Spectrum map optimization method and device based on smooth constraint, equipment and storage medium
By introducing smooth constraints in spectrum map optimization, combining model-based and data-based methods to optimize the error of spectrum maps, the problems of difficulty in selecting models and high computing resources in the existing technology are solved, and more efficient spectrum map generation and A wider range of application scenarios.
Patent Information
- Application Number
- CN202411134824.8
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2024-08-19
- Publication Date
- 2024-11-15
Smart Images

Figure CN118972006A_ABST
Abstract
Description
Technical Field
[0001] The present application relates to the technical field of spectrum mapping, and more specifically, to a spectrum map optimization method, apparatus, device and storage medium based on smoothness constraints. Background Art
[0002] Spectrum maps have a wide range of applications in the field of wireless communications, such as network planning, interference coordination and suppression, power control, resource allocation, multi-hop routing, dynamic spectrum access, etc., and have become one of the key technologies for electromagnetic environment cognition. At present, spectrum map generation technology can be divided into three categories: model-based methods, data-based methods, and methods combining models and data. Model-based methods do not rely on spectrum sampling data, but require prior information, while data-based methods do not rely on signal prior information, but may have lower performance when the amount of spectrum sampling data is small.
[0003] The method of combining models and data to construct spectrum maps can combine the advantages of both methods. However, there are relatively few studies on the combination of models and data. The existing methods of combining models and data are difficult to select appropriate mathematical models for spectrum map estimation in specific scenarios, and have higher requirements for computing resources, which limits their application scenarios and scope. Summary of the invention
[0004] In view of at least one defect or improvement need of the prior art, the present invention provides a spectrum map optimization method, apparatus, device and storage medium based on smoothness constraints, which can solve at least one of the problems existing in the prior art.
[0005] To achieve the above object, according to a first aspect of the present invention, a spectrum map optimization method based on smoothness constraints is provided, the method comprising:
[0006] A model-based approach is used based on the prior information of the target area to obtain the estimated value of the global spectrum data;
[0007] Obtaining a quantized estimation error based on an estimation value of current spectrum data and an estimation value of global spectrum data;
[0008] Obtaining a regularized error based on signal strength values of the estimated spectrum data of each point in the target area and its surrounding points;
[0009] Obtaining a joint error based on the quantization estimation error and the regularization error;
[0010] A frequency spectrum map that minimizes the joint error is obtained, that is, an optimal frequency spectrum map of the target area.
[0011] Furthermore, in the above-mentioned spectrum map optimization method based on smoothness constraint, the step of obtaining the quantized estimation error specifically includes:
[0012] The difference between the actual spectrum data obtained by sensor sampling and the estimated value of the global spectrum data is recorded as The root mean square error is used to estimate the quantization error,
[0013]
[0014] Among them, κ represents the estimated value of the current spectrum data, which is updated in real time under the gradient descent method, and H(κ) represents the real spectrum data intensity value. represents a spectrum detection data set, that is, a set of spectrum detection data corresponding one-to-one to discrete points in the set of deployed sensor locations, and N represents the number of sensors.
[0015] Furthermore, in the above-mentioned spectrum map optimization method based on smoothness constraint, the obtaining of regularization error specifically includes:
[0016] The regularized error is expressed as have
[0017]
[0018] in, represents the estimated value of the global spectrum data, (l x ,l y ) represents the position coordinates of the grid points in the target area, (L x ,L y ) represents the maximum number of grid points in the target area in the x-axis and y-axis directions, (l x +1,l y ) means (l x ,l y )The position coordinate index on the right side, (l x ,l y +1) means (l x ,l y ) is the position coordinate index on the upper side, and p represents the signal strength at the position of the corresponding point.
[0019] Furthermore, in the above-mentioned spectrum map optimization method based on smoothness constraint, the step of obtaining the joint error specifically includes:
[0020] The joint error is expressed as have
[0021]
[0022] Among them, λ and μ represent weight factors.
[0023] Furthermore, in the above-mentioned spectrum map optimization method based on smoothness constraint, the step of obtaining the spectrum map that minimizes the joint error, that is, the optimal spectrum map of the target area, specifically includes:
[0024] The optimal spectrum map of the target area is represented as κ * ,have
[0025]
[0026] That is, the optimal spectrum map of the target area is the spectrum map that minimizes the joint error.
[0027] Furthermore, in the above-mentioned spectrum map optimization method based on smoothness constraint, the step of obtaining the spectrum map that minimizes the joint error specifically includes performing optimization based on a random search method to obtain the spectrum map that minimizes the joint error.
[0028] Furthermore, in the above spectrum map optimization method based on smoothness constraints, the random search method for optimization specifically includes:
[0029] The spectrum map output by the i-th iteration optimization is represented as κ i , initialize the spectrum map is the spectrum map estimated in the first stage, then the update formula of iterative optimization is expressed as follows
[0030]
[0031] in, represents the gradient of the loss function and α is the learning rate.
[0032] According to a second aspect of the present invention, there is also provided a spectrum map optimization device based on smoothness constraints, comprising:
[0033] A model building module, used for obtaining a global spectrum data estimation value by adopting a model-based method based on prior information of a target area;
[0034] A first acquisition module is used to acquire a quantized estimation error based on real spectrum data obtained by sensor sampling and the estimated value of the global spectrum data;
[0035] A second acquisition module is used to acquire a regularization error based on the signal strength values of the estimated spectrum data of each point in the target area and its surrounding points;
[0036] A third acquisition module, configured to acquire a joint error based on the quantization estimation error and the regularization error;
[0037] The optimization module is used to obtain the spectrum map that minimizes the joint error, that is, the optimal spectrum map of the target area.
[0038] According to the third aspect of the present invention, there is also provided a spectrum map optimization device based on smoothness constraints, which includes at least one processing unit and at least one storage unit, wherein the storage unit stores a computer program, and when the computer program is executed by the processing unit, the processing unit executes the steps of any one of the above-mentioned methods.
[0039] According to the fourth aspect of the present invention, a storage medium is also provided, which stores a computer program that can be executed by a spectrum map optimization device based on a smoothness constraint. When the computer program runs on the spectrum map optimization device based on a smoothness constraint, the spectrum map optimization device based on a smoothness constraint executes the steps of any one of the above-mentioned methods.
[0040] In general, the above technical solutions conceived by the present invention can achieve the following beneficial effects compared with the prior art:
[0041] The present invention provides a spectrum map optimization method based on smoothness constraints, which realizes the complementary advantages of model and data by combining the two stages of model-based estimation and data-based optimization. By making full use of prior information through model-based estimation, an estimated spectrum map is obtained, and a quantitative estimation error representing accuracy and a regularization error representing smoothness are introduced to optimize the joint error. This can improve the accuracy of the spectrum map and obtain the optimal spectrum map. Since the architecture is universal, it can be applied to different scenarios. BRIEF DESCRIPTION OF THE DRAWINGS
[0042] In order to more clearly illustrate the technical solutions in the embodiments of the present application, the drawings required for use in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.
[0043] Figure 1 A schematic diagram of a flow chart of a spectrum map optimization method based on smoothness constraints provided in an embodiment of the present application;
[0044] Figure 2 A schematic diagram of the IRT method provided in an embodiment of the present application before and after combined with smoothness-constrained data-driven optimization;
[0045] Figure 3 A schematic diagram of the RadioUnet method provided in an embodiment of the present application before and after combined with smoothness-constrained data-driven optimization;
[0046] Figure 4Schematic diagram of error variation curves of different construction methods at different sampling rates provided in the embodiments of the present application. DETAILED DESCRIPTION
[0047] In order to make the purpose, technical solutions and advantages of the present invention more clearly understood, the present invention is further described in detail below in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present invention and are not intended to limit the present invention. In addition, the technical features involved in the various embodiments of the present invention described below can be combined with each other as long as they do not conflict with each other.
[0048] The terms "first", "second", "third", etc. in the specification and claims of this application and the above-mentioned drawings are used to distinguish different objects, rather than to describe a specific order. In addition, the terms "including" and "having" and any variations thereof are intended to cover non-exclusive inclusions. For example, a process, method, system, product or device that includes a series of steps or units is not limited to the listed steps or units, but optionally includes steps or units that are not listed, or optionally includes other steps or units inherent to these processes, methods, products or devices.
[0049] Figure 1 A schematic diagram of a flow chart of a spectrum map optimization method based on smoothness constraints provided in an embodiment of the present application is shown in FIG. Figure 1 As shown, a spectrum map optimization method based on smoothness constraints provided in an embodiment of the present application specifically includes:
[0050] 101. A model-based approach is used based on the prior information of the target area to obtain an estimate of the global spectrum data.
[0051] Specifically, M is defined as the geographical environment information such as buildings and elevations in the target area, and P T , L T are the signal propagation prior information such as transmitter power and position, respectively. Based on the model construction method φ, the global spectrum data estimation value can be obtained Right now
[0052]
[0053] Where S is a set of coordinate positions, representing a set of L discrete coordinates in the target area, and the model-based method φ is based on the data optimization method and the spatial interpolation method ψ. The model-based method may include a propagation model or a ray tracing method, which may be selected according to actual engineering needs and is not specifically limited in this application.
[0054] 102. Obtain a quantized estimation error based on an estimation value of current spectrum data and an estimation value of global spectrum data.
[0055] Specifically, the number of spectrum sensors in the target area is N (N < M), and it is defined that is the set of N sensor deployment positions, that is, the sampling network coordinate set, is a subset of S, and there is
[0056]
[0057] where represents the coordinate of the nth grid sampling position, that is, the coordinate of the nth spectrum sensor.
[0058] represents the deployment position set corresponding sampling spectrum data set, that is, the set of spectrum detection data corresponding one-to-one to the discrete points in the deployment position set, the spectrum data in is in one-to-one correspondence with the coordinate points in is a subset of κ, and there is
[0059]
[0060] where represents the received signal strength value sampled at the coordinate position.
[0061] Let H(·) denote the sampling operator determined by the following formula,
[0062]
[0063] where represents the received signal strength value detected at the position s n There is indicating that the data sampled from the true spectrum data is the measured value corresponding to the sensor deployment position, κ true represents the unknown true spectrum map data.
[0064] Based on the data optimization method using the sampled spectrum data to optimize the spectrum map estimated by the model-based estimation method φ to further improve the accuracy of the spectrum map, and finally obtain the optimal spectrum map κ * , there is
[0065]
[0066] where the data optimization method The specific selection can be made according to actual engineering needs, and this application does not make any specific limitations here.
[0067] Considering that the sparse spectrum data obtained by sensor sampling represents the true value, in order to optimize the spectrum map, it is necessary to minimize the estimation error and reduce the difference between the sampled data and the spectrum map. The quantitative estimation error can be obtained based on the true spectrum data obtained by sensor sampling and the estimated value of the global spectrum data. The method for obtaining the quantitative estimation error can be selected according to the actual engineering needs, and this application does not make specific limitations here.
[0068] Optionally, the root mean square error (RMSE) can be used to quantify the estimation error, and the difference between the actual spectrum data obtained by sensor sampling and the estimated value of the global spectrum data is recorded as The root mean square error is used to estimate the quantization error,
[0069]
[0070] Among them, κ represents the estimated value of the current spectrum data, which is updated in real time under the gradient descent method. That is to say, κ is equal to the result of model estimation at the beginning. After optimization, the optimal spectrum map is output; H(κ) represents the real spectrum data intensity value, represents a spectrum detection data set, that is, a set of spectrum detection data corresponding one-to-one to discrete points in the set of deployed sensor locations, and N represents the number of sensors.
[0071] 103. Obtain a regularized error based on signal strength values of the estimated spectrum data of each point in the target area and its surrounding points.
[0072] Specifically, considering that wireless signals gradually attenuate in space, the spectrum map area without building occlusion is smooth. In the field of image restoration, total variation regularization can process the smooth area in the image by applying regularization constraints, and the image is smoothed while retaining previous information. Based on the signal strength value of the estimated spectrum data of each point in the target area and its surrounding points, the regularization error is obtained. The method for obtaining the regularization error can be selected according to the actual engineering needs, and this application does not make specific limitations here.
[0073] Optionally, the regularized error is expressed as By introducing the regularization error To improve the smoothness of the RF map, minimize the total variation of the signal, where the total variation is the sum of the absolute differences between adjacent elements. In the discrete case, the regularization error is expressed as have
[0074]
[0075] in, represents the estimated value of the global spectrum data, (l x ,l y ) represents the position coordinates of the grid points in the target area, (L x ,L y ) represents the maximum number of grid points in the target area in the x-axis and y-axis directions, (l x +1,l y ) means 9l x ,l y )The position coordinate index on the right side, (l x ,l y +1) means (l x ,l y ) is the position coordinate index on the upper side, and p represents the signal strength at the position of the corresponding point.
[0076] 104. Based on the quantized estimation error and the regularized error, a joint error is obtained.
[0077] Specifically, the joint error can be obtained by adjusting the proportion of the quantization estimation error and the regularization error.
[0078] Alternatively, the joint error can be expressed as have
[0079]
[0080] Among them, λ and μ represent weight factors, and the selection of λ and μ can be selected according to actual engineering needs, and this application does not make specific limitations here.
[0081] 105. Obtain a spectrum map that minimizes the joint error, that is, an optimal spectrum map of the target area.
[0082] Specifically, an optimization may be performed based on a random search method to obtain a spectrum map that minimizes the joint error. The present application does not specifically limit the optimization method.
[0083] Alternatively, the optimal spectrum map of the target area can be expressed as κ * ,have
[0084]
[0085] That is, the optimal spectrum map of the target area is the spectrum map that minimizes the joint error. The smallest spectrum map is the optimal spectrum map κ * .
[0086] Optionally, an optimization may be performed based on a random search method to obtain a frequency spectrum map that minimizes the joint error.
[0087] Optionally, the Adam optimization algorithm can be used for optimization, which is a stochastic objective function optimization method based on first-order gradients, implemented by adaptively estimating low-order moments. It is an extension of the stochastic gradient descent algorithm to minimize the loss function in the equation.
[0088] The spectrum map output by the i-th iteration optimization is represented as κ i , initialize the spectrum map is the spectrum map estimated in the first stage, then the update formula of iterative optimization is expressed as follows
[0089]
[0090] in, represents the gradient of the loss function and α is the learning rate.
[0091] A spectrum map optimization method based on smoothness constraints provided in an embodiment of the present application realizes complementary advantages of the model and data by combining the two stages of model-based estimation and data-based optimization. By making full use of prior information through model-based estimation, an estimated spectrum map is obtained, and a quantitative estimation error representing accuracy and a regularization error representing smoothness are introduced to optimize the joint error. This can improve the accuracy of the spectrum map and obtain the optimal spectrum map. Since its architecture is universal, it can be applied to different scenarios.
[0092] The spectrum map optimization method based on smoothness constraint proposed in the present invention is simulated and verified based on a specific embodiment. The following is the experimental setting and result display.
[0093] 1. Dataset and Experimental Setup
[0094] The experiment used the RaioMapSeer dataset, which includes city maps, transmitter locations, and simulated spectrum maps. The height of the transmitter and receiver is set to 1.5 meters, and the building height is 25 meters. Each map is in a 256×256m 2 The image is represented in a two-dimensional grid and converted into a morphological two-dimensional image.
[0095] The experiments consider scenarios with two sensor deployment methods: random deployment and grid-based uniform deployment. The measurements of the sensors are treated as sampled spectral data. The sampling rate is expressed as the ratio of the number of sensors to the number of grids in the target area. The sampling rate ranges from 5% to 25%.
[0096] The root mean square error (RMSE) is used to evaluate the accuracy of spectrum map construction, and the calculation formula is as follows:
[0097]
[0098] Among them, ‖·‖ means finding the L2 norm, and L is the cardinality of the sample point set, that is, the number of coordinate points in the set.
[0099] 2. Benchmark Algorithms
[0100] Two model-based methods are introduced as baseline algorithms:
[0101] ①IRT (Intelligent Ray Tracing). A three-dimensional ray tracing method for predicting path loss and delay spread in urban environments. It simulates the propagation path of electromagnetic waves based on precise physical principles and mathematical models.
[0102] ②RadioUnet. A deep learning-based radio map estimation method that uses a city map (i.e., the geometry of the urban environment) and transmitter locations as input. One version of this method is called RadioUNet. C , no spectrum sampling data is required, and this method can be classified as a model-based method.
[0103] Figure 2 and Figure 3 Results of different methods of constructing the spectrum map using a 15% sampling rate are shown. Figure 2 Depicted is the comparison between the IRT estimation results before and after the joint data-driven optimization, where the left side is the true RF map, the middle is the estimated RF map based on IRT, and the right side is the RF map constructed after the smoothness constraint-based data-driven optimization. Figure 3 The experimental results of using RadioUnet in the model estimation stage are shown, with the real RF map on the left, the estimated RF map based on RadioUnet in the middle, and the RF map constructed after data-driven optimization based on smoothness constraints on the right. Compared with the real spectrum map, it can be observed that the RF map generated after data-driven optimization shows enhanced refinement in specific details, especially in the areas where shadows and non-shadows intersect in the image.
[0104] Figure 4 The RMSE results are shown as the sampling rate increases considering different model methods and sensor deployment scenarios. Figure 4 (a) shows the error change curve based on the IRT method. Figure 4(b) shows the error change curve based on the RadioUnet method. The IRT and RadioUnet methods maintain a consistent error level because they do not rely on spectrum sampling data, which is presented as a horizontal line. As the sampling rate increases, the RMSE decreases, and when the sampling rate is fixed, the RMSE of uniform deployment is smaller than that of random deployment. When the sampling rate and sensor deployment scenario are the same, the RMSE of the RadioUnet method is smaller than that of the IRT method. In addition, when the model-based estimation method and the sampling rate are the same, the RMSE of the uniform deployment (UD) method is lower than that of the random deployment (RD) method, indicating that for the data-driven optimization method with smooth constraints designed by us, the uniform deployment (UD) method is more helpful than the random deployment (RD) method to improve the overall performance.
[0105] The present application also provides a spectrum map optimization device based on smoothness constraints, comprising:
[0106] A model building module, used for obtaining a global spectrum data estimation value by adopting a model-based method based on prior information of a target area;
[0107] A first acquisition module is used to acquire a quantized estimation error based on real spectrum data and global spectrum data estimation values obtained by sensor sampling;
[0108] A second acquisition module is used to acquire a regularization error based on the signal strength values of the estimated spectrum data of each point in the target area and its surrounding points;
[0109] A third acquisition module is used to acquire a joint error based on the quantized estimation error and the regularized error;
[0110] The optimization module is used to obtain the spectrum map that minimizes the joint error, that is, the optimal spectrum map of the target area.
[0111] The present application also provides a computer-readable storage medium on which a computer program is stored, and when the program is executed by a processor, the steps of the above method are implemented. The computer-readable storage medium may include, but is not limited to, any type of disk, including a floppy disk, an optical disk, a DVD, a CD-ROM, a micro drive, and a magneto-optical disk, a ROM, a RAM, an EPROM, an EEPROM, a DRAM, a VRAM, a flash memory device, a magnetic card or an optical card, a nanosystem (including a molecular memory IC), or any type of medium or device suitable for storing instructions and / or data.
[0112] It should be noted that, for the aforementioned method embodiments, for the sake of simplicity, they are all expressed as a series of action combinations, but those skilled in the art should be aware that the present application is not limited by the described order of actions, because according to the present application, certain steps can be performed in other orders or simultaneously. Secondly, those skilled in the art should also be aware that the embodiments described in the specification are all preferred embodiments, and the actions and modules involved are not necessarily required by the present application.
[0113] In the above embodiments, the description of each embodiment has its own emphasis. For parts that are not described in detail in a certain embodiment, reference can be made to the relevant descriptions of other embodiments.
[0114] In the several embodiments provided in the present application, it should be understood that the disclosed devices can be implemented in other ways. For example, the device embodiments described above are only schematic, such as the division of the units, which is only a logical function division. There may be other division methods in actual implementation, such as multiple units or components can be combined or integrated into another system, or some features can be ignored or not executed. Another point is that the mutual coupling or direct coupling or communication connection shown or discussed can be through some service interfaces, and the indirect coupling or communication connection of devices or units can be electrical or other forms.
[0115] The units described as separate components may or may not be physically separated, and the components shown as units may or may not be physical units, that is, they may be located in one place or distributed on multiple network units. Some or all of the units may be selected according to actual needs to achieve the purpose of the solution of this embodiment.
[0116] In addition, each functional unit in each embodiment of the present application may be integrated into one processing unit, or each unit may exist physically separately, or two or more units may be integrated into one unit. The above-mentioned integrated unit may be implemented in the form of hardware or in the form of software functional units.
[0117] If the integrated unit is implemented in the form of a software functional unit and sold or used as an independent product, it can be stored in a computer-readable memory. Based on this understanding, the technical solution of the present application is essentially or the part that contributes to the prior art or all or part of the technical solution can be embodied in the form of a software product, and the computer software product is stored in a memory, including a number of instructions to enable a computer device (which can be a personal computer, server or network device, etc.) to perform all or part of the steps of the method described in each embodiment of the present application. The aforementioned memory includes: U disk, read-only memory (ROM), random access memory (RAM), mobile hard disk, disk or optical disk and other media that can store program code.
[0118] A person of ordinary skill in the art can understand that all or part of the steps in the various methods of the above embodiments can be completed by instructing related hardware through a program, and the program can be stored in a computer-readable memory, which can include: a flash drive, a read-only memory (ROM), a random access memory (RAM), a magnetic disk or an optical disk, etc.
[0119] The above is only an exemplary embodiment of the present disclosure, and the scope of the present disclosure cannot be limited thereto. That is, any equivalent changes and modifications made according to the teachings of the present disclosure are still within the scope of the present disclosure. After considering the specification and practicing the disclosure here, those skilled in the art will easily think of the implementation scheme of the present disclosure. This application is intended to cover any modification, use or adaptation of the present disclosure, which follows the general principles of the present disclosure and includes common knowledge or customary technical means in the technical field not recorded in the present disclosure. The description and examples are regarded as exemplary only, and the scope and spirit of the present disclosure are defined by the claims.
[0120] The technical features of the above embodiments may be arbitrarily combined. To make the description concise, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.
[0121] It will be easily understood by those skilled in the art that the above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions and improvements made within the spirit and principles of the present invention should be included in the protection scope of the present invention.
Claims
1. A spectrum map optimization method based on smoothness constraints, characterized in that: include: A model-based approach is used based on the prior information of the target area to obtain the estimated value of the global spectrum data; Obtaining a quantized estimation error based on an estimation value of current spectrum data and an estimation value of global spectrum data; Obtaining a regularized error based on signal strength values of the estimated spectrum data of each point in the target area and its surrounding points; Obtaining a joint error based on the quantization estimation error and the regularization error; A frequency spectrum map that minimizes the joint error is obtained, that is, an optimal frequency spectrum map of the target area.
2. The spectrum map optimization method based on smoothness constraint according to claim 1, characterized in that: The obtaining of the quantization estimation error specifically includes: The difference between the estimated value of the current spectrum data and the estimated value of the global spectrum data is recorded as The root mean square error is used to estimate the quantization error, Among them, κ represents the estimated value of the current spectrum data, which is updated in real time under the gradient descent method, and H(κ) represents the real spectrum data intensity value. represents a spectrum detection data set, that is, a set of spectrum detection data corresponding one-to-one to discrete points in the set of deployed sensor locations, and N represents the number of sensors.
3. The spectrum map optimization method based on smoothness constraint according to claim 2, characterized in that: The obtaining of the regularization error specifically includes: The regularized error is expressed as have in, represents the estimated value of the global spectrum data, (l x ,l y ) represents the position coordinates of the grid points in the target area, (L x ,L y ) represents the maximum number of grid points in the target area in the x-axis and y-axis directions, (l x +1,l y ) means (l x ,l y )The position coordinate index on the right side, (l x ,l y +1) means (l x ,l y ) is the position coordinate index on the upper side, and p represents the signal strength at the position of the corresponding point.
4. The spectrum map optimization method based on smoothness constraint according to claim 3, characterized in that: The obtaining of the joint error specifically includes: The joint error is expressed as have Among them, λ and μ represent weight factors.
5. The spectrum map optimization method based on smoothness constraint according to claim 4, characterized in that: The obtaining of the spectrum map that minimizes the joint error, that is, the optimal spectrum map of the target area, specifically includes: The optimal spectrum map of the target area is represented as κ * ,have That is, the optimal spectrum map of the target area is the spectrum map that minimizes the joint error.
6. The spectrum map optimization method based on smoothness constraint according to claim 5, characterized in that: The obtaining of the frequency spectrum map that minimizes the joint error specifically includes performing optimization based on a random search method to obtain the frequency spectrum map that minimizes the joint error.
7. The spectrum map optimization method based on smoothness constraint according to claim 6, characterized in that: The random search method is used to find the optimal solution, specifically including: The spectrum map output by the i-th iteration optimization is represented as κ i , initialize the spectrum map is the spectrum map estimated in the first stage, then the update formula of iterative optimization is expressed as follows in, represents the gradient of the loss function and α is the learning rate.
8. A spectrum map optimization device based on smoothness constraints, characterized in that: include: A model building module, used for obtaining a global spectrum data estimation value by adopting a model-based method based on prior information of a target area; A first acquisition module is used to acquire a quantized estimation error based on real spectrum data obtained by sensor sampling and the estimated value of the global spectrum data; A second acquisition module is used to acquire a regularization error based on the signal strength values of the estimated spectrum data of each point in the target area and its surrounding points; A third acquisition module, configured to acquire a joint error based on the quantization estimation error and the regularization error; The optimization module is used to obtain the spectrum map that minimizes the joint error, that is, the optimal spectrum map of the target area.
9. A spectrum map optimization device based on smoothness constraints, characterized in that: The method comprises at least one processing unit and at least one storage unit, wherein the storage unit stores a computer program, and when the computer program is executed by the processing unit, the processing unit executes the steps of the method according to any one of claims 1 to 7.
10. A storage medium, characterized in that: It stores a computer program executable by a spectrum map optimization device based on smoothness constraints. When the computer program runs on the spectrum map optimization device based on smoothness constraints, the spectrum map optimization device based on smoothness constraints executes the steps of the method described in any one of claims 1 to 7.
