CN116975589B - Feature generation method, device, equipment and storage medium for network field intensity prediction - Google Patents
Feature generation method, device, equipment and storage medium for network field intensity predictionInfo
- Publication number
- CN116975589B CN116975589B CN202210796088.7A CN202210796088A CN116975589B CN 116975589 B CN116975589 B CN 116975589B CN 202210796088 A CN202210796088 A CN 202210796088A CN 116975589 B CN116975589 B CN 116975589B
- Authority
- CN
- China
- Prior art keywords
- feature
- network field
- field intensity
- sample
- intensity prediction
- 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.)
- Active
Links
Classifications
-
- G—PHYSICS
- G06—COMPUTING OR CALCULATING; COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q10/00—Administration; Management
- G06Q10/04—Forecasting or optimisation specially adapted for administrative or management purposes, e.g. linear programming or "cutting stock problem"
-
- G—PHYSICS
- G06—COMPUTING OR CALCULATING; COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/08—Learning methods
-
- G—PHYSICS
- G06—COMPUTING OR CALCULATING; COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q50/00—Information and communication technology [ICT] specially adapted for implementation of business processes of specific business sectors, e.g. utilities or tourism
- G06Q50/08—Construction
Landscapes
- Engineering & Computer Science (AREA)
- Business, Economics & Management (AREA)
- Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- Human Resources & Organizations (AREA)
- Economics (AREA)
- Strategic Management (AREA)
- General Physics & Mathematics (AREA)
- General Health & Medical Sciences (AREA)
- Marketing (AREA)
- Tourism & Hospitality (AREA)
- General Business, Economics & Management (AREA)
- Health & Medical Sciences (AREA)
- Biomedical Technology (AREA)
- Computing Systems (AREA)
- Artificial Intelligence (AREA)
- Primary Health Care (AREA)
- Biophysics (AREA)
- Computational Linguistics (AREA)
- Data Mining & Analysis (AREA)
- Evolutionary Computation (AREA)
- Molecular Biology (AREA)
- Life Sciences & Earth Sciences (AREA)
- General Engineering & Computer Science (AREA)
- Mathematical Physics (AREA)
- Software Systems (AREA)
- Development Economics (AREA)
- Game Theory and Decision Science (AREA)
- Entrepreneurship & Innovation (AREA)
- Operations Research (AREA)
- Quality & Reliability (AREA)
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
Abstract
本发明属于人工智能技术领域，公开了一种网络场强预测的特征生成方法、装置、设备及存储介质。该方法包括：获取网络场强预测样本；对所述网络场强预测样本进行标签插值处理生成栅格化标签特征值；根据所述网络场强预测样本确定基础特征样本，并根据所述基础特征样本生成基础特征值；根据所述栅格化标签特征值和基础特征值生成网络场强预测特征，所述网络场强预测特征用于对网络场强预测模型进行训练。通过上述方法，将多项特征值应用于特征向量的提取中，生成用于移动网络场强预测的特征，输入神经网络中进行模型训练，相较于简单的特征提取过程，其预测精度有大幅提高，模型训练效果更好。
This invention belongs to the field of artificial intelligence technology and discloses a method, apparatus, device, and storage medium for generating features for network field strength prediction. The method includes: acquiring network field strength prediction samples; performing label interpolation on the network field strength prediction samples to generate rasterized label feature values; determining basic feature samples based on the network field strength prediction samples and generating basic feature values based on the basic feature samples; generating network field strength prediction features based on the rasterized label feature values and the basic feature values, wherein the network field strength prediction features are used to train a network field strength prediction model. Through the above method, multiple feature values are applied to feature vector extraction to generate features for mobile network field strength prediction, which are then input into the neural network for model training. Compared to a simple feature extraction process, the prediction accuracy is significantly improved, and the model training effect is better.
Description
Technical Field
The present invention relates to the field of artificial intelligence technologies, and in particular, to a method, an apparatus, a device, and a storage medium for generating characteristics of network field strength prediction.
Background
The wireless network planning is to search for a network deployment scheme which meets the requirements and has reasonable cost by using methods such as service scene positioning, user density analysis and the like. Link budget is a key basis for coverage planning, and requires the selection of a propagation model for the appropriate physical environment. However, the existing propagation model does not mention a method for generating a grid label in the process of rasterizing sample data, a simple mean method is generally adopted to calculate a grid label value, and the obtained grid label value and the error between samples are larger, so that the model training is greatly influenced, and the training effect is poor.
The foregoing is provided merely for the purpose of facilitating understanding of the technical solutions of the present invention and is not intended to represent an admission that the foregoing is prior art.
Disclosure of Invention
The invention mainly aims to provide a characteristic generation method, device, equipment and storage medium for network field intensity prediction, and aims to solve the technical problem of poor training effect of a model in the prior art.
In order to achieve the above object, the present invention provides a feature generation method for network field strength prediction, the method comprising the steps of:
Acquiring a network field intensity prediction sample;
Performing label interpolation processing on the network field intensity prediction sample to generate a rasterized label characteristic value;
determining a basic characteristic sample according to the network field intensity prediction sample, and generating a basic characteristic value according to the basic characteristic sample;
Generating a network field intensity prediction feature according to the rasterized tag characteristic value and the basic characteristic value, wherein the network field intensity prediction feature is used for training a network field intensity prediction model.
Optionally, before generating the network field strength prediction feature according to the rasterized tag feature value and the basic feature value, the method further includes:
performing representation learning processing on the network field intensity prediction sample to obtain preset ground object relation characteristics;
the generating a network field intensity prediction feature according to the rasterized tag feature value and the basic feature value comprises the following steps:
And generating a network field intensity prediction feature according to the preset ground object relation feature, the rasterized tag feature value and the basic feature value.
Optionally, the performing a representation learning process on the network field intensity prediction sample to obtain a preset feature relation feature includes:
Determining feature visibility parameters and feature category parameters according to the network field intensity prediction samples;
and determining preset feature relation characteristics according to the feature visibility parameters and the feature category parameters.
Optionally, before generating the network field strength prediction feature according to the rasterized tag feature value and the basic feature value, the method includes:
Determining the district topography type characteristic according to the network field intensity prediction sample;
The generating network field intensity prediction features according to the rasterized tag feature values and the basic feature values further comprises:
generating a network field intensity prediction feature according to the cell topography type feature, the rasterized tag feature value and the basic feature value;
Or alternatively, the first and second heat exchangers may be,
And generating a network field intensity prediction feature according to the preset ground object relation feature, the district topography type feature, the rasterized tag feature value and the basic feature value.
Optionally, the determining the characteristic of the district topography type according to the network field intensity prediction sample includes:
acquiring preset type information;
determining a current sample cell parameter according to the network field intensity prediction sample;
and matching the current sample cell parameters with preset type information to obtain the cell topography type characteristics.
Optionally, the performing label interpolation processing on the network field intensity prediction sample to generate a rasterized label feature value includes:
rasterizing the network field intensity prediction sample to obtain a transmitting end coordinate and grid information of each signal grid;
and carrying out label interpolation on each signal grid according to the transmitting end coordinates and the grid information to obtain a rasterized label characteristic value.
Optionally, the determining a basic feature sample according to the network field intensity prediction sample, and generating a basic feature value according to the basic feature sample, includes:
Screening the network field intensity prediction samples to obtain basic characteristic samples, wherein the basic characteristic samples at least comprise height difference data of a sending end and a receiving end, distance data of the sending end and the receiving end and angle data of the sending end and a signal direction;
And generating a basic characteristic value according to the basic characteristic sample.
In addition, in order to achieve the above object, the present invention also provides a feature generation device for network field intensity prediction, the feature generation device for network field intensity prediction includes:
the acquisition module is used for acquiring a network field intensity prediction sample;
the processing module is used for carrying out label interpolation processing on the network field intensity prediction sample to generate a rasterized label characteristic value;
the processing module is further used for determining a basic characteristic sample according to the network field intensity prediction sample and generating a basic characteristic value according to the basic characteristic sample;
The processing module is further used for generating a network field intensity prediction feature according to the rasterized tag characteristic value and the basic characteristic value, and the network field intensity prediction feature is used for training a network field intensity prediction model.
In order to achieve the above object, the invention also proposes a network field strength prediction feature generation device comprising a memory, a processor and a network field strength prediction feature generation program stored on the memory and executable on the processor, the network field strength prediction feature generation program being configured to implement the steps of the network field strength prediction feature generation method as described above.
In addition, in order to achieve the above object, the present invention also proposes a storage medium having stored thereon a network field intensity predicted feature generation program which, when executed by a processor, implements the steps of the network field intensity predicted feature generation method as described above.
The method comprises the steps of obtaining a network field intensity prediction sample, carrying out label interpolation processing on the network field intensity prediction sample to generate a rasterized label characteristic value, determining a basic characteristic sample according to the network field intensity prediction sample, generating a basic characteristic value according to the basic characteristic sample, and generating a network field intensity prediction characteristic according to the rasterized label characteristic value and the basic characteristic value, wherein the network field intensity prediction characteristic is used for training a network field intensity prediction model. By the method, multiple feature values are applied to the extraction of the feature vectors, the features for predicting the field intensity of the mobile network are generated, and the features are input into the neural network for model training.
Drawings
FIG. 1 is a schematic diagram of a feature generation device for network field strength prediction for a hardware operating environment according to an embodiment of the present invention;
FIG. 2 is a flow chart of a first embodiment of a feature generation method for network field strength prediction according to the present invention;
FIG. 3 is a schematic block diagram illustrating an embodiment of a feature generation method for network field strength prediction according to the present invention;
FIG. 4 is a schematic diagram of distribution of sampling points according to an embodiment of a feature generation method for network field strength prediction according to the present invention;
FIG. 5 is a schematic view of a profile of engineering parameters according to an embodiment of a feature generation method for network field strength prediction according to the present invention;
FIG. 6 is a schematic view of a direction angle of an embodiment of a feature generation method for network field strength prediction according to the present invention;
FIG. 7 is a flow chart of a second embodiment of a feature generation method for network field strength prediction according to the present invention;
FIG. 8 is a schematic diagram illustrating EMD histogram calculation according to an embodiment of a feature generation method for network field strength prediction of the present invention;
fig. 9 is a block diagram showing the construction of a first embodiment of the characteristic generating apparatus for network field strength prediction according to the present invention.
The achievement of the objects, functional features and advantages of the present invention will be further described with reference to the accompanying drawings, in conjunction with the embodiments.
Detailed Description
It should be understood that the specific embodiments described herein are for purposes of illustration only and are not intended to limit the scope of the invention.
Referring to fig. 1, fig. 1 is a schematic structural diagram of a feature generating device for predicting network field intensity of a hardware running environment according to an embodiment of the present invention.
As shown in fig. 1, the characteristic generating device for network field strength prediction may comprise a processor 1001, such as a central processing unit (Central Processing Unit, CPU), a communication bus 1002, a user interface 1003, a network interface 1004, a memory 1005. Wherein the communication bus 1002 is used to enable connected communication between these components. The user interface 1003 may include a Display, an input unit such as a Keyboard (Keyboard), and the optional user interface 1003 may further include a standard wired interface, a wireless interface. The network interface 1004 may optionally include a standard wired interface, a Wireless interface (e.g., a Wireless-Fidelity (Wi-Fi) interface). The Memory 1005 may be a high-speed random access Memory (Random Access Memory, RAM) Memory or a stable Non-Volatile Memory (NVM), such as a disk Memory. The memory 1005 may also optionally be a storage device separate from the processor 1001 described above.
Those skilled in the art will appreciate that the structure shown in fig. 1 does not constitute a limitation of the feature generation apparatus for network field strength prediction, and may include more or fewer components than shown, or may combine certain components, or may be a different arrangement of components.
As shown in fig. 1, a memory 1005, which is a storage medium, may include an operating system, a network communication module, a user interface module, and a feature generation program for network field strength prediction.
In the network field intensity predicted feature generating device shown in fig. 1, the network interface 1004 is mainly used for carrying out data communication with a network server, the user interface 1003 is mainly used for carrying out data interaction with a user, and the processor 1001 and the memory 1005 in the network field intensity predicted feature generating device can be arranged in the network field intensity predicted feature generating device, and the network field intensity predicted feature generating device calls a network field intensity predicted feature generating program stored in the memory 1005 through the processor 1001 and executes the network field intensity predicted feature generating method provided by the embodiment of the invention.
The embodiment of the invention provides a characteristic generation method for network field intensity prediction, and referring to fig. 2, fig. 2 is a flow chart diagram of a first embodiment of the characteristic generation method for network field intensity prediction.
In this embodiment, the method for generating the characteristics of the network field intensity prediction includes the following steps:
and S10, acquiring a network field intensity prediction sample.
It should be noted that, the execution body of the embodiment is an intelligent terminal including a feature extraction function, where the intelligent terminal may be a server, a computer, or other intelligent devices with the same or similar functions as the server.
It should be noted that, in this embodiment, the feature extraction process of sample data before being applied to the network field intensity prediction model is optimized to ensure that the feature data input to the model to be trained is more comprehensive and reliable, and in order to overcome the defects in the prior art, the embodiment provides a feature generation method for mobile network field intensity prediction, which realizes feature extraction of original sample data and generates feature engineering which can be applied to a deep learning network and has strong correlation with mobile network field intensity prediction. This is because feature engineering plays an important role in deep learning neural networks in order to extract features from raw data to the maximum for algorithms and models. In the past, most of feature selection is carried out by adopting a single method, the correlation between feature engineering and a label value is low, and the model training effect is poor, but the feature engineering comprehensively analyzes the discrete degree of the feature, the correlation between the feature engineering and a predicted variable and the importance degree of the predicted variable, so that the process of feature selection is more reasonable, the training result can be more accurately close to the predicted value, and the good training effect is achieved.
In a specific implementation, as shown in fig. 3, the feature vector is sent to the input layer of the network based on the deep learning neural network model, for example, the layers are fully connected in the deep learning neural network, and assuming that the model is provided with 3 hidden layers, each neuron of the next layer is connected with all neurons of the current input layer, in this way, features are integrated, and field strength predicted values applied to the mobile network are obtained through the 3 hidden layers until the output layer. The training process based on the above example network model proposes a preferred sample data feature extraction process.
The network field strength prediction sample is a sample containing a large amount of data, for example, a spatial position of a base station in a map, a spatial position of each signal detection point, a topography in a detection area, a type of each block (the sample is divided into a grid form), a signal condition of each signal cell, and the like.
And S20, performing label interpolation processing on the network field intensity prediction sample to generate a rasterized label characteristic value.
It should be noted that the tag interpolation process is divided into two parts, firstly, a grid number and a grid center coordinate are required to be set, a sample area is set with the base station as the center according to the position of the base station, the signal area in the network field intensity prediction sample with the base station as the center is divided, if the signal grid exists, the grids are required to be integrated, the grid size is adjusted to be in line with the grid size of the subsequent model to be trained, then the center coordinate of each grid is used as the grid coordinate, secondly, the grid tag value is required to be calculated and associated with the corresponding grid number, wherein the grid tag value is the influence factor of N sample points in the grid on the grid center point field intensity value, therefore, the obtained grid tag feature value is the feature value for reflecting the signal condition of each area around the base station as the center, and the tag interpolation generating module in fig. 3 is used for processing.
In the embodiment, the network field intensity prediction sample is subjected to rasterization processing to obtain a transmitting end coordinate and grid information of each signal grid, and label interpolation is performed on each signal grid according to the transmitting end coordinate and the grid information to obtain a rasterized label characteristic value. The transmitting end is the signal base station.
In a specific implementation, the present embodiment proposes a preferred scheme for generating the characteristic value of the rasterized tag and describes this as an example, for example, first, the map is rasterized with a base station as the center, and the resolution is 5m×5m. And determining the mark of the grid where each sample value is located according to the geographic coordinate value of the sample value. Let the coordinates of the cell center point be (cell_x, cell_y), r be the preset cell coverage radius, x be the preset grid size (default 5 m), then the upper left corner coordinates of the cell coverage:
x_min=cell_x-r
y_min=cell_y-r
The labels of the grids are respectively determined to be 0,1, & gt, (p x p-1) in sequence from left to right and from top to bottom (assuming that p=2r/x is the result is rounded up), and according to the geographic coordinate value of the sample value and the preset size of the grid, the label of the grid where each sample value is located is determined to be that the coordinate of the sample value is assumed to be (point_x j,point_yj), and the label i of the grid where the sample value is located is:
i=p*(floor((point_yj-y_min)/x))+floor((point_xj-x_min)/x)
The coordinates of the ith grid are the center coordinates (grid_x i,grid_yi) of the ith grid, and the calculation formula is:
grid_xi=x_min+x*(i%p+1/2)
grid_yi=y_min+x*(floor(i/p)+1/2)
Wherein,% represents the remainder operation, floor represents the rounding-down operation, the sample rasterization process is completed in the above manner, and the numbering of each grid is completed.
Furthermore, the grid label value is required to be further calculated on the basis, the influence factors of N sample points in the grid on the grid center point field intensity value are calculated by using a label interpolation algorithm, and the field intensity label value of the grid center point is obtained by interpolation according to the influence factors. Interpolation is calculated using BiCubic basis functions:
Wherein a is-0.5.
When the label interpolation method is applied, the abscissa of the sample point is calculated separately, the abscissa coefficient corresponding to the field intensity label value A (point_x i,point_yi) of the sample center point is calculated to be k_x i, and the ordinate coefficient is calculated to be k_y i. The calculation formula is as follows:
And similarly, obtaining coefficients corresponding to all sample points, and obtaining the weight corresponding to each sample point by the product of the abscissa weight coefficients. Finally, the field intensity label value of the grid center point is obtained through a summation formula, wherein the summation formula is as follows:
wherein, the For the field strength tag value of the sample point, B (X, Y) is the field strength tag value of the grid center point, k_x i represents the weight on the sample point abscissa, and k_y i represents the weight on the sample point ordinate. As shown in fig. 4, a large dot B represents a desired grid center point, and dots A1 to A5 distributed randomly around represent sample points within the grid point.
It should be noted that, in this embodiment, it is proposed that the label feature is calculated by the label interpolation algorithm, the calculated target value error is smaller, and the predicted result is closer to the true value when the calculated target value error is sent to the neural network to perform model training, that is, the effect of model training is improved, so that the predicted value of the model is more true when the model is used.
And step S30, determining a basic characteristic sample according to the network field intensity prediction sample, and generating a basic characteristic value according to the basic characteristic sample.
It should be noted that, the basic feature module in fig. 3 is a module for processing basic feature values, where the basic feature module performs feature combination on engineering parameters, map data, and the like in the network field intensity prediction sample to construct grid height features, angle features, and distance features as data bases for basic feature processing. The spatial relationship between each sampling point in the sample data and the base station is because the relative spatial relationship between the transmitting end (base station) and the receiving end (signal sampling point) affects the signal, for example, the relative elevation difference pitch angle affects the signal propagation, and because the signal contacts the ground to generate reflection to enhance or cancel the signal. The effects of this can be predicted more accurately by the underlying feature values.
In this embodiment, the network field intensity prediction sample is screened to obtain a basic feature sample, where the basic feature sample at least includes height difference data between a transmitting end and a receiving end, distance data between the transmitting end and the receiving end, and angle data between the transmitting end and a signal direction, and a basic feature value is generated according to the basic feature sample.
It should be noted that the required data needs to be screened out based on the network field intensity prediction samples, for example, engineering parameters, geographical data and the like, and profile features, altitude features, distance features and angle features are extracted as shown in fig. 5. Further feature extraction is performed, where the basic feature sample includes at least height difference data between the sending end and the receiving end, distance data between the sending end and the receiving end, and angle data between the sending end and the signal line direction, and may also include other parameters, which are not limited in this embodiment, where the angle data between the sending end and the signal line direction is an included angle between a grid center point and signal strength, such as a horizontal direction angle MsAzimuth in fig. 6.
In fig. 6, the signal direction angle data is the antenna horizontal direction angle, the user horizontal direction angle, the grid center point, the north direction angle, and the like.
In a specific implementation, this embodiment proposes a preferred scheme for obtaining a basic feature value in conjunction with fig. 5, and is described by taking this as an example, where feature extraction is performed according to the heights of the transmitting end, the receiving end and the relevant section first, where fig. 5 mainly includes the height features of the base station position and the grid center point, and the height features on the connection line of the two. The specific formula is as follows:
TRRelHeight=TxHeight+TxAltitude-TRHeight-TRAltitude
TxHeightProfile1=TxHeight+TxAltitude-TxHeight1-TxAltitude1
TxHeightProfile2=TxHeight+TxAltitude–TxHeight2–TxAltitude2
RxHeightProfile1=TxHeight+TxAltitude-RxHeight1-RxAltitude1
RxHeightProfile2=TxHeight+TxAltitude–RxHeight2–RxAltitude2
TRMaxEdgeHeight=TxHeight+TxAltitude–MaxHeight–MaxAltitude
TRRELHEIGHT is the relative height difference between the base station and the grid center point, txHeight is the height of the base station, because the base station is generally a base station tower and has a certain height to consider, txAltitude is the altitude of the base station, TRHeight is the actual height of the grid center point, and TRAltitude is the altitude of the grid center point;
TxHeightProfile1 is the height of the sampling point 5m away from the base station on the connecting line of the base station and the grid central point, txHeight is the height of the building where the sampling point 5m away from the base station is located, and TxAltitude is the altitude of the sampling point 5m away from the base station;
TxHeightProfile2 is the height of a sampling point which is 10m away from the base station on the connecting line of the base station and the grid central point, txHeight is the height of a building where the sampling point which is 10m away from the base station is located, and TxAltitude2 is the altitude of the sampling point which is 10m away from the base station;
RxHeightProfile1 is the height of the sampling point 5m away from the grid center point on the connection line of the base station and the grid center point, rxHeight is the height of the building where the sampling point 5m away from the grid center point is located, and RxAltitude1 is the altitude of the sampling point 5m away from the grid center point;
RxHeightProfile2 is the height of a sampling point which is 10m away from the grid center point on the connecting line of the base station and the grid center point, rxHeight is the height of a building where the sampling point is 10m away from the grid center point, and RxAltitude is the altitude of the sampling point which is 10m away from the grid center point;
TRMaxEdgeHeight is the relative height of the highest peak of the connection line of the base station and the grid center point, maxHeight is the building height of the highest peak of the connection line of the base station and the grid center point, and MaxAltitude is the altitude of the highest peak of the connection line of the base station and the grid center point.
Further, feature extraction needs to be performed on the transmitting end, the receiving end and the related profile distances, as shown in fig. 5, specifically including construction of the 2D plane distance TRDISTANCE between the grid position where the receiving point is located and the transmitting point, and the horizontal 2D distance EdgeDistanceToRx between the highest peak of the receiving line and the receiving point. Wherein, (CellX, cellY) is the geographic location of the base station, (X, Y) is the grid center point coordinate, and (X_max, Y_max) is the grid location where the highest peak of the transceiving connection line is located. The specific calculation formula is as follows:
Finally, feature extraction is required to be performed on the grid center point and the signal angle, as shown in fig. 6, the feature extraction mainly comprises an included angle MsAzimuth between the grid center point and the direction in which the signal intensity is located, and an elevation angle MsElevation between the grid center point and the top end of the base station. And introducing an offset angle between the direction in which the signal intensity is located and the north direction, namely an antenna horizontal angle Azimuth, and enabling an angle between the grid center point and the north direction to be theta. The specific formula is as follows:
MsAzimuth=θ-Azimuth
MsElevation=arctan(TRRelHeight/TRDistance)
the linear characteristics (height characteristics, distance characteristics and angle characteristics) are subjected to normal distribution normalization, so that the distribution of the characteristics can be uniform, the convergence rate of a model is increased, and a specific normal distribution normalization formula is as follows:
Wherein x represents a sample value, namely the linear feature vector, mu is a sample mean value, sigma is a sample standard deviation, and z is a normalized result. The above linear features (height feature, distance feature, angle feature) are normalized to obtain a basic feature value, which is not limited to the present embodiment, and may be directly used as the basic feature value without normalization, but the effect is relatively poor.
And S40, generating a network field intensity prediction characteristic according to the characteristic value and the basic characteristic value of the rasterized label, wherein the network field intensity prediction characteristic is used for training a network field intensity prediction model.
The network field intensity prediction feature is generated according to the rasterized tag feature value and the basic feature value and is used as the feature data input training to train the network field intensity prediction model, and the trained network field intensity prediction model is ensured by introducing the spatial position relation and the tag value of each grid, so that the network field intensity of the base station can be predicted accurately under the topography.
It should be noted that, based on the label interpolation algorithm, the grid digitization processing is performed on the original sample data, and the field intensity value of the mobile network corresponding to the grid is generated, so that the error of the field intensity value averaging can be reduced, and the training result is more accurate. And the profile characteristics, the height characteristics, the distance characteristics and the angle characteristics are extracted aiming at sample data, engineering parameters, geographic data and the like, so that the correlation between the profile characteristics, the height characteristics, the distance characteristics and the angle characteristics and field intensity sample values is improved, the model convergence speed is accelerated, and a more universal and accurate intelligent propagation model is obtained.
The embodiment obtains a network field intensity prediction sample, performs label interpolation processing on the network field intensity prediction sample to generate a rasterized label feature value, determines a basic feature sample according to the network field intensity prediction sample, generates a basic feature value according to the basic feature sample, and generates a network field intensity prediction feature according to the rasterized label feature value and the basic feature value, wherein the network field intensity prediction feature is used for training a network field intensity prediction model. By the method, multiple feature values are applied to the extraction of the feature vectors, the features for predicting the field intensity of the mobile network are generated, and the features are input into the neural network for model training.
Referring to fig. 7, fig. 7 is a flowchart of a second embodiment of a feature generation method for network field strength prediction according to the present invention.
Based on the first embodiment, the feature generation method of network field strength prediction according to the present embodiment further includes, before the step S40 and in the step S40:
and S401, determining the terrain type characteristics of the cell according to the network field intensity prediction sample.
It should be noted that different terrains also affect the reception and propagation of signals, so that it is advantageous to overcome this by extracting the characteristics of the cell topography type of each cell and adding them to the characteristics for training, and to improve the accuracy of model prediction.
In the embodiment, preset type information is acquired, a current sample cell parameter is determined according to the network field intensity prediction sample, and the current sample cell parameter is matched with the preset type information to obtain a cell topography type characteristic.
It should be noted that, because the cell topography type has uncertainty, different preset cell types need to be set in advance to help the extraction of the cell topography type features. The cell classification module is used for extracting the topographic type features of the cell as shown in fig. 3.
In a specific implementation, the embodiment provides a preferred scheme for extracting the topographic type features of the cell, for example, the topographic type features of the cell are judged based on an EMD algorithm. In the field intensity prediction of the mobile network, some characteristics which cannot be learned by the neural network but have great influence on a prediction result exist, and the characteristics are required to be extracted for unified processing, so that a cell is divided into 20 preset types based on a large amount of sample data and priori knowledge, and the type of the topography of the cell is judged by utilizing a bulldozer distance algorithm. The specific algorithm is shown in fig. 8.
Wherein, p= < P 1,p2,…,pN>,Q＝<q1,q2,…,qN > is two kinds of terrain distribution, u i is each weight in X, v j is each weight in Y. The EMD was used to calculate the similarity between the two terrain distributions.
Wherein the flow between p i and q j is f ij,pi and the distance between q j is d ij, and
∑ifij＝vj,∑jfij＝ui,dij＝abs(pi-qj)
The similarity between the cell and 20 preset cell topographic distribution types can be respectively obtained based on the algorithm, and then the distribution type of the cell is judged.
The method is applied to the extraction of the feature vectors, generates the feature engineering for predicting the field intensity of the mobile network, inputs the feature engineering into the neural network for model training, and compared with the simple feature engineering, the prediction accuracy is greatly improved, and the model training effect is better. The feature of the district topography category is extracted based on an EMD algorithm, and the feature mobile network field intensity prediction has strong correlation, so that the model training result has higher precision and better training effect.
And S41, generating a network field intensity prediction feature according to the district topography type feature, the rasterization tag feature value and the basic feature value, or generating a network field intensity prediction feature according to the preset ground feature relation feature, the district topography type feature, the rasterization tag feature value and the basic feature value.
It should be noted that, the characteristics of the district topography type can be combined with the above-mentioned rasterized tag characteristic values and basic characteristic value generating network field intensity prediction characteristics to be used as the characteristics of the training model to be input into the network model for training. And generating a network field intensity prediction characteristic according to the preset ground object relation characteristic, the cell topography type characteristic, the rasterized tag characteristic value and the basic characteristic value for model training.
The preset feature relation feature is a relation between a space acquired in advance and a receiving end of a transmitting end, and can be directly input, or can be obtained by extracting features of a network field intensity prediction sample according to a representation learning module shown in fig. 3. Similarly, the preset feature relation feature can be independently combined with the rasterized tag feature value and the basic feature value to be used as the feature of the network model training.
In the embodiment, the network field intensity prediction sample is subjected to representation and learning processing to obtain a preset ground object relation characteristic, and the generation of the network field intensity prediction characteristic according to the rasterized tag characteristic value and the basic characteristic value comprises the generation of the network field intensity prediction characteristic according to the preset ground object relation characteristic, the rasterized tag characteristic value and the basic characteristic value.
It can be understood that the network field intensity prediction sample is subjected to representation learning processing, and the representation learning method is mainly used for representing the ground object relationship. The ground object relation is based on the ground object distribution between the base station and the actual signal receiving grid, for example, ocean, wetland, lake, forest and the like are included between the base station and the actual signal receiving grid, and the propagation efficiency of signals is affected.
In the embodiment, the feature visibility parameter and the feature category parameter are determined according to the network field intensity prediction sample, and the preset feature relation characteristic is determined according to the feature visibility parameter and the feature category parameter.
It should be noted that the embodiment proposes a preferred scheme for extracting feature relation features, for example, firstly, considering a three-dimensional topographic structure and a ray tracing algorithm to determine whether a transmitting point and a receiving grid are visible, obtaining feature visibility parameters whether the transmitting point and the receiving grid are visible, and characterizing the determination result by a local representation method. The local representation may be generally represented in the form of One-Hot vectors. And secondly, extracting feature vectors of the ground features based on the ground feature distribution between the base station and the actual signal receiving grid as shown in table 1, and carrying out characterization by using a distributed representation learning method. The distributed representation may be generally represented as a dense vector of low dimensions, i.e. a neural network is used to map the local representation space of high dimensions to a distributed representation space of very low dimensions. In this low-dimensional space, each feature is no longer a point on a coordinate axis, but is dispersed throughout the low-dimensional space, maintaining topological relationships between different feature values.
The embodiment provides a classification method of feature indexes, which has 20 categories in total, and each index value represents a different feature category, as shown in table 1.
Based on the distance, it is to be noted that, based on a distributed representation learning method in representation learning, feature index features on the connection line of the transmitting points and the receiving grid points are recorded, and the feature indexes are respectively mapped into 6-dimensional dense vectors by adopting the distributed representation learning method. The distributed representation method has the advantages that the representation capability of the feature index features is stronger, the vector dimension is reduced, the high-dimensional sparse problem is solved, and the generalization capability of the neural network under sparse input data is improved.
TABLE 1 surface feature index meaning Table
It should be noted that, based on the occlusion judgment result, the representation is performed by adopting a local representation learning method, and the discrete representation mode has good interpretation and very high calculation efficiency when being used in a linear model.
The embodiment determines the district topography type characteristic according to the network topography predicting sample, generates the network topography predicting characteristic according to the district topography type characteristic, the grating label characteristic value and the basic characteristic value, or generates the network topography predicting characteristic according to the preset ground feature relation characteristic, the district topography type characteristic, the grating label characteristic value and the basic characteristic value. The adaptability of the network field intensity prediction model is improved by extracting a plurality of characteristics and using the characteristics to train the network field intensity prediction model, so that the prediction of the network field intensity prediction model is not influenced by regional environment any more, and the network field intensity prediction model still has high accuracy under the condition of complex terrains or obstacles.
In addition, the embodiment of the invention also provides a storage medium, wherein the storage medium is stored with a characteristic generating program for network field intensity prediction, and the characteristic generating program for network field intensity prediction realizes the steps of the characteristic generating method for network field intensity prediction when being executed by a processor.
Referring to fig. 9, fig. 9 is a block diagram showing the configuration of a first embodiment of a feature generation apparatus for network field strength prediction according to the present invention.
As shown in fig. 9, the feature generating device for network field intensity prediction according to the embodiment of the present invention includes an obtaining module 10, configured to obtain a network field intensity prediction sample.
And the processing module 20 is used for performing label interpolation processing on the network field intensity prediction samples to generate rasterized label characteristic values.
The processing module 20 is further configured to determine a basic feature sample according to the network field strength prediction sample, and generate a basic feature value according to the basic feature sample.
The processing module 20 is further configured to generate a network field strength prediction feature according to the rasterized tag characteristic value and the basic characteristic value, where the network field strength prediction feature is used to train a network field strength prediction model.
It should be understood that the foregoing is illustrative only and is not limiting, and that in specific applications, those skilled in the art may set the invention as desired, and the invention is not limited thereto.
The method comprises the steps of obtaining a network field intensity prediction sample by the obtaining module 10, carrying out label interpolation processing on the network field intensity prediction sample by the processing module 20 to generate a rasterized label characteristic value, determining a basic characteristic sample according to the network field intensity prediction sample by the processing module 20, generating a basic characteristic value according to the basic characteristic sample, and generating a network field intensity prediction characteristic according to the rasterized label characteristic value and the basic characteristic value by the processing module 20, wherein the network field intensity prediction characteristic is used for training a network field intensity prediction model. By the method, multiple feature values are applied to the extraction of the feature vectors, the features for predicting the field intensity of the mobile network are generated, and the features are input into the neural network for model training.
In an embodiment, the processing module 20 is further configured to perform a representation learning process on the network field strength prediction sample to obtain a preset feature relation feature;
the generating a network field intensity prediction feature according to the rasterized tag feature value and the basic feature value comprises the following steps:
And generating a network field intensity prediction feature according to the preset ground object relation feature, the rasterized tag feature value and the basic feature value.
In an embodiment, the processing module 20 is further configured to determine a feature visibility parameter and a feature class parameter according to the network field strength prediction sample;
and determining preset feature relation characteristics according to the feature visibility parameters and the feature category parameters.
In an embodiment, the processing module 20 is further configured to determine a cell terrain type feature according to the network field strength prediction samples;
generating a network field intensity prediction feature according to the cell topography type feature, the rasterized tag feature value and the basic feature value;
Or alternatively, the first and second heat exchangers may be,
And generating a network field intensity prediction feature according to the preset ground object relation feature, the district topography type feature, the rasterized tag feature value and the basic feature value.
In an embodiment, the processing module 20 is further configured to obtain preset type information;
determining a current sample cell parameter according to the network field intensity prediction sample;
and matching the current sample cell parameters with preset type information to obtain the cell topography type characteristics.
In an embodiment, the processing module 20 is further configured to rasterize the network field strength prediction sample to obtain a transmitting end coordinate and grid information of each signal grid;
and carrying out label interpolation on each signal grid according to the transmitting end coordinates and the grid information to obtain a rasterized label characteristic value.
In an embodiment, the processing module 20 is further configured to screen the network field strength prediction samples to obtain basic feature samples, where the basic feature samples at least include height difference data between a transmitting end and a receiving end, distance data between the transmitting end and the receiving end, and angle data between the transmitting end and a signal direction;
And generating a basic characteristic value according to the basic characteristic sample.
It should be noted that the above-described working procedure is merely illustrative, and does not limit the scope of the present invention, and in practical application, a person skilled in the art may select part or all of them according to actual needs to achieve the purpose of the embodiment, which is not limited herein.
In addition, technical details which are not described in detail in the present embodiment may refer to the feature generation method of network field intensity prediction provided in any embodiment of the present invention, and are not described herein.
Furthermore, it should be noted that, in this document, the terms "comprises," "comprising," or any other variation thereof, are intended to cover a non-exclusive inclusion, such that a process, method, article, or system that comprises a list of elements does not include only those elements but may include other elements not expressly listed or inherent to such process, method, article, or system. Without further limitation, an element defined by the phrase "comprising one does not exclude the presence of other like elements in a process, method, article, or system that comprises the element.
The foregoing embodiment numbers of the present invention are merely for the purpose of description, and do not represent the advantages or disadvantages of the embodiments.
From the above description of the embodiments, it will be clear to those skilled in the art that the above-described embodiment method may be implemented by means of software plus a necessary general hardware platform, but of course may also be implemented by means of hardware, but in many cases the former is a preferred embodiment. Based on such understanding, the technical solution of the present invention may be embodied essentially or in a part contributing to the prior art in the form of a software product stored in a storage medium (e.g. Read Only Memory)/RAM, magnetic disk, optical disk) and including several instructions for causing a terminal device (which may be a mobile phone, a computer, a server, or a network device, etc.) to perform the method according to the embodiments of the present invention.
The foregoing description is only of the preferred embodiments of the present invention, and is not intended to limit the scope of the invention, but rather is intended to cover any equivalents of the structures or equivalent processes disclosed herein or in the alternative, which may be employed directly or indirectly in other related arts.
Claims (9)
1. The characteristic generation method of the network field intensity prediction is characterized by comprising the following steps of:
Acquiring a network field intensity prediction sample;
Performing label interpolation processing on the network field intensity prediction sample to generate a rasterized label characteristic value, wherein the rasterized label characteristic value is obtained by calculating influence factors of N sample points in a grid on a grid center point field intensity value by using a label interpolation algorithm and interpolating according to the influence factors;
determining a basic characteristic sample according to the network field intensity prediction sample, and generating a basic characteristic value according to the basic characteristic sample;
Generating a network field intensity prediction feature according to the rasterized tag characteristic value and the basic characteristic value, wherein the network field intensity prediction feature is used for training a network field intensity prediction model;
The method for determining the basic characteristic sample according to the network field intensity prediction sample and generating the basic characteristic value according to the basic characteristic sample comprises the following steps:
Screening the network field intensity prediction samples to obtain basic characteristic samples, wherein the basic characteristic samples at least comprise height difference data of a sending end and a receiving end, distance data of the sending end and the receiving end and angle data of the sending end and a signal direction;
And generating a basic characteristic value according to the basic characteristic sample.
2. The method of claim 1, wherein prior to generating the network field strength prediction feature from the rasterized tag characteristic value and the base characteristic value, further comprising:
performing representation learning processing on the network field intensity prediction sample to obtain preset ground object relation characteristics;
the generating a network field intensity prediction feature according to the rasterized tag feature value and the basic feature value comprises the following steps:
And generating a network field intensity prediction feature according to the preset ground object relation feature, the rasterized tag feature value and the basic feature value.
3. The method of claim 2, wherein the performing a representation learning process on the network field strength prediction samples to obtain a preset feature relationship feature comprises:
Determining feature visibility parameters and feature category parameters according to the network field intensity prediction samples;
and determining preset feature relation characteristics according to the feature visibility parameters and the feature category parameters.
4. A method according to any one of claims 1 to 3, wherein before generating a network field strength prediction feature from the rasterized tag eigenvalues and base eigenvalues, comprising:
Determining the district topography type characteristic according to the network field intensity prediction sample;
The generating network field intensity prediction features according to the rasterized tag feature values and the basic feature values further comprises:
generating a network field intensity prediction feature according to the cell topography type feature, the rasterized tag feature value and the basic feature value;
Or alternatively, the first and second heat exchangers may be,
And generating a network field intensity prediction feature according to the preset ground object relation feature, the district topography type feature, the grid tag feature value and the basic feature value.
5. The method of claim 4, wherein said determining a cell terrain type feature from said network field strength prediction samples comprises:
acquiring preset type information;
determining a current sample cell parameter according to the network field intensity prediction sample;
and matching the current sample cell parameters with preset type information to obtain the cell topography type characteristics.
6. A method as claimed in any one of claims 1 to 3, wherein said performing a label interpolation process on said network field strength prediction samples to generate rasterized label eigenvalues comprises:
rasterizing the network field intensity prediction sample to obtain a transmitting end coordinate and grid information of each signal grid;
and carrying out label interpolation on each signal grid according to the transmitting end coordinates and the grid information to obtain a rasterized label characteristic value.
7. A characteristic generating device for network field intensity prediction, characterized in that the characteristic generating device for network field intensity prediction comprises:
the acquisition module is used for acquiring a network field intensity prediction sample;
the processing module is used for carrying out label interpolation processing on the network field intensity prediction samples to generate a rasterized label characteristic value, wherein the rasterized label characteristic value is obtained by calculating influence factors of N sample points in a grid on a grid center point field intensity value by using a label interpolation algorithm and interpolating according to the influence factors;
the processing module is further used for determining a basic characteristic sample according to the network field intensity prediction sample and generating a basic characteristic value according to the basic characteristic sample;
The processing module is further used for generating a network field intensity prediction characteristic according to the rasterized tag characteristic value and the basic characteristic value, and the network field intensity prediction characteristic is used for training a network field intensity prediction model;
The processing module is further used for screening the network field intensity prediction samples to obtain basic characteristic samples, wherein the basic characteristic samples at least comprise height difference data of a sending end and a receiving end, distance data of the sending end and the receiving end and angle data of the sending end and a signal direction;
And generating a basic characteristic value according to the basic characteristic sample.
8. A network field strength predicted feature generation device comprising a memory, a processor and a network field strength predicted feature generation program stored on the memory and executable on the processor, the network field strength predicted feature generation program being configured to implement the steps of the network field strength predicted feature generation method of any one of claims 1 to 6.
9. A storage medium having stored thereon a network field strength predicted feature generation program which, when executed by a processor, implements the steps of the network field strength predicted feature generation method of any one of claims 1 to 6.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202210796088.7A CN116975589B (en) | 2022-07-07 | 2022-07-07 | Feature generation method, device, equipment and storage medium for network field intensity prediction |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202210796088.7A CN116975589B (en) | 2022-07-07 | 2022-07-07 | Feature generation method, device, equipment and storage medium for network field intensity prediction |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN116975589A CN116975589A (en) | 2023-10-31 |
| CN116975589B true CN116975589B (en) | 2025-10-31 |
Family
ID=88471891
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202210796088.7A Active CN116975589B (en) | 2022-07-07 | 2022-07-07 | Feature generation method, device, equipment and storage medium for network field intensity prediction |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN116975589B (en) |
Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103078690A (en) * | 2013-01-09 | 2013-05-01 | 中国电子科技集团公司第二十二研究所 | Correction method used for coverage field strength estimation of broadcast television system |
| WO2020125349A1 (en) * | 2018-12-20 | 2020-06-25 | 中兴通讯股份有限公司 | Field strength testing method |
Family Cites Families (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US6687657B2 (en) * | 2000-09-27 | 2004-02-03 | David N. Levin | Self-referential method and apparatus for creating stimulus representations that are invariant under systematic transformations of sensor states |
| US7583961B2 (en) * | 2004-04-28 | 2009-09-01 | Avaya Inc. | Method and apparatus for generating a signal strength model for an access point at an arbitrary location |
| CN106528956B (en) * | 2016-10-19 | 2019-10-25 | 天津大学 | Field strength prediction method based on data interpolation method based on ray tracing model |
| CN106599908A (en) * | 2016-11-29 | 2017-04-26 | 天津大学 | Field intensity prediction method based on modularized neural network |
| CN111507897B (en) * | 2019-01-30 | 2023-12-01 | 国家无线电监测中心陕西监测站 | Urban field intensity reconstruction method based on self-adaptive hybrid interpolation |
| CN110336629B (en) * | 2019-07-18 | 2021-11-12 | 上海创远仪器技术股份有限公司 | Method for realizing field intensity distribution prediction processing based on frequency spectrum drive test data and corresponding system |
| CN111062466B (en) * | 2019-12-11 | 2023-08-04 | 南京华苏科技有限公司 | Method for predicting field intensity distribution of cell after antenna adjustment based on parameters and neural network |
| CN112367129B (en) * | 2020-07-27 | 2022-02-11 | 南京邮电大学 | 5G reference signal received power prediction method based on geographic information |
-
2022
- 2022-07-07 CN CN202210796088.7A patent/CN116975589B/en active Active
Patent Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103078690A (en) * | 2013-01-09 | 2013-05-01 | 中国电子科技集团公司第二十二研究所 | Correction method used for coverage field strength estimation of broadcast television system |
| WO2020125349A1 (en) * | 2018-12-20 | 2020-06-25 | 中兴通讯股份有限公司 | Field strength testing method |
Also Published As
| Publication number | Publication date |
|---|---|
| CN116975589A (en) | 2023-10-31 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN108830870B (en) | Satellite image high-precision farmland boundary extraction method based on multi-scale structure learning | |
| Jin et al. | A point-based fully convolutional neural network for airborne LiDAR ground point filtering in forested environments | |
| CN108320329A (en) | A kind of 3D map creating methods based on 3D laser | |
| US20220067336A1 (en) | System and process for integrative computational soil mapping | |
| CN111428627B (en) | A remote sensing extraction method and system for mountainous landforms | |
| JP5981886B2 (en) | Point cloud analysis processing device, point cloud analysis processing method, and program | |
| US10080143B2 (en) | Method of placing an antenna of a radio access network (RAN) asset in a wireless communication network | |
| JP2019211342A (en) | Weather analyzer, weather analysis method, and program | |
| CN116127327B (en) | Forest ground biomass inversion method, device, equipment and storage medium | |
| CN110532963B (en) | Vehicle-mounted laser radar point cloud driven road marking accurate extraction method | |
| CN117935075A (en) | A group target detection method based on visible light remote sensing images | |
| CN116401827A (en) | A Method for Optimizing Distributed Detection Stations Based on Environment Awareness | |
| CN116975589B (en) | Feature generation method, device, equipment and storage medium for network field intensity prediction | |
| CN116819560A (en) | Multi-box accurate detection positioning system and method based on laser radar | |
| CN117994549B (en) | Ultra-high pressure maintenance safety control method, device, equipment and storage medium | |
| Sarkar et al. | Radiance: Radio-frequency adversarial deep-learning inference for automated network coverage estimation | |
| CN119920060A (en) | A landslide monitoring and early warning method and system for complex environments | |
| CN114390534A (en) | Beam pattern determination method, beam pattern determination device, electronic equipment and storage medium | |
| CN115802490A (en) | High-precision wireless positioning method, system and storage medium for positioning base station | |
| CN117241239A (en) | Indoor positioning method and system for position fingerprints based on one-dimensional convolutional neural network | |
| CN108694720A (en) | A kind of high score image building area extracting method based on multi-dimension texture space clustering | |
| Lukač et al. | Genetic algorithm for optimizing meteorological sensors placement using airborne LiDAR data | |
| CN119380105B (en) | A Power Line Target Extraction Method Based on Random Forest and Class-A Classification Strategy | |
| CN119692025B (en) | A Joint Surveying and Mapping Data Processing Information System | |
| Daniş et al. | Tracking a Mobile Beacon: A Purely Probabilistic Approach |
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 |