management and control in a complex environment, and become a new tool in the field of future spectrum management and control.
Description
Technical Field
The invention belongs to the field of cognitive radio of wireless communication technology, and particularly relates to a spectrum knowledge graph and an intelligent spectrum management and control framework based on the spectrum knowledge graph.
Background
The electromagnetic spectrum is an electromagnetic wave family which is continuously arranged according to the frequency of electromagnetic waves, is an ideal medium for wireless information transmission, and is also one of the main activity spaces of human society in the information age and the intelligent age. The frequency range of the electromagnetic spectrum is generally considered to be 3Hz to 3000GHz, while the frequency range of electromagnetic waves that humans are currently able to use is mainly concentrated in the range of 30Hz to 40GHz, mostly below 6 GHz. Seemingly abundant electromagnetic spectrum resources are actually as important but scarce national strategic resources as water, forests, land, and mineral deposits. In recent years, with the continuous emergence of emerging technologies and services such as mobile internet, internet of things, big data, robots and the like, the number of various frequency-using devices and systems such as communication, sensing and the like is increased explosively, and how to reasonably configure and efficiently utilize limited electromagnetic spectrum resources is of great importance.
The spectrum management and control refers to planning, organizing, coordinating and controlling the use of electromagnetic spectrum by comprehensively using the means of administration, technology, engineering and the like so as to avoid mutual interference among frequency-using equipment, systems and services. Early spectrum management and control mainly depends on manual establishment of spectrum policies and frequency utilization rules, is realized by uniformly dividing frequency bands and assigning frequencies for frequency utilization equipment, systems and services, and is suitable for the conditions of limited frequency utilization requirements and relatively simple electromagnetic environment. The static management and control mode of the strip division emphasizes planning distribution and passive response, has high labor cost, lower management and control efficiency and poorer timeliness, is difficult to adapt to explosive growth of application frequency requirements and rapid change of electromagnetic environment, and can also cause the problems of unbalanced frequency spectrum use, lower frequency spectrum utilization rate and the like.
To solve the above problem, a spectrum hole can be used to realize dynamic spectrum access. Cognitive radio is a key technology for realizing dynamic spectrum access, and provides an ability for unauthorized users or cognitive users to share wireless spectrum resources with authorized users in an opportunistic manner. Under the condition that part of frequency spectrum is fixedly divided to authorized users, the cognitive users can sense the electromagnetic environment, detect unused frequency spectrum (namely frequency spectrum holes), estimate channel state information and predict channel capacity, select the optimal available frequency from the unused frequency spectrum and access the optimal available frequency, and realize frequency spectrum sharing with the authorized users. In this process, a cognitive user is an agent that is endowed with the capabilities of observation, learning, adaptation, decision making, and the like, as well as reconfigurability to support the transmission and reception of signals on different frequencies. This also establishes the main framework of "spectrum sensing-spectrum decision-spectrum sharing-spectrum movement" for the intellectualization of spectrum management.
Further, the explosive development of artificial intelligence technology brings new opportunities for the revolution of the spectrum management and control mode, and emerging technologies such as deep learning, group intelligence, block chains and the like show superiority in the aspects of spectrum data analysis, control channel allocation, anti-cheating decision and the like, so that the capability of spectrum opportunity discovery and utilization is continuously improved. Spectrum regulation is undergoing a transition from manual/manual to machine automated/autonomous intelligence, from static closed allocation to dynamic open sharing, from centralized unified assignment to distributed autonomous coordination, but the following challenges are still faced in this process:
firstly, the frequency spectrum space modeling representation mode is single, and the complicated electromagnetic environment is difficult to adapt. Currently, a spectrum situation is often depicted by a spectrogram, and time-frequency-space distribution of available spectrum resources is mainly concerned during modeling, and the spectrum situation is specifically represented by a busy-idle state, radiation power, an access protocol, a modulation mode and the like of a spectrum. With the increasing types and number of frequency-using devices/systems, the electromagnetic spectrum space is increasingly complex and complicated, and develops into a complex system which is composed of multiple subjects, multiple factors and multiple variables and is used for mutual input and output. The current spectrum modeling and characterization mode is difficult to clarify the multivariate relation among all main bodies in a spectrum space and the deep influence of the main bodies by factors and variables, and lacks the refining of systematic spectrum knowledge.
Secondly, the spectrum management and control mode has strong dependence on human experience, and the automatic and intelligent spectrum management and control effect is difficult to realize. The traditional spectrum management and control mode of static partitioning relies on manual operation and human experience. In the dynamic cooperative intelligent spectrum management and control mode, although each link of spectrum sensing, spectrum prediction, spectrum decision and the like has a technical solution for different optimization targets, the input and the output of an intelligent algorithm among the links still depend on manual connection and supervision, many operation skills and practical experiences are only stored in the brains of frequency management personnel, frequency equipment cannot understand the semantics of data flowing in a spectrum cognitive loop, thinking cannot be performed by combining the skill experiences on the basis of data calculation, and automatic spectrum sharing is not realized.
And thirdly, the frequency spectrum control efficiency is low, and the requirements of precise and real-time frequency spectrum control are difficult to meet. The existing spectrum control method mainly comprises the steps of establishing a statistical model from spectrum data, and mining statistical rules to perform spectrum prediction and spectrum decision, wherein the model-driven methods have inherent contradictions that model complexity, accuracy, resolvability and the like are difficult to reconcile. Moreover, for different spectrum management and control scenes, due to the lack of systematic spectrum knowledge, the generalization capability of the existing model is poor. In addition, for a large amount of optimized calculation in spectrum management and control, the formats of diversified spectrum data are difficult to unify, the calculation capability of spectrum management and control is limited, the results of spectrum prediction and spectrum decision are often lagged, and the timeliness is poor.
In summary, in order to promote the spectrum management and control to transition from a static, inefficient manual mode as a main mode to a dynamic, precise intelligent mode, exploring a new spectrum management and control mode is a difficult problem to be solved urgently at present.
Disclosure of Invention
The invention provides an intelligent spectrum management and control framework based on a spectrum knowledge graph, and aims to apply the knowledge graph theory and technology to spectrum management and control, define the concept, knowledge system and representation method of the spectrum knowledge graph, construct the intelligent spectrum management and control framework based on the spectrum knowledge graph, and discuss typical applications such as frequency use recommendation, spectrum search, spectrum question and answer and the like based on the spectrum knowledge graph.
The intelligent spectrum management and control framework based on the spectrum knowledge graph comprises: the spectrum knowledge graph is used as a graph layer of a driving kernel, the intelligent frequency-using equipment is used as an equipment layer of an execution unit, and the spectrum management and control application is used as an actually presented scene layer.
The spectrum layer is a multi-domain associated spectrum knowledge spectrum, original data is used for initial construction of the spectrum knowledge spectrum, monitoring data, image information and a control log which are obtained by frequency equipment through spectrum sensing, decision and action are input as new data to be used for updating the spectrum knowledge spectrum, and the spectrum knowledge spectrum enables heterogeneous spectrum data and model/expert experience to be effectively fused and converged into a knowledge base of the frequency equipment;
the device layer is an intelligent frequency-using device configured with a spectrum knowledge graph, and the intelligent frequency-using device has the following capabilities: spectrum sensing of knowledge map energizing, spectrum decision of knowledge map energizing and spectrum action of knowledge map energizing;
the scene layer is a diversified spectrum management and control application scene, the intelligent frequency management center determines a spectrum management and control target under the guidance of scene knowledge contained in a spectrum knowledge map, selects a proper technical method according to scene-technology knowledge in the spectrum knowledge map, and issues a frequency management task to intelligent frequency equipment; the intelligent frequency utilization equipment completes the frequency management task in a distributed cooperation mode and reports information to the intelligent frequency management center.
Preferably, the construction process of the spectrum knowledge graph specifically comprises the following steps:
firstly, acquiring knowledge of unstructured spectrum data with wide sources and various types, wherein the data sources comprise practice experience of professional spectrum management and control personnel, a textual spectrum management and control policy, frequency equipment information, a formatted spectrum management and control log and a spectrum monitoring data report, and a spectrum situation image; in the process, the spectrum knowledge system provides basis and reference for knowledge acquisition, namely an instance layer is created by a mode layer, wherein the mode layer stores refined concepts or entities and relationship types between the concept entities; the example layer corresponds to a specific example object extracted from the actual data and a relation between the example objects; then, the extracted spectrum knowledge is subjected to knowledge fusion with the structured spectrum data to form a unified spectrum knowledge representation; storing the unified spectrum knowledge to form an available spectrum knowledge map; and (4) complementing and perfecting missing knowledge links in the constructed spectrum knowledge graph through knowledge reasoning, and finally serving for specific application.
Preferably, in the construction process of the spectrum knowledge graph, example knowledge is described in a form of a triple, namely < head entity, relation, tail entity >, which is marked as (h, r, t), wherein the head entity h is a concept or an entity in a knowledge system, and the tail entity t is the concept or the entity, or is an attribute value field of the entity; correspondingly, the relation r is a predicate or a top-bottom relation of the connector and the tail entity, or an attribute relation between the description entity object and the attribute value field thereof; a complete triple is called a fact, denoted as F ═ h, r, t, from the perspective of the graph model, all entities and fields in the spectrum knowledge graph correspond to nodes in the graph, and relationships between entities or attributes of entities correspond to edges in the graph, forming a mesh graph data model, closely organizing the dispersed spectrum knowledge together.
Frequency utilization recommendation in the frequency spectrum management and control is to take frequency utilization equipment as a user, take frequency spectrum resources as articles, and analyze frequency utilization requirements and behavior patterns of the frequency utilization equipment by mining the evolution characteristics of the frequency spectrum resources, so as to recommend available and well-utilized frequency spectrum resources for the frequency utilization equipment. The frequency recommendation is helpful for breaking through the limit of sensing capability, reducing the cost of sensing time, making up the defects of a spectrum availability model in the aspects of complexity, accuracy, interpretability and the like, providing active, predictive and enhanced information support for spectrum decision, assisting frequency management personnel in optimizing a spectrum management and control policy, and improving the spectrum management and control efficiency. The method can realize modeling representation of the multivariate relation in the spectrum space, improve the intelligent level of spectrum management and control in a complex environment, and become a new tool in the field of future spectrum management and control.
Drawings
FIG. 1 is a schematic diagram of a spectral knowledge graph construction process.
Fig. 2 is a diagram of a spectrum knowledge graph-based intelligent spectrum management framework.
Fig. 3 is an exemplary diagram of an air-to-ground spectrum sharing knowledge graph.
Fig. 4 is a schematic diagram of link quality matrix completion based on a spectrum knowledge graph.
Detailed Description
The spectrum knowledge graph is a new concept for representing a complex spectrum space, and is defined as follows:
the spectrum knowledge graph is a domain knowledge graph, represents spectrum resources, concepts in frequency space of frequency-using equipment and the like and complex relations among entities by integrating multi-source heterogeneous spectrum data, realizes representation, extraction, storage and reasoning of spectrum knowledge, and serves the automatic, intelligent and accurate requirements of future spectrum management and control.
The spectrum knowledge refers to experience accumulated in spectrum management and control practice, established rules or various facts and information appearing in spectrum management and control. The spectrum knowledge system refines and summarizes the spectrum knowledge and determines a basic framework for describing the spectrum knowledge. From the aspect of a control object, spectrum control is developed around important scarce spectrum resources and frequency utilization equipment with various types; in terms of control content, scene configuration of spectrum control and specific tasks under each scene are different; from the management and control means, the spectrum management and control relates to a plurality of intelligent algorithms facing different task targets. Based on this, the main concepts in the spectrum knowledge system can be divided into four categories, namely resources, devices, scenes and technologies, and the concept category includes specific entities which have multiple attributes.
The process of constructing a spectrum knowledge graph according to the present invention is shown in fig. 1. Firstly, knowledge acquisition is carried out on unstructured spectrum data with wide sources and various types, and the data sources comprise practice experience of professional spectrum management and control personnel, a textual spectrum management and control policy, frequency equipment information, a formatted spectrum management and control log, a spectrum monitoring data report, even a spectrum situation image and the like. In the process, the spectrum knowledge system provides basis and reference for knowledge acquisition, namely an instance layer is created by a mode layer, wherein the mode layer stores refined concepts or entities and relationship types among the refined concepts or entities, and the instance layer corresponds to specific instance objects extracted from actual data and relationships of the specific instance objects. And then, the extracted spectrum knowledge is subjected to knowledge fusion with the structured spectrum data to form a unified spectrum knowledge representation. And storing the spectrum knowledge, namely forming a usable spectrum knowledge map. And the missing knowledge links in the established spectrum knowledge graph can be complemented and perfected through knowledge reasoning, and finally, the spectrum knowledge graph is used for specific application.
The spectrum knowledge graph described in the invention adopts a triple form to describe example knowledge, namely < head entity, relation, tail entity >, which is marked as (h, r, t). The head entity h is generally a concept or an entity in a knowledge system, and the tail entity t can be the concept or the entity or an attribute value field of the entity; correspondingly, the relation r can be predicates or upper and lower relations of the connectors and the tail entities, and can also describe the attribute relation between the entity object and the attribute value field thereof. A complete triplet is called a fact, denoted F ═ h, r, t. From the perspective of the graph model, all entities and fields in the spectrum knowledge graph correspond to nodes in the graph, relationships among the entities or attributes of the entities correspond to edges in the graph, a mesh graph data model is formed, and dispersed spectrum knowledge is closely organized together.
The intelligent spectrum management and control framework based on the spectrum knowledge graph is shown in fig. 2. The map layer, the equipment layer and the scene layer are sequentially arranged from left to right. The future spectrum management and control is a mode of combining centralized control and distributed autonomy, the spectrum management and control center is an intelligent control center with the capabilities of storage, calculation, understanding, thinking and the like, the frequency-using equipment is an intelligent agent with the capabilities of perception, understanding, memory, thinking and the like, and the spectrum knowledge graph is a knowledge base in the spectrum management and control center and the brain of the frequency-using equipment. Under the guidance of the spectrum knowledge graph, the intelligent frequency management center can determine a spectrum management and control target according to a scene where the intelligent frequency management center is located, select a proper intelligent method, and issue a frequency management task for the intelligent frequency utilization equipment in the scene, and the intelligent frequency utilization equipment can automatically call an intelligent algorithm according to the received frequency management task and the role of the intelligent frequency management task in the scene, so that a determined spectrum management and control target is realized, and information is reported to the intelligent frequency management center. The intelligent frequency-using equipment queries, retrieves, infers and updates the spectrum knowledge graph through spectrum sensing, spectrum decision, spectrum action and the like in the process of completing the frequency management task, and the intelligent frequency management center expands and perfects the spectrum knowledge graph according to diversified scenes and tasks.
The map layer is a driving kernel of an intelligent spectrum management and control framework, namely a multi-domain associated spectrum knowledge map. And the spectrum control policy, expert experience, part of historical spectrum control logs, historical spectrum monitoring data and other raw data are used for initial construction of the spectrum knowledge graph. Monitoring data, image information and control logs obtained by the frequency utilization equipment through spectrum sensing, decision making and action are used as new data input for updating the spectrum knowledge graph. The spectrum knowledge graph enables heterogeneous spectrum data and model/expert experience to be effectively fused and converged into a knowledge base of frequency-using equipment.
The device layer is an execution unit of an intelligent spectrum management and control framework and mainly refers to intelligent frequency utilization devices configured with spectrum knowledge maps. In addition to the cognitive ability and reconfigurability of the frequency utilization equipment in the traditional cognitive radio technology, the frequency utilization equipment in the intelligent spectrum management and control framework also needs to have the ability of understanding, memorizing and thinking, and the spectrum knowledge graph plays a role in the brain of the frequency utilization equipment. The intelligent frequency utilization equipment mainly completes three tasks: spectrum sensing of knowledge map energizing, spectrum decision of knowledge map energizing and spectrum action of knowledge map energizing, wherein the knowledge map energizing is embodied in spectrum sensing, spectrum decision, spectrum action spectrum knowledge map application:
1) spectrum sensing with knowledge map energizing: the data information obtained by sensing is extracted and processed to form new spectrum knowledge which is stored in a spectrum knowledge map, and the existing knowledge in the spectrum knowledge map can be used as prior information to assist frequency equipment in extracting knowledge based on an attention mechanism.
2) Spectrum decision of knowledge map energizing: the spectrum knowledge graph can be inquired by the frequency utilization equipment to obtain the relation fact between the entities, and the spectrum knowledge graph can also be used for reasoning the relation or the fact between the entities to assist the decision making.
3) Spectrum action with knowledge map energizing: the fact that the spectrum is shared or moved will further update the spectrum knowledge graph, and inference based on the spectrum knowledge graph can also evaluate the effect of the spectrum action.
The scene layer is an application presentation of the intelligent spectrum management and control framework. Facing to diversified application scenarios such as spectrum order management and control, spectrum countermeasure management and control, spectrum sharing management and control, an intelligent frequency management center firstly determines a target of spectrum management and control under the guidance of scene knowledge contained in a spectrum knowledge graph, for example, in an air-to-ground spectrum sharing management and control scenario in fig. 2, communication of an unmanned aerial vehicle network and intelligent frequency utilization equipment is interfered by an interference station, and the target of spectrum management and control is to realize spectrum safety sharing. The intelligent frequency management center selects a proper technical method according to scene-technology knowledge in the spectrum knowledge map, and issues a frequency management task to the intelligent frequency equipment. The intelligent frequency utilization equipment completes the frequency management task in a distributed cooperation mode and reports information to the intelligent frequency management center.
Example 1: a specific embodiment of the spectrum knowledge graph according to the present invention is described below, which lists main items of a spectrum knowledge system, types of relationships between entities, and ranges of relationship functions, as shown in tables 1 and 2, and shows a representation of the spectrum knowledge graph by taking an air-to-ground spectrum sharing scenario in spectrum management and control as an example, as shown in fig. 3.
Table 1 example spectrum knowledge system
Table 2 example of relationships in a spectrum knowledge system
At the device-resource level, each oval node in fig. 3 is an example of a concept in a resource or device class; each rectangular node is a field that describes an attribute value. Relationships between entities or attributes of entities are represented by directed edges between nodes. In the scene-technical level, the relationship between the scene type, the specific task and the key technology is shown in fig. 2, and the scene-technical level and the device-resource level are linked by including the device and the resource in the scene task. In addition, the nodes of the knowledge graph are set as entities or the fields can be flexibly adjusted according to actual situations, such as "direct sequence spread spectrum", if the nodes are only used for describing the modulation mode attribute of the signal transmitted by the equipment, the nodes can be set as fields, and if the nodes in the knowledge graph also relate to specific attribute parameters of the modulation mode, such as spreading code length, code rate and the like, the nodes are set as entities.
Example 2: a specific embodiment of the intelligent spectrum management and control framework based on the spectrum knowledge graph is described below, and takes intelligent frequency recommendation as an example to embody a knowledge guiding role of the spectrum knowledge graph in intelligent spectrum management and control.
Frequency utilization recommendation in the frequency spectrum management and control is to take frequency utilization equipment as a user, take frequency spectrum resources as articles, and analyze frequency utilization requirements and behavior patterns of the frequency utilization equipment by mining the evolution characteristics of the frequency spectrum resources, so that available and well-utilized frequency spectrum resources are recommended for the frequency utilization equipment. The frequency recommendation is helpful for breaking through the limit of sensing capability, reducing the cost of sensing time, making up the defects of a spectrum availability model in the aspects of complexity, accuracy, interpretability and the like, providing active, predictive and enhanced information support for spectrum decision, assisting frequency management personnel in optimizing a spectrum management and control policy, and improving the spectrum management and control efficiency.
In communication, the interactive data between the user and the article can be expressed as the evaluation of the transmission quality of the frequency spectrum resource by the frequency utilization equipment. For example, in a short-wave communication system, since an ionosphere reflecting a short-wave signal is a typical time-varying transmission medium, the ionosphere has different reflection and absorption capacities for radio waves of different frequencies, characteristics thereof are also changed by multiple factors such as day and night alternation, season alternation, sun and night periods, geographical positions and the like, and a channel is seriously interfered, so that communication is extremely unstable. The short-wave frequency equipment usually performs link quality analysis on a preset frequency set, and then adaptively selects communication frequency according to a link quality evaluation result. This makes it easy to conceive of short-wave frequency recommendation using interactive data such as link quality evaluation.
On the one hand, however, when the short-wave frequency device obtains link quality through a probing channel, it may be affected by interference or noise to cause a failure in probing; on the other hand, the detection capability of frequency-using devices is limited, and in practical systems, frequency-using devices mainly detect near their usual frequencies. Therefore, the link quality matrix is incomplete, and the short-wave frequency recommendation also faces the problem of sparse interactive data. The embodiment completes the link quality matrix by using the spectrum knowledge graph, and serves for short-wave frequency recommendation.
For spectrum control professionals, link quality is affected by main factors, which time-varying trend is presented generally, the practical experience belongs to spectrum knowledge, the spectrum knowledge can be represented through a spectrum knowledge map, the guiding effect of the spectrum knowledge in intelligent spectrum control is exerted, and frequency use recommendation driven by data and experience in a mixed mode is achieved.
Specifically, in this embodiment, a resource class entity in the spectrum knowledge graph is further refined based on embodiment 1, and link quality is modeled as an entity. Consider first "knowledge 1": the link quality is mainly affected by frequency, time, transmission distance and other factors, and the numerical value of the link quality, the corresponding time, frequency, the link starting point, the link ending point and the like are modeled as the attributes of the entity. As shown in fig. 4, the above-mentioned graph modeling approach has been formed for any one of the known elements in the link quality matrix (q)i,j,ValueIs,23)、(qi,j,TimeIs,06/01_08)、(qi,j,FrequencyIs,f2)、(qi,jTrlocalities, a) and the like; for missing elements in the link quality matrix, since the position of the element in the matrix is unambiguous, i.e. information such as time, frequency, start and end of the link, etc. is known, only (q) is formedm,n,TimeIs,06/01_03)、(qm,n,FrequencyIs,f5)、(qm,nTrlocalities, a), the link quality matrix completion problem is converted into a triplet (q) representing the link quality valuem,nValueis,? ) The link prediction problem of (1). On this basis, the following "knowledge 2" can be further considered: link quality is greatly affected by day and night alternation, and for the same link, link quality at the same time on different days should have similarity, and triplets such as (06/01_07, TimeSimiar, 06/11_07) can be added as known information.
Because the spectrum knowledge graph has a large number of many-to-one relations (N-to-1), the spectrum knowledge graph is considered to be represented and learned by adopting a TransR model, and all entities, relations and vector representation of corresponding mapping matrixes in the spectrum are learned. Then, the link prediction problem (q) is addressedm,nValueis,? ) For each missing tail entity in the triple to be predicted, all entities in the corresponding category in the knowledge graph are used as candidate item calculation score functions, and the entity with the highest score rank is used as a prediction result, namely the value of the missing element in the link quality matrix.
The experimental data of the embodiment are from simulation data of short-wave medium-and-long-term forecasting software VOACAP. The data set 1 simulates the receiving signal-to-noise ratio (assuming that the receiving signal-to-noise ratio is used as a link quality index) of a short-wave communication link from Nanjing to Xiamen, and records the link quality in the whole period of the first day (9 days in total) of every ten days from 6 months to 8 months in 2020, wherein the link quality is recorded once per hour every day, the link frequency is set to 9 frequency points which are uniformly distributed in a short-wave frequency band, and the signal-to-noise ratio value is preprocessed and then subjected to 5-level uniform quantization, so that a complete link quality matrix is a 9 x 216 numerical matrix. The data set 2 simulates the receiving signal-to-noise ratio of a short wave communication link from Nanjing to Haikou, the recording time is the link quality of the 2015 year in the whole period of 15 days every month, the link quality is recorded once every hour every day, the link frequency is set to 9 frequency points, the signal-to-noise ratio value is quantized in 5 levels, and then the complete link quality matrix of the data set 2 is a 9 x 288 numerical matrix. For the two data sets, the proportion of missing elements in the link quality matrix is set to be 80%. In this embodiment, a TransR model is used to perform vector representation learning on the constructed spectrum knowledge graph, wherein the dimensions of the entity vector, the relationship vector and the mapping matrix are set to 100, 100 and 100 × 100, respectively, the learning rate of the stochastic gradient descent algorithm is set to 0.02, the marginal parameter is set to 4, the scale of each batch of training data is set to 5000, and the training frequency is set to 5000.
The results of the experiments are shown in table 3,
TABLE 3 accuracy of Link quality matrix completion
Therefore, the frequency spectrum knowledge graph can play a knowledge guide role in frequency use recommendation, and the performance of link quality matrix completion can be effectively improved by adding knowledge.
Claims (8)
1. An intelligent spectrum management and control framework based on spectrum knowledge graph, the intelligent spectrum management and control framework comprising: the method comprises the steps of taking a spectrum knowledge graph as a graph layer of a driving kernel, taking intelligent frequency utilization equipment as an equipment layer of an execution unit, and taking spectrum management and control application as an actually presented scene layer;
the spectrum layer is a multi-domain associated spectrum knowledge spectrum, original data is used for initial construction of the spectrum knowledge spectrum, monitoring data, image information and a control log which are obtained by frequency equipment through spectrum sensing, decision and action are input as new data to be used for updating the spectrum knowledge spectrum, and the spectrum knowledge spectrum enables heterogeneous spectrum data and model/expert experience to be effectively fused and converged into a knowledge base of the frequency equipment;
the device layer is an intelligent frequency-using device configured with a spectrum knowledge graph, and the intelligent frequency-using device has the following capabilities: spectrum sensing of knowledge map energizing, spectrum decision of knowledge map energizing and spectrum action of knowledge map energizing;
the scene layer is a diversified spectrum management and control application scene, the intelligent frequency management center determines a spectrum management and control target under the guidance of scene knowledge contained in a spectrum knowledge map, selects a proper technical method according to scene-technology knowledge in the spectrum knowledge map, and issues a frequency management task to intelligent frequency equipment; the intelligent frequency utilization equipment completes the frequency management task in a distributed cooperation mode and reports information to the intelligent frequency management center.
2. The intelligent spectrum management and control framework based on spectrum knowledge graph according to claim 1, wherein the spectrum knowledge graph is constructed by the following specific steps:
firstly, acquiring knowledge of unstructured spectrum data with wide sources and various types, wherein the data sources comprise practice experience of professional spectrum management and control personnel, a textual spectrum management and control policy, frequency equipment information, a formatted spectrum management and control log and a spectrum monitoring data report, and a spectrum situation image; in the process, the spectrum knowledge system provides basis and reference for knowledge acquisition, namely an instance layer is created by a mode layer, wherein the mode layer stores refined concepts or entities and relationship types between the concept entities; the example layer corresponds to a specific example object extracted from the actual data and a relation between the example objects; then, the extracted spectrum knowledge is subjected to knowledge fusion with the structured spectrum data to form a unified spectrum knowledge representation; storing the unified spectrum knowledge to form an available spectrum knowledge map; and (4) complementing and perfecting missing knowledge links in the constructed spectrum knowledge graph through knowledge reasoning, and finally serving for specific application.
3. The intelligent spectrum management and control framework based on the spectrum knowledge graph according to claim 2, wherein example knowledge is described in a triple form in the construction process of the spectrum knowledge graph, namely < head entity, relationship, tail entity >, which is denoted as (h, r, t), wherein the head entity h is a concept or an entity in a knowledge system, and the tail entity t is a concept or an entity, or is an attribute value field of an entity; correspondingly, the relation r is a predicate or a top-bottom relation of the connector and the tail entity, or an attribute relation between the description entity object and the attribute value field thereof; a complete triple is called a fact, denoted as F ═ h, r, t, from the perspective of the graph model, all entities and fields in the spectrum knowledge graph correspond to nodes in the graph, and relationships between entities or attributes of entities correspond to edges in the graph, forming a mesh graph data model, closely organizing the dispersed spectrum knowledge together.
4. The intelligent spectrum management and control framework based on spectrum knowledge graph of claim 1, wherein spectrum sensing enabled by knowledge graph spectrum: the data information obtained by sensing is extracted and processed to form new spectrum knowledge, the new spectrum knowledge is stored in a spectrum knowledge graph, the existing knowledge in the spectrum knowledge graph is used as prior information, and frequency equipment is assisted to extract knowledge based on an attention mechanism.
5. The intelligent spectrum management and control framework based on spectrum knowledge graph of claim 1, wherein knowledge graph spectrum-energized spectrum decision: the spectrum knowledge graph can be inquired by the frequency utilization equipment to obtain the relation fact between the entities, and the spectrum knowledge graph can also be used for reasoning the relation or the fact between the entities to assist the decision making.
6. The intelligent spectrum management and control framework based on spectrum knowledge graph of claim 1, wherein spectrum actions enabled by knowledge graph are: the fact that the spectrum is shared or moved will further update the spectrum knowledge graph, and inference based on the spectrum knowledge graph can also evaluate the effect of the spectrum action.
7. The spectrum intellectual property graph based on intelligent spectrum management and control framework of claim 1, wherein the raw data at least comprises spectrum management and control policies, expert experience, and part of historical spectrum management and control logs and historical spectrum monitoring data.
8. The framework of claim 1, wherein the application scenarios of spectrum knowledge graph management comprise spectrum order management, spectrum countermeasure management, and spectrum sharing management.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202110257241.4A CN113038614B (en) | 2021-03-09 | 2021-03-09 | Intelligent frequency spectrum management and control system based on frequency spectrum knowledge graph |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202110257241.4A CN113038614B (en) | 2021-03-09 | 2021-03-09 | Intelligent frequency spectrum management and control system based on frequency spectrum knowledge graph |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN113038614A true CN113038614A (en) | 2021-06-25 |
| CN113038614B CN113038614B (en) | 2023-04-18 |
Family
ID=76467459
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202110257241.4A Active CN113038614B (en) | 2021-03-09 | 2021-03-09 | Intelligent frequency spectrum management and control system based on frequency spectrum knowledge graph |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN113038614B (en) |
Cited By (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN113992288A (en) * | 2021-10-29 | 2022-01-28 | 南京航空航天大学 | 6G communication network-oriented super-large-scale spectrum knowledge graph construction method |
| CN114880484A (en) * | 2022-05-11 | 2022-08-09 | 军事科学院系统工程研究院网络信息研究所 | Satellite communication frequency-orbit resource map construction method based on vector mapping |
| CN116340533A (en) * | 2023-02-23 | 2023-06-27 | 中国人民解放军军事科学院系统工程研究院 | Satellite-borne electromagnetic spectrum big data intelligent processing system based on knowledge graph |
Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN106304095A (en) * | 2016-08-31 | 2017-01-04 | 中国人民解放军重庆通信学院 | A kind of dynamic spectrum management system and method |
| CN107040325A (en) * | 2017-05-19 | 2017-08-11 | 中国电子科技集团公司第七研究所 | Frequency spectrum resource dynamic allocation method, device and system based on strategy |
| CN111078868A (en) * | 2019-06-04 | 2020-04-28 | 中国人民解放军92493部队参谋部 | Knowledge graph analysis-based equipment test system planning decision method and system |
-
2021
- 2021-03-09 CN CN202110257241.4A patent/CN113038614B/en active Active
Patent Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN106304095A (en) * | 2016-08-31 | 2017-01-04 | 中国人民解放军重庆通信学院 | A kind of dynamic spectrum management system and method |
| CN107040325A (en) * | 2017-05-19 | 2017-08-11 | 中国电子科技集团公司第七研究所 | Frequency spectrum resource dynamic allocation method, device and system based on strategy |
| CN111078868A (en) * | 2019-06-04 | 2020-04-28 | 中国人民解放军92493部队参谋部 | Knowledge graph analysis-based equipment test system planning decision method and system |
Non-Patent Citations (3)
| Title |
|---|
| 孙佳琛: "频谱知识图谱：面向未来频谱管理的智能引擎", 《通信学报》 * |
| 田家强等: "动态频谱管理技术发展研究", 《通信技术》 * |
| 荣华: "战术通信网频谱管理系统设计", 《通讯世界》 * |
Cited By (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN113992288A (en) * | 2021-10-29 | 2022-01-28 | 南京航空航天大学 | 6G communication network-oriented super-large-scale spectrum knowledge graph construction method |
| CN114880484A (en) * | 2022-05-11 | 2022-08-09 | 军事科学院系统工程研究院网络信息研究所 | Satellite communication frequency-orbit resource map construction method based on vector mapping |
| CN116340533A (en) * | 2023-02-23 | 2023-06-27 | 中国人民解放军军事科学院系统工程研究院 | Satellite-borne electromagnetic spectrum big data intelligent processing system based on knowledge graph |
| CN116340533B (en) * | 2023-02-23 | 2023-09-01 | 中国人民解放军军事科学院系统工程研究院 | Satellite-borne electromagnetic spectrum big data intelligent processing system based on knowledge graph |
Also Published As
| Publication number | Publication date |
|---|---|
| CN113038614B (en) | 2023-04-18 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN113038614B (en) | Intelligent frequency spectrum management and control system based on frequency spectrum knowledge graph | |
| Zhang et al. | Game theory based real‐time shop floor scheduling strategy and method for cloud manufacturing | |
| CN114896472B (en) | Knowledge graph machine reasoning system and method based on multi-source space-time data | |
| US20140324747A1 (en) | Artificial continuously recombinant neural fiber network | |
| CN110662245B (en) | Base station load early warning method and device based on deep learning | |
| Petrenko et al. | Problem of developing an early-warning cybersecurity system for critically important governmental information assets | |
| Colizzi et al. | Introduction to agricultural IoT | |
| CN112905340B (en) | System resource allocation method, device and equipment | |
| Khanum et al. | Towards a semantically enriched computational intelligence (SECI) framework for smart farming | |
| EP1650686A2 (en) | Method and system for simulating non-linear dependencies between physical entities and measured data on influencing factors, governed by a local/global modelling approach using objects with embedded probabilistic nets | |
| CN106910075B (en) | The intelligent processing system and method that client mobile communication is complained | |
| Rozhnov et al. | Investigation of the joint semantic environment for heterogeneous robotics | |
| CN102624782A (en) | Internal concentrated harmonization system/information and communication technology (ICH/ICT) information fusion basis of internet | |
| Mendula et al. | Energy-aware edge federated learning for enhanced reliability and sustainability | |
| Bryndin | Collaboration robots as digital doubles of person for communication in public life and space | |
| Tomforde | An architectural framework for self-configuration and self-improvement at runtime | |
| Wu et al. | Unifying principles of generalization: past, present, and future | |
| CN102710732A (en) | Internet holographic collaborative system information fusion foundation | |
| Hemamalini et al. | An analysis on multi-agent based distributed data mining system | |
| CN116782399A (en) | Rule-driven intelligent spectrum management method based on reasoning engine | |
| Guillaume et al. | Open source software for modelling using agro-environmental georeferenced data | |
| Díaz González | Federated Learning for Time Series Forecasting Using LSTM Networks: Exploiting Similarities Through Clustering | |
| Newhouse et al. | Integrating the Analytic Hierarchy Process (AHP) in process engineering for infrastructure Modelling and Simulation (M&S) | |
| Ташатов et al. | RESEARCH AND DEVELOPMENT METHODS OF A SOFTWARE SYSTEM FOR THE INTEGRATED ANALYSIS OF METEOROLOGICAL DATA | |
| Skobelev et al. | Methods and Tools for Designing of a Multi-service Platform for Agricultural Enterprises |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |