CN115226112B - Network planning method, device, equipment and storage medium based on machine learning - Google Patents
Network planning method, device, equipment and storage medium based on machine learning Download PDFInfo
- Publication number
- CN115226112B CN115226112B CN202110420946.3A CN202110420946A CN115226112B CN 115226112 B CN115226112 B CN 115226112B CN 202110420946 A CN202110420946 A CN 202110420946A CN 115226112 B CN115226112 B CN 115226112B
- Authority
- CN
- China
- Prior art keywords
- processed
- data
- preset
- machine learning
- network planning
- 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
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04W—WIRELESS COMMUNICATION NETWORKS
- H04W16/00—Network planning, e.g. coverage or traffic planning tools; Network deployment, e.g. resource partitioning or cells structures
- H04W16/18—Network planning tools
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N20/00—Machine learning
Abstract
The application discloses a network planning method, a device, equipment and a storage medium based on machine learning, wherein the method comprises the following steps: when a network planning instruction is detected, determining a road measuring point corresponding to the base station to obtain azimuth data to be processed corresponding to the network planning instruction; determining to-be-processed environment data based on a preset ray tracking mode, and obtaining to-be-processed input data based on the to-be-processed azimuth data and the to-be-processed environment data; inputting the input data to be processed into a preset machine learning model, and based on the preset machine learning model, carrying out prediction processing on the propagation link loss from the base station to the road point on the input data to be processed to obtain target predicted road loss; the preset machine learning model is obtained by performing iterative training on a preset model to be trained based on training data with preset loss labels. The method and the device accurately predict the path loss of the base station under the condition of small prediction calculation amount, and improve the prediction efficiency.
Description
Technical Field
The present application relates to the field of wireless network technologies, and in particular, to a network planning method, apparatus, device and storage medium based on machine learning.
Background
With the continuous development of technology, more and more technologies are applied to the field of network construction, but the network construction also puts higher requirements on the technologies, such as higher requirements on the accuracy of the network construction.
In network construction, the precision of base station network planning can influence the development process of the whole network, so the method has very important status and effect, and in order to improve the precision of base station network planning, all relevant characteristics of a base station and a road point are input into a network prediction model as input data, so that the technical problems of large calculated amount and low efficiency exist in the prediction process.
Disclosure of Invention
The application mainly aims to provide a network planning method, device, equipment and storage medium based on machine learning, and aims to solve the technical problems of large network path loss prediction calculation amount and low efficiency of a base station in the prior art.
In order to achieve the above object, the present application provides a network planning method based on machine learning, which is applied to a base station, and the network planning method based on machine learning includes:
When a network planning instruction is detected, determining a road measuring point corresponding to the base station to obtain azimuth data to be processed corresponding to the network planning instruction;
determining to-be-processed environment data based on a preset ray tracking mode, and obtaining to-be-processed input data based on the to-be-processed azimuth data and the to-be-processed environment data;
inputting the input data to be processed into a preset machine learning model, and based on the preset machine learning model, carrying out prediction processing on the propagation link loss from the base station to the road point on the input data to be processed to obtain target predicted road loss;
the preset machine learning model is obtained by performing iterative training on a preset model to be trained based on training data with preset loss labels.
Optionally, the step of obtaining the input data to be processed based on the azimuth data to be processed and the environmental data to be processed includes:
extracting azimuth characteristic parameters from the network planning instruction, and extracting to-be-processed characteristic data from the to-be-processed azimuth data and the to-be-processed environment data based on the azimuth characteristic parameters;
and performing data dimension reduction processing on the feature data to be processed to obtain input data to be processed.
Optionally, before the step of inputting the input data to be processed into a preset machine learning model, performing prediction processing of propagation link loss on the input data to be processed based on the preset machine learning model to obtain the target predicted path loss, the method includes:
acquiring historical data, performing data processing on the historical data based on the preset ray tracing mode to obtain training input data, and acquiring an actual road loss label of the training input data;
inputting the training input data into a preset model to be trained, and carrying out prediction processing on the training input data based on the preset model to be trained to obtain predicted path loss;
comparing the predicted path loss with the actual path loss to obtain a comparison result;
and based on the comparison result, if the preset training condition of the preset model to be trained is not finished, returning to the step of inputting the training input data into the preset model to be trained until the preset machine learning model is obtained.
Optionally, the step of obtaining the history data, performing data processing on the history data based on the preset ray tracing mode to obtain training input data includes:
According to a preset ray tracking mode, simulation data which correspond to base station information, road point information, map information, material configuration information and antenna information in the historical data together are determined, and training environment characteristics are obtained based on the simulation data;
determining antenna gain based on the historical data and a preset antenna gain dictionary;
acquiring historical data, and determining Euclidean distance between the base station and the road point based on the historical data;
acquiring historical data, and determining a relative horizontal angle and a relative vertical angle based on the historical data;
and setting the training environment characteristics, the antenna gain, the Euclidean distance, the relative horizontal angle and the relative vertical angle as the training input data.
Optionally, if it is determined that the model to be trained does not complete the preset training condition, the step of inputting the training input data into the model to be trained is returned until the preset machine learning model is obtained, where the step includes:
based on the comparison result, if the preset training condition of the preset model to be trained is not completed, adjusting weight parameters and bias parameters in the preset machine learning model;
And returning to the step of inputting the training input data into the adjusted preset model to be trained until the preset machine learning model is obtained.
Optionally, after the step of inputting the input data to be processed into a preset machine learning model, and performing prediction processing of propagation link loss on the input data to be processed based on the preset machine learning model to obtain the target predicted path loss, the method includes:
determining whether the target predicted path loss is within a preset path loss range;
and if the coverage prediction is within the preset path loss range, performing coverage prediction processing on the network based on a preset coverage prediction model.
The application also provides a network planning method based on machine learning, which is applied to the road measuring points and comprises the following steps:
when a network planning instruction is detected, a base station corresponding to the road measuring point is in communication connection so as to receive a wireless signal sent by the base station;
and acquiring the road point information based on the wireless signals, and sending the road point information to the base station so that a preset machine learning model in the base station can predict the propagation link loss based on the road point information to obtain target predicted road loss.
The application also provides a network planning device based on machine learning, which is applied to a base station, and comprises:
the first determining module is used for determining a road point corresponding to the base station when the network planning instruction is detected so as to obtain azimuth data to be processed corresponding to the network planning instruction;
the second determining module is used for determining to-be-processed environment data based on a preset ray tracing mode and obtaining to-be-processed input data based on the to-be-processed azimuth data and the to-be-processed environment data;
the first input module is used for inputting the input data to be processed into a preset machine learning model, and based on the preset machine learning model, carrying out prediction processing on the propagation link loss from the base station to the road point on the input data to be processed to obtain target prediction road loss;
the preset machine learning model is obtained by performing iterative training on a preset model to be trained based on training data with preset loss labels.
Optionally, the second determining module includes:
the first determining unit is used for extracting azimuth characteristic parameters from the network planning instruction, and extracting to-be-processed characteristic data from the to-be-processed azimuth data and the to-be-processed environment data based on the azimuth characteristic parameters;
And the dimension reduction unit is used for carrying out data dimension reduction processing on the feature data to be processed to obtain input data to be processed.
Optionally, the machine learning based network planning apparatus further includes:
the acquisition module is used for acquiring historical data, carrying out data processing on the historical data based on the preset ray tracing mode to obtain training input data, and acquiring an actual road loss label of the training input data;
the second input module is used for inputting the training input data into a preset model to be trained, and carrying out prediction processing on the training input data based on the preset model to be trained to obtain predicted path loss;
the comparison module is used for comparing the predicted path loss with the actual path loss to obtain a comparison result;
and the return module is used for returning to the step of inputting the training input data into the preset model to be trained until the preset machine learning model is obtained if the preset model to be trained is determined to not complete the preset training condition based on the comparison result.
Optionally, the acquiring module includes:
the first acquisition unit is used for determining simulation data which correspond to base station information, road point information, map information, material configuration information and antenna information in the historical data together according to a preset ray tracking mode, and obtaining training environment characteristics based on the simulation data;
A second determining unit, configured to determine an antenna gain based on the history data and a preset antenna gain dictionary;
the second acquisition unit is used for acquiring historical data and determining the Euclidean distance between the base station and the road point based on the historical data;
a third acquisition unit configured to acquire history data, and determine a relative horizontal angle and a relative vertical angle based on the history data;
and the setting unit is used for setting the training environment characteristics, the antenna gain, the Euclidean distance, the relative horizontal angle and the relative vertical angle as the training input data.
Optionally, the return module includes:
the adjusting unit is used for adjusting weight parameters and bias parameters in the preset machine learning model if the preset model to be trained is determined to not finish the preset training conditions based on the comparison result;
and the return unit is used for returning to the step of inputting the training input data into the adjusted preset model to be trained until the preset machine learning model is obtained.
Optionally, the machine learning based network planning apparatus further includes:
the third determining module is used for determining whether the target predicted path loss is in a preset path loss range;
And the coverage prediction module is used for performing coverage prediction processing on the network based on a preset coverage prediction model if the coverage prediction module is in the preset path loss range.
The application also provides a network planning device based on machine learning, which is applied to the road measuring points, and the network planning device based on machine learning comprises:
the detection module is used for carrying out communication connection with the base station corresponding to the road point when the network planning instruction is detected so as to receive the wireless signal sent by the base station;
the acquisition module is used for acquiring the road point information based on the wireless signals and sending the road point information to the base station so that a preset machine learning model in the base station can conduct prediction processing of propagation link loss based on the road point information to obtain target predicted road loss.
The application also provides a network planning device based on machine learning, which is entity node device, comprising: the computer-readable storage medium comprises a memory, a processor and a program of the machine-learning-based network planning method stored on the memory and executable on the processor, wherein the program of the machine-learning-based network planning method, when executed by the processor, can implement the steps of the machine-learning-based network planning method as described above.
The present application also provides a storage medium having stored thereon a program for implementing the above-mentioned machine learning based network planning method, which when executed by a processor implements the steps of the above-mentioned machine learning based network planning method.
The application also provides a computer program product comprising a computer program which, when executed by a processor, implements the steps of the machine learning based network planning method described above.
Compared with the prior art that the number of the features extracted by a machine learning algorithm model is large, so that the network route loss prediction calculation amount of a base station is large and the efficiency is low, in the application, when a network planning instruction is detected, only the to-be-processed azimuth data and the to-be-processed environment data obtained based on a preset ray tracing mode are needed to be obtained, the to-be-processed input data can be obtained and then input into the preset machine learning model to obtain the target predicted route loss, namely, in the application, the to-be-processed environment data (environment features) are introduced by an RT (ray tracing mode), and a small number of other azimuth features, namely, the to-be-processed azimuth data, can be input into the preset machine learning model to accurately obtain the corresponding target predicted route loss.
Drawings
The accompanying drawings, which are incorporated in and constitute a part of this specification, illustrate embodiments consistent with the application and together with the description, serve to explain the principles of the application.
In order to more clearly illustrate the embodiments of the application or the technical solutions of the prior art, the drawings which are used in the description of the embodiments or the prior art will be briefly described, and it will be obvious to a person skilled in the art that other drawings can be obtained from these drawings without inventive effort.
FIG. 1 is a flowchart of a first embodiment of a machine learning based network planning method according to the present application;
fig. 2 is a detailed step flow diagram of step S20 in the machine learning-based network planning method of the present application;
fig. 3 is a schematic device structure diagram of a hardware running environment according to an embodiment of the present application.
The achievement of the objects, functional features and advantages of the present application 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 application.
In a first embodiment of the machine learning-based network planning method according to the present application, referring to fig. 1, the machine learning-based network planning method includes:
Step S10, when a network planning instruction is detected, determining a road point corresponding to the base station to obtain azimuth data to be processed corresponding to the network planning instruction;
step S20, determining to-be-processed environment data based on a preset ray tracing mode, and obtaining to-be-processed input data based on the to-be-processed azimuth data and the to-be-processed environment data;
step S30, inputting the input data to be processed into a preset machine learning model, and based on the preset machine learning model, carrying out prediction processing on the propagation link loss from the base station to the road point on the input data to be processed to obtain target predicted road loss;
the preset machine learning model is obtained by performing iterative training on a preset model to be trained based on training data with preset loss labels.
The method comprises the following specific steps:
step S10, when a network planning instruction is detected, determining a road point corresponding to the base station to obtain azimuth data to be processed corresponding to the network planning instruction;
in this embodiment, it should be noted that the machine learning-based network planning method may be applied to site planning (site planning of a base station) belonging to a machine learning-based network planning system subordinate to a machine learning-based network planning apparatus.
For a network planning system based on machine learning corresponding to a certain site planning area, a preset machine learning model is built in, or preset machine learning models in other systems can be called, so that the route loss of a base station is predicted, the coverage prediction of the planning area is further performed, and the fact that the preset machine learning model is a trained model is required to be directly applied is required.
It should be noted that, the site planning area may also be built in or call other models, such as an overlay prediction model or an azimuth characteristic parameter determining model, which are not limited herein.
In this embodiment, a specific application scenario may be:
the network of a certain area needs to be planned, and in the planning process, the actual network coverage effect needs to be simulated, so that the rationality of site planning is verified, and the actual network coverage effect is related to network route loss. Therefore, it is necessary to determine the predicted network path loss, specifically, after the base station location is determined, it is necessary to determine the wireless network path loss at each path point, and if the wireless network path loss is within a preset acceptable range, the network coverage can be performed on the path point, or the design of the base station is reasonable at this time, so that the accurate determination of the wireless network path loss of the base station is a key point for performing network planning.
Specifically, for example, a base station is set at the a, a route point is designed at the B, and in the process of transmitting a wireless network signal of the base station to the B, if the route loss is within a preset range, the subsequent network coverage is performed, or the setting of the base station is determined to be reasonable at this time.
Specifically, in this embodiment, when a network planning instruction of a base station is detected, a road measurement point or a road measurement terminal (may be one or more) corresponding to the base station is determined, and to-be-processed azimuth data corresponding to the network planning instruction is obtained, where the to-be-processed azimuth data corresponding to the network planning instruction may be obtained:
mode one: when a network planning instruction of a base station is detected, directly extracting to-be-processed azimuth data from the network planning instruction, namely, in the embodiment, the to-be-processed azimuth data are carried in the instruction (the to-be-processed azimuth data are tested in advance);
in this embodiment, the triggering mode of the network planning instruction may be a voice mode or a touch mode. That is, in this embodiment, the network planning instructions may be triggered on a machine learning based network planning system interface.
Mode two: triggering a preset road test flow when a network planning instruction of a base station is detected, and acquiring azimuth data to be processed, which are measured in real time by a road test point, based on the preset road test flow;
In this embodiment, it should be noted that the azimuth data to be processed may be azimuth raw data to be processed or azimuth features to be processed, where the azimuth features to be processed may be in a vector form or a matrix form.
That is, in this embodiment, when a network planning instruction of a base station is detected, obtaining azimuth data to be processed corresponding to the network planning instruction includes:
when a network planning instruction of a base station is detected, acquiring a feature of a direction vector to be processed corresponding to the network planning instruction;
or when the network planning instruction of the base station is detected, acquiring the azimuth matrix characteristics to be processed corresponding to the network planning instruction.
In this embodiment, if the azimuth data to be processed is the azimuth raw data to be processed, the azimuth raw data to be processed is preprocessed, so as to obtain the azimuth vector feature to be processed or the azimuth matrix feature to be processed. The preprocessing includes vectorized preprocessing or matrixing preprocessing.
Step S20, determining to-be-processed environment data based on a preset ray tracing mode, and obtaining to-be-processed input data based on the to-be-processed azimuth data and the to-be-processed environment data;
In this embodiment, based on a preset ray tracing manner, the to-be-processed environmental data is determined, and the specific process may be: based on a preset ray tracing algorithm in a preset ray tracing mode, base station information, road point information, map information, material configuration information, antenna information and the like, predicting the signal intensity of the base station, generating a simulation intermediate file correspondingly in the process of predicting the signal intensity of the receiving point, and extracting environment data to be processed from the simulation intermediate file, and specifically extracting to obtain environment characteristics (matrix or vector form) to be processed.
In this embodiment, the to-be-processed input data is obtained by combining the to-be-processed azimuth data and the to-be-processed environmental data, and specifically, the to-be-processed azimuth data and the to-be-processed environmental data have the same data form, such as a matrix form or a vector form.
In this embodiment, the azimuth data to be processed specifically includes a relative horizontal angle (alpha), a relative vertical angle, a direct path corresponding to an antenna gain, a transmission path corresponding to an antenna gain, a diffraction path corresponding to an antenna gain, and the like.
In this embodiment, after the azimuth feature to be processed (azimuth data to be processed) is obtained, the environmental feature to be processed of the base station is obtained based on the preset ray tracing mode, so that input data can be obtained and input into a preset machine learning model instead of inputting all features (including invalid features) into the preset machine learning model in a unified way, thereby reducing data processing amount and improving data processing efficiency.
In this embodiment, because the to-be-processed environmental feature of the base station is obtained based on the preset ray tracing mode instead of measuring the complex environmental feature, the time for obtaining the environmental data can be reduced, and the data processing efficiency can be further improved.
In this embodiment, the direct path corresponds to the antenna gain, the transmission path corresponds to the antenna gain, and the diffraction path corresponds to the antenna gain, and the euclidean distance, the relative horizontal angle (alpha), and the relative vertical angle (beta) belong to azimuth data to be processed; the direct path order and the transmission path order belong to the environmental data to be processed, namely the direct path order and the transmission path order, the diffraction path order is an environmental parameter (also called multipath order, which represents the number of intersection points of the ray and a scene) introduced by a preset ray tracking mode, namely RT, and in the outdoor environment, when the direct path exists between the receiving and transmitting points (the base station and the path measuring point), the transmission path and the diffraction path do not exist, and when the transmission path and the diffraction path exist, the direct path does not exist.
Referring to fig. 2, the step of obtaining the input data to be processed based on the azimuth data to be processed and the environmental data to be processed includes the following steps S21 to S22:
S21, extracting azimuth characteristic parameters from the network planning instruction, and extracting to-be-processed characteristic data from the to-be-processed azimuth data and the to-be-processed environment data based on the azimuth characteristic parameters;
in this embodiment, first, an azimuth characteristic parameter is extracted from the network planning instruction, where the azimuth characteristic parameter may be preset in the network planning instruction, or the azimuth characteristic parameter is determined based on a configuration file.
In this embodiment, the to-be-processed feature data is extracted from the to-be-processed azimuth data and the to-be-processed environment data by the names of the azimuth feature parameters, that is, the parameter contents corresponding to the azimuth feature parameters are extracted from the to-be-processed azimuth data and the to-be-processed environment data.
And S22, performing data dimension reduction processing on the feature data to be processed to obtain input data to be processed.
In this embodiment, the data dimension reduction process is performed on the feature data to be processed to obtain input data to be processed, where the dimension reduction mode may be a k-neighbor dimension reduction mode or the like.
In this embodiment, after the feature data to be processed is obtained, the data dimension reduction processing is further performed, so that the data processing amount can be further reduced, and the data processing efficiency can be improved.
Step S30, inputting the input data to be processed into a preset machine learning model, and based on the preset machine learning model, carrying out prediction processing on the propagation link loss from the base station to the road point on the input data to be processed to obtain target predicted road loss;
the preset machine learning model is obtained by performing iterative training on a preset model to be trained based on training data with preset loss labels.
In this embodiment, the input data to be processed is input into a preset machine learning model, and the prediction processing of the propagation link loss from the base station to the road point is performed on the input data to be processed based on the preset machine learning model to obtain the target predicted path loss, that is, because the preset machine learning model is a model capable of accurately predicting the path loss obtained by performing iterative training on the preset model to be trained based on training data with a preset loss label, the target predicted path loss of the input data to be processed can be accurately predicted.
In this embodiment, it should be noted that the training data based on the preset wear-out tag may also be processed based on the preset ray tracing mode, that is, in this embodiment, the training data is processed by the preset ray tracing algorithm.
Before the step of inputting the input data to be processed into a preset machine learning model and performing prediction processing of propagation link loss on the input data to be processed based on the preset machine learning model to obtain a target predicted path loss, the method comprises the following steps S01-S04:
step S01, acquiring historical data, performing data processing on the historical data based on the preset ray tracing mode to obtain training input data, and acquiring an actual road loss label of the training input data;
in this embodiment, historical data (including base station information, road point information, map information, material configuration information, antenna information, and the like) is obtained, the historical data is basic data that is not processed, data processing is performed on the historical data based on the preset ray tracing mode, training input data is obtained, and after the training input data is obtained, an actual road loss label of the training input data is obtained, where the actual road loss label of the training input data may be a label recorded in history, or a label marked manually after manual testing.
The step of obtaining the history data, performing data processing on the history data based on a preset ray tracing mode to obtain training input data and obtaining an actual road loss label of the training input data comprises the following steps:
Acquiring historical data, and performing data processing on the historical data based on a preset ray tracing mode to obtain initial input data;
randomly extracting data with preset proportion, such as 30%, from the initial input data to serve as training input data, and acquiring an actual path loss label of the training input data. Namely, during training, only 30% of the data set is taken as training data, and a corresponding target model can be obtained.
The base station is in communication connection with the base station, the step of obtaining historical data, performing data processing on the historical data based on a preset ray tracing mode to obtain training input data comprises the following steps of A1-A5:
a1, determining simulation data which correspond to base station information, road point information, map information, material configuration information and antenna information in the historical data together according to a preset ray tracking mode, and obtaining training environment characteristics based on the simulation data;
a2, determining antenna gain based on the historical data and a preset antenna gain dictionary;
antenna gain is one of the most important parameters of an antenna, and the size of the antenna gain is proportional to the coverage area of a signal, so that the operation quality of wireless communication is affected, and therefore, the antenna gain between the base station and the road measuring point needs to be determined.
Specifically, in the present embodiment, the antenna gain is determined based on the history data and a preset antenna gain dictionary.
In this embodiment, the specific process of obtaining the antenna gain may be: traversing according to an interpolation method to obtain antenna gain dictionaries in horizontal [ -180, 180] directions and vertical [ -90, 90] directions, then respectively obtaining EOD (projection angle) and AOD (pitch angle) in a direct path, a projection path and a diffraction path, obtaining antenna gains in corresponding paths according to the EOD and AOD table look-up (preset antenna gain dictionary), and converting the antenna gains from a log domain to a linear domain, and normalizing, namely the antenna gain characteristics in the characteristics.
The specific method comprises the following steps: set Gain ue Antenna gain obtained by a dB domain (namely log domain) interpolation method;
Gain max maximum antenna gain for dB domain;
let P be ue Antenna gain for the linear domain;
P max maximum antenna gain for the linear domain;
the way to shift the antenna gain from the dB domain to the linear domain is:
then normalizing the antenna gain based on the linear domain
Step A3, acquiring historical data, and determining the Euclidean distance between the base station and the road measuring point based on the historical data;
if n is the number of the road points or the receiving points, the Euclidean Distance (Distance) between the base station and the road points is:
Wherein, (xi, yi, zi) is the coordinates of the road point, and (X, Y, Z) is the coordinates of the road point.
Step A4, acquiring historical data, and determining a relative horizontal angle and a relative vertical angle based on the historical data;
acquiring historical data, and acquiring Tx and Rx based on the historical data, wherein Tx (X0, y0, z 0) is a transmitting point coordinate, rx (X1, y1, z 1) is a receiving point coordinate, and the relative horizontal angle is the included angle between a projection line segment of Rx to the horizontal plane and the X axisThe relative vertical angle is the included angle between the line segment between the receiving and transmitting points and the projection line segment of Rx to the horizontal plane
And step A5, setting the training environment characteristics, the antenna gain, the Euclidean distance, the relative horizontal angle and the relative vertical angle as the training input data.
In this embodiment, environmental parameters are introduced by combining a preset three-dimensional map corresponding to a base station in a preset ray tracing manner, so as to obtain training environmental features, and a specific process for obtaining the training environmental features may be:
determining base station information, road point information, such as information of transmitting power, frequency and the like (determined by inputting an original file or setting);
determining map information of the base station, such as landform type, ground elevation, building height and the like (determined by inputting original files or setting);
Determining material configuration information such as loss of rays through an obstacle, etc. (determined by inputting an original file or set-up);
antenna information such as antenna configuration information, antenna patterns, etc. is determined (by inputting an original file or setting).
In this embodiment, after the drive test base station information, the map information, the material configuration information and the antenna information are determined, the signal intensity of the receiving point is predicted by using a ray tracing algorithm, and in the process of predicting the signal intensity of the receiving point, a simulation intermediate file is generated correspondingly, and the environmental data to be processed is extracted from the simulation intermediate file, specifically, the environmental characteristics (matrix or vector form) to be processed are extracted. In the present embodiment, the original file input by the ray tracing method is: the method comprises the steps of obtaining a result matrix (including some parameters) and a simulation result file based on the original file, and further obtaining the environmental characteristics to be processed. Namely, the type of the path between the transmitting and receiving points and the corresponding direct path order, transmission path order and diffraction path order (the multi-path order represents the number of times of rays and scene collision points).
Step S02, inputting the training input data into a preset model to be trained, and carrying out prediction processing on the training input data based on the preset model to be trained to obtain predicted path loss;
In this embodiment, after obtaining training input data, i.e. obtaining history data processed by a preset ray tracing mode, the training input data is input into a preset model to be trained, and prediction processing is performed on the training input data based on the preset model to be trained, so as to obtain predicted path loss.
In this embodiment, if the model to be trained is a three-layer fully-connected neural network structure, that is, only one hidden layer, one input layer and one output layer are included, where the number of neurons in the input layer may be 9, the number of neurons in the hidden layer may be 10, and the number of neurons in the output layer is 1. Before training starts, initializing parameters of a preset model to be trained, for example, adopting truncated front distribution for weight initialization, setting standard deviation to be 0.1 for initial standard deviation, setting bias to be a constant, setting the bias to be 0.1, and adjusting the corresponding learning rate according to actual conditions.
The algorithm training process sends training input data into the network in a batch mode, and the batch number is 256.
S03, comparing the predicted path loss with the actual path loss label to obtain a comparison result;
the specific training process is as follows:
The network input feature vector (training input data) is x= (X1, X2, X3, X4, X5, X6, X7, X8, X9), corresponding to the direct path order, direct path corresponding antenna gain, transmission path order, transmission path corresponding antenna gain, diffraction path order, diffraction path corresponding antenna gain, euclidean distance, relative horizontal angle (alpha), relative vertical angle (beta).
The output of the hidden layer is: f (X) =relu (W 1 X+b 1 ) Wherein W is 1 And b 1 The weights and biases of the neurons of the first layer of the network are represented, respectively. The activation function employed by the hidden layer is ReLU.
The output layer outputs: y=w 2 f(X)+b 2 Wherein W is 2 And b 2 Representing the weight and bias of the second layer of the network. The output layer has no activation function.
Representing the target predicted path loss of network output as y (i) The actual path loss is expressed asThe error between the network output and the actual path loss is calculated, and the MSE loss function is selected and expressed as:
in this embodiment, the predicted path loss is compared with the actual path loss label to obtain a comparison result, and the prediction accuracy is obtained based on the comparison result.
And step S04, based on the comparison result, if the preset training condition is not completed by the preset model to be trained, returning to the step of inputting the training input data into the preset model to be trained until the preset machine learning model is obtained.
And based on the comparison result, if the preset training condition of the preset model to be trained is not finished, returning to the step of inputting the training input data into the preset model to be trained until the preset machine learning model is obtained.
And if the comparison result is based on that the preset training condition is not completed by the preset model to be trained, returning to the step of inputting the training input data into the preset model to be trained until the preset machine learning model is obtained, wherein the step comprises the following steps:
based on the comparison result, if the preset training condition of the preset model to be trained is not finished, returning to the step of inputting the training input data into the preset model to be trained;
and if the preset model to be trained is determined to finish the preset training conditions, verifying the trained model based on the initial input data and the verification path loss label of the initial input data, and obtaining the preset machine learning model when verification passes.
Compared with the prior art that the number of the features extracted by a machine learning algorithm model is large, so that the network route loss prediction calculation amount of a base station is large and the efficiency is low, in the application, when a network planning instruction is detected, only the to-be-processed azimuth data and the to-be-processed environment data obtained based on a preset ray tracing mode are needed to be obtained, the to-be-processed input data can be obtained and then input into the preset machine learning model to obtain the target predicted route loss, namely, in the application, the to-be-processed environment data (environment features) are introduced by an RT (ray tracing mode), and a small number of other azimuth features, namely, the to-be-processed azimuth data, can be input into the preset machine learning model to accurately obtain the corresponding target predicted route loss.
Further, based on the first embodiment of the present application, another embodiment of the present application is provided, in this embodiment, based on the comparison result, if it is determined that the preset model to be trained does not complete the preset training condition, the step of inputting the training input data into the preset model to be trained is returned until the preset machine learning model is obtained, including:
step B1, based on the comparison result, if the preset training condition of the preset model to be trained is not completed, adjusting weight parameters and bias parameters in the preset machine learning model;
and step B2, returning to the step of inputting the training input data into the adjusted preset model to be trained until the preset machine learning model is obtained.
In this embodiment, if it is determined that the preset model to be trained does not have the preset training conditions, the weight parameters and the bias parameters in the preset machine learning model are adjusted, specifically, the weight parameters and the bias parameters in the preset machine learning model are directionally adjusted based on the comparison result and the actual path loss label, and training is continued based on the adjusted model until the preset machine learning model is obtained.
In this embodiment, because based on the comparison result, if it is determined that the preset model to be trained does not complete the preset training condition, the weight parameter and the bias parameter in the preset machine learning model are adjusted, and the step of inputting the training input data into the adjusted preset model to be trained is returned until the preset machine learning model is obtained. In this embodiment, accurate adjustment of the preset model to be trained is achieved, so as to quickly obtain a preset machine learning model.
Further, according to the first embodiment and the second embodiment of the present application, there is provided another embodiment of the present application, in which the step of inputting the input data to be processed into a preset machine learning model, and performing prediction processing of propagation link loss on the input data to be processed based on the preset machine learning model to obtain a target predicted path loss, the method includes:
step S40, determining whether the target predicted path loss is within a preset path loss range;
and S50, if the coverage prediction is within the preset path loss range, performing coverage prediction processing on the network based on a preset coverage prediction model.
In this embodiment, it may also be determined whether the target predicted path loss is within a preset path loss range, specifically, if the target predicted path loss is within the preset path loss range, performing coverage prediction processing on the base station based on a preset coverage prediction model, specifically, generating an integrated file based on the target predicted path loss, and sending the integrated file to the intermediate station, and performing coverage prediction on the planned base station based on a trained machine learning model.
In this embodiment, the step of determining whether the target predicted path loss is within the preset path loss range may not be performed, and the coverage prediction processing may be directly performed on the network based on the target predicted path loss and the preset coverage prediction model.
In this embodiment, it is further determined whether the target predicted path loss is within a preset path loss range; and if the coverage prediction is within the preset path loss range, performing coverage prediction processing on the network based on a preset coverage prediction model. Therefore, a foundation is laid for network construction.
Further, based on the above embodiment of the present application, another embodiment of the present application is provided, in which the network planning method based on machine learning is applied to a road point, and includes:
step C1, when a network planning instruction is detected, a base station corresponding to the road point is in communication connection so as to receive a wireless signal sent by the base station;
and step C2, acquiring road point information based on the wireless signals, and sending the road point information to the base station so that a preset machine learning model in the base station can conduct prediction processing of propagation link loss based on the road point information to obtain target predicted road loss.
In this embodiment, the network planning method based on machine learning is further applied to a road station, where the road station may be a road test terminal, and when the road test terminal detects a network planning instruction, a base station corresponding to the road station performs communication connection to receive a wireless signal sent by the base station, further, based on the wireless signal, collect road station information such as a position of the road station, and send the road station information to the base station, so that a preset machine learning model in the base station performs prediction processing of propagation link loss based on the road station information to obtain a target predicted path loss. In this embodiment, the road point information is accurately obtained by the road point, so that a foundation is laid for accurately obtaining the target predicted road loss by the preset machine learning model in the base station.
Referring to fig. 3, fig. 3 is a schematic device structure diagram of a hardware running environment according to an embodiment of the present application.
As shown in fig. 3, the machine learning based network planning apparatus may include: a processor 1001, such as a CPU, memory 1005, and a communication bus 1002. Wherein a communication bus 1002 is used to enable connected communication between the processor 1001 and a memory 1005. The memory 1005 may be a high-speed RAM memory or a stable memory (non-volatile memory), such as a disk memory. The memory 1005 may also optionally be a storage device separate from the processor 1001 described above.
Optionally, the machine learning based network planning device may further include a rectangular user interface, a network interface, a camera, an RF (Radio Frequency) circuit, a sensor, an audio circuit, a WiFi module, and the like. The rectangular user interface may include a Display screen (Display), an input sub-module such as a Keyboard (Keyboard), and the optional rectangular user interface may also include a standard wired interface, a wireless interface. The network interface may optionally include a standard wired interface, a wireless interface (e.g., WI-FI interface).
Those skilled in the art will appreciate that the machine learning based network planning device structure shown in fig. 3 does not constitute a limitation of the machine learning based network planning device and may include more or fewer components than shown, or may combine certain components, or a different arrangement of components.
As shown in fig. 3, an operating system, a network communication module, and a machine learning-based network planning program may be included in the memory 1005 as one type of storage medium. An operating system is a program that manages and controls machine-learning based network planning device hardware and software resources, supporting machine-learning based network planning programs and the execution of other software and/or programs. The network communication module is used to enable communication between components within the memory 1005 and other hardware and software in the machine learning based network planning system.
In the machine learning based network planning device shown in fig. 3, a processor 1001 is configured to execute a machine learning based network planning program stored in a memory 1005, to implement the steps of the machine learning based network planning method described in any of the above.
The specific implementation manner of the network planning device based on machine learning is basically the same as that of each embodiment of the network planning method based on machine learning, and is not repeated here.
The application also provides a network planning device based on machine learning, which comprises:
the network planning device based on machine learning is applied to a base station and comprises:
the first determining module is used for determining a road point corresponding to the base station when the network planning instruction is detected so as to obtain azimuth data to be processed corresponding to the network planning instruction;
the second determining module is used for determining to-be-processed environment data based on a preset ray tracing mode and obtaining to-be-processed input data based on the to-be-processed azimuth data and the to-be-processed environment data;
the first input module is used for inputting the input data to be processed into a preset machine learning model, and based on the preset machine learning model, carrying out prediction processing on the propagation link loss from the base station to the road point on the input data to be processed to obtain target prediction road loss;
The preset machine learning model is obtained by performing iterative training on a preset model to be trained based on training data with preset loss labels.
Optionally, the second determining module includes:
the first determining unit is used for extracting azimuth characteristic parameters from the network planning instruction, and extracting to-be-processed characteristic data from the to-be-processed azimuth data and the to-be-processed environment data based on the azimuth characteristic parameters;
and the dimension reduction unit is used for carrying out data dimension reduction processing on the feature data to be processed to obtain input data to be processed.
Optionally, the machine learning based network planning apparatus further includes:
the acquisition module is used for acquiring historical data, carrying out data processing on the historical data based on the preset ray tracing mode to obtain training input data, and acquiring an actual road loss label of the training input data;
the second input module is used for inputting the training input data into a preset model to be trained, and carrying out prediction processing on the training input data based on the preset model to be trained to obtain predicted path loss;
the comparison module is used for comparing the predicted path loss with the actual path loss to obtain a comparison result;
And the return module is used for returning to the step of inputting the training input data into the preset model to be trained until the preset machine learning model is obtained if the preset model to be trained is determined to not complete the preset training condition based on the comparison result.
Optionally, the acquiring module includes:
the first acquisition unit is used for determining simulation data which correspond to base station information, road point information, map information, material configuration information and antenna information in the historical data together according to a preset ray tracking mode, and obtaining training environment characteristics based on the simulation data;
a second determining unit, configured to determine an antenna gain based on the history data and a preset antenna gain dictionary;
the second acquisition unit is used for acquiring historical data and determining the Euclidean distance between the base station and the road point based on the historical data;
a third acquisition unit configured to acquire history data, and determine a relative horizontal angle and a relative vertical angle based on the history data;
and the setting unit is used for setting the training environment characteristics, the antenna gain, the Euclidean distance, the relative horizontal angle and the relative vertical angle as the training input data.
Optionally, the return module includes:
the adjusting unit is used for adjusting weight parameters and bias parameters in the preset machine learning model if the preset model to be trained is determined to not finish the preset training conditions based on the comparison result;
and the return unit is used for returning to the step of inputting the training input data into the adjusted preset model to be trained until the preset machine learning model is obtained.
Optionally, the machine learning based network planning apparatus further includes:
the third determining module is used for determining whether the target predicted path loss is in a preset path loss range;
and the coverage prediction module is used for performing coverage prediction processing on the network based on a preset coverage prediction model if the coverage prediction module is in the preset path loss range.
The specific implementation manner of the network planning device based on machine learning of the present application is basically the same as the above embodiments of the network planning method based on machine learning, and will not be described herein.
The application also provides a network planning device based on machine learning, which is applied to the road measuring points, and the network planning device based on machine learning comprises:
the detection module is used for carrying out communication connection with the base station corresponding to the road point when the network planning instruction is detected so as to receive the wireless signal sent by the base station;
The acquisition module is used for acquiring the road point information based on the wireless signals and sending the road point information to the base station so that a preset machine learning model in the base station can conduct prediction processing of propagation link loss based on the road point information to obtain target predicted road loss.
The specific implementation manner of the network planning device based on machine learning of the present application is basically the same as the above embodiments of the network planning method based on machine learning, and will not be described herein.
Embodiments of the present application provide a storage medium, and the storage medium stores one or more programs, which are further executable by one or more processors for implementing the steps of the machine learning based network planning method described in any of the above.
The specific implementation manner of the storage medium of the present application is basically the same as the above embodiments of the network planning method based on machine learning, and will not be described herein again.
The application also provides a computer program product comprising a computer program which, when executed by a processor, implements the steps of the machine learning based network planning method described above.
The specific implementation manner of the computer program product of the present application is basically the same as the above embodiments of the network planning method based on machine learning, and will not be described herein.
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 apparatus 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 apparatus. 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 apparatus that comprises the element.
The foregoing embodiment numbers of the present application 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 application 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. ROM/RAM, magnetic disk, optical disk) comprising instructions for causing a terminal device (which may be a mobile phone, a computer, a server, an air conditioner, or a network device, etc.) to perform the method according to the embodiments of the present application.
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. A machine learning-based network planning method, applied to a base station, comprising:
when a network planning instruction is detected, determining a road measuring point corresponding to the base station, and determining azimuth data to be processed of the base station and the road measuring point;
predicting signal intensity of a base station based on a preset ray tracking algorithm in a preset ray tracking mode, base station information, road point information, map information, material configuration information and antenna information, generating a simulation intermediate file correspondingly in the process of predicting signal intensity of a receiving point, extracting environment data to be processed from the simulation intermediate file, and obtaining input data to be processed based on the azimuth data to be processed and the environment data to be processed, wherein a direct path corresponds to antenna gain, a transmission path corresponds to antenna gain and a diffraction path corresponds to antenna gain, and Euclidean distance, relative horizontal angle (alpha) and relative vertical angle (beta) belong to azimuth data to be processed; the direct radial order and the transmission radial order, and the diffraction radial order belongs to the environmental data to be processed;
Inputting the input data to be processed into a preset machine learning model, and based on the preset machine learning model, carrying out prediction processing on the propagation link loss from the base station to the road point on the input data to be processed to obtain target predicted road loss;
the preset machine learning model is obtained by performing iterative training on a preset model to be trained based on training data with preset loss labels;
the step of obtaining the input data to be processed based on the azimuth data to be processed and the environmental data to be processed includes:
extracting azimuth characteristic parameters from the network planning instruction, and extracting to-be-processed characteristic data from the to-be-processed azimuth data and the to-be-processed environment data based on the azimuth characteristic parameters;
and performing data dimension reduction processing on the feature data to be processed to obtain input data to be processed.
2. The machine learning based network planning method of claim 1, wherein the step of inputting the input data to be processed into a preset machine learning model, and performing prediction processing of propagation link loss on the input data to be processed based on the preset machine learning model to obtain a target predicted path loss, the method comprises:
Acquiring historical data, performing data processing on the historical data based on the preset ray tracing mode to obtain training input data, and acquiring an actual road loss label of the training input data;
inputting the training input data into a preset model to be trained, and carrying out prediction processing on the training input data based on the preset model to be trained to obtain predicted path loss;
comparing the predicted path loss with the actual path loss to obtain a comparison result;
and based on the comparison result, if the preset training condition of the preset model to be trained is not finished, returning to the step of inputting the training input data into the preset model to be trained until the preset machine learning model is obtained.
3. The machine learning based network planning method of claim 2 wherein the step of obtaining historical data, performing data processing on the historical data based on the preset ray tracing mode, and obtaining training input data comprises the steps of:
according to a preset ray tracking mode, simulation data which correspond to base station information, road point information, map information, material configuration information and antenna information in the historical data together are determined, and training environment characteristics are obtained based on the simulation data;
Determining antenna gain based on the historical data and a preset antenna gain dictionary, wherein the antenna gain dictionary comprises an antenna gain parameter correlation table;
acquiring historical data, and determining Euclidean distance between the base station and the road point based on the historical data;
acquiring historical data, and determining a relative horizontal angle and a relative vertical angle based on the historical data;
and setting the training environment characteristics, the antenna gain, the Euclidean distance, the relative horizontal angle and the relative vertical angle as the training input data.
4. The machine learning based network planning method of claim 2, wherein the step of returning the training input data to the preset model to be trained until the preset machine learning model is obtained if it is determined that the preset model to be trained does not complete the preset training condition based on the comparison result comprises:
based on the comparison result, if the preset training condition of the preset model to be trained is not completed, adjusting weight parameters and bias parameters in the preset machine learning model;
and returning to the step of inputting the training input data into the adjusted preset model to be trained until the preset machine learning model is obtained.
5. The machine learning based network planning method according to any one of claims 1 to 4, wherein after the step of inputting the input data to be processed into a preset machine learning model, performing prediction processing of propagation link loss on the input data to be processed based on the preset machine learning model, and obtaining a target predicted path loss, the method includes:
determining whether the target predicted path loss is within a preset path loss range;
and if the coverage prediction is within the preset path loss range, performing coverage prediction processing on the network based on a preset coverage prediction model.
6. A network planning method based on machine learning, which is applied to a road point, the network planning method based on machine learning comprises:
when a network planning instruction is detected, a base station corresponding to the road measuring point is in communication connection so as to receive a wireless signal sent by the base station;
acquiring road point information based on the wireless signals, and sending the road point information to the base station so that a preset machine learning model in the base station can predict propagation link loss based on the road point information to obtain target predicted road loss;
The base station predicts the signal intensity of the base station based on a preset ray tracking algorithm in a preset ray tracking mode, base station information, road point information, map information, material configuration information and antenna information, generates a simulation intermediate file correspondingly in the process of predicting the signal intensity of a receiving point, extracts environment data to be processed from the simulation intermediate file, and obtains input data to be processed based on the azimuth data to be processed and the environment data to be processed, wherein a direct path corresponds to the antenna gain, a transmission path corresponds to the antenna gain, a diffraction path corresponds to the antenna gain, and a Euclidean distance, a relative horizontal angle (alpha) and a relative vertical angle (beta) belong to the azimuth data to be processed; the direct radial order and the transmission radial order, and the diffraction radial order belongs to the environmental data to be processed; inputting the input data to be processed into a preset machine learning model, and based on the preset machine learning model, carrying out prediction processing on the propagation link loss from the base station to the road point on the input data to be processed to obtain target predicted road loss; the preset machine learning model is obtained by performing iterative training on a preset model to be trained based on training data with preset loss labels;
The step of obtaining the input data to be processed based on the azimuth data to be processed and the environmental data to be processed includes:
extracting azimuth characteristic parameters from the network planning instruction, and extracting to-be-processed characteristic data from the to-be-processed azimuth data and the to-be-processed environment data based on the azimuth characteristic parameters;
and performing data dimension reduction processing on the feature data to be processed to obtain input data to be processed.
7. A machine-learning-based network planning apparatus, the machine-learning-based network planning apparatus comprising:
the first determining module is used for determining a road measuring point corresponding to a base station when a network planning instruction is detected so as to obtain azimuth data to be processed corresponding to the network planning instruction;
the second determining module is used for predicting the signal intensity of the base station based on a preset ray tracing algorithm in a preset ray tracing mode, base station information, road point information, map information, material configuration information and antenna information, generating a simulation intermediate file correspondingly in the process of predicting the signal intensity of the receiving point, extracting environment data to be processed from the simulation intermediate file, and obtaining input data to be processed based on the azimuth data to be processed and the environment data to be processed, wherein the direct path corresponds to the antenna gain, the transmission path corresponds to the antenna gain, the diffraction path corresponds to the antenna gain, and the Euclidean distance, the relative horizontal angle (alpha) and the relative vertical angle (beta) belong to the azimuth data to be processed; the direct radial order and the transmission radial order, and the diffraction radial order belongs to the environmental data to be processed;
The first input module is used for inputting the input data to be processed into a preset machine learning model, and based on the preset machine learning model, carrying out prediction processing on the propagation link loss from the base station to the road point on the input data to be processed to obtain target prediction road loss;
the preset machine learning model is obtained by performing iterative training on a preset model to be trained based on training data with preset loss labels;
the second determining module is further configured to implement:
extracting azimuth characteristic parameters from the network planning instruction, and extracting to-be-processed characteristic data from the to-be-processed azimuth data and the to-be-processed environment data based on the azimuth characteristic parameters;
and performing data dimension reduction processing on the feature data to be processed to obtain input data to be processed.
8. A machine-learning-based network planning device, the machine-learning-based network planning device comprising: a memory, a processor and a program stored on the memory for implementing the machine learning based network planning method,
the memory is used for storing a program for realizing a network planning method based on machine learning;
The processor is configured to execute a program implementing the machine learning based network planning method to implement the steps of the machine learning based network planning method according to any one of claims 1 to 6.
9. A storage medium having stored thereon a program for implementing a machine learning based network planning method, the program for implementing the machine learning based network planning method being executed by a processor to implement the steps of the machine learning based network planning method of any of claims 1 to 6.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202110420946.3A CN115226112B (en) | 2021-04-19 | 2021-04-19 | Network planning method, device, equipment and storage medium based on machine learning |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202110420946.3A CN115226112B (en) | 2021-04-19 | 2021-04-19 | Network planning method, device, equipment and storage medium based on machine learning |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN115226112A CN115226112A (en) | 2022-10-21 |
| CN115226112B true CN115226112B (en) | 2023-09-19 |
Family
ID=83605255
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202110420946.3A Active CN115226112B (en) | 2021-04-19 | 2021-04-19 | Network planning method, device, equipment and storage medium based on machine learning |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN115226112B (en) |
Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101931984A (en) * | 2009-06-18 | 2010-12-29 | 华为技术有限公司 | Method, user equipment and base station for drive test measurement |
| CN108810911A (en) * | 2018-06-04 | 2018-11-13 | 南京邮电大学 | A kind of low-power consumption WAN network planing method based on data mining |
| CN110418354A (en) * | 2019-08-06 | 2019-11-05 | 北京邮电大学 | It is a kind of that propagation model wireless network planning method is exempted from based on machine learning |
| CN111147163A (en) * | 2019-12-17 | 2020-05-12 | 南京航空航天大学 | Wireless communication link loss prediction method based on DNN neural network |
Family Cites Families (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US11157002B2 (en) * | 2017-12-28 | 2021-10-26 | Intel Corporation | Methods, systems, articles of manufacture and apparatus to improve autonomous machine capabilities |
-
2021
- 2021-04-19 CN CN202110420946.3A patent/CN115226112B/en active Active
Patent Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101931984A (en) * | 2009-06-18 | 2010-12-29 | 华为技术有限公司 | Method, user equipment and base station for drive test measurement |
| CN108810911A (en) * | 2018-06-04 | 2018-11-13 | 南京邮电大学 | A kind of low-power consumption WAN network planing method based on data mining |
| CN110418354A (en) * | 2019-08-06 | 2019-11-05 | 北京邮电大学 | It is a kind of that propagation model wireless network planning method is exempted from based on machine learning |
| CN111147163A (en) * | 2019-12-17 | 2020-05-12 | 南京航空航天大学 | Wireless communication link loss prediction method based on DNN neural network |
Non-Patent Citations (6)
| Title |
|---|
| 基于多特征排序模型的网络课程推荐算法研究与应用;朱华;CNKI优秀硕士学位论文全文库;全文 * |
| 基于时间序列分析的网络安全态势预测模型研究;王雪;CNKI优秀硕士学位论文全文库;全文 * |
| 基于深度学习的特征检测与识别;巴特尔;CNKI优秀硕士学位论文全文库;全文 * |
| 强化学习在路测覆盖分析中的实践;杨洁艳;何新平;邓巍;李衡;;信息通信技术(第02期);全文 * |
| 无线通信中新型天线及天线阵技术研究;朱华;CNKI博士学位论文全文库;全文 * |
| 深度文本匹配与排序的研究与实现;刘玮;CNKI优秀硕士学位论文全文库;全文 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN115226112A (en) | 2022-10-21 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| US20230077304A1 (en) | Cooperative positioning method and apparatus, device, and storage medium | |
| Yang et al. | Efficient object localization using sparsely distributed passive RFID tags | |
| CN112218330B (en) | Positioning method and communication device | |
| EP3404439A1 (en) | Cluster-based magnetic positioning method, device and system | |
| CN106454747B (en) | Wireless positioning method of mobile phone terminal | |
| CN108732559B (en) | Positioning method, positioning device, electronic equipment and readable storage medium | |
| CN113329437B (en) | Wireless network signal propagation path loss prediction method and electronic equipment | |
| CN110798886B (en) | Positioning method and device | |
| US9660740B2 (en) | Signal strength distribution establishing method and wireless positioning system | |
| JP6696859B2 (en) | Quality estimation device and quality estimation method | |
| Olama et al. | Position and velocity tracking in mobile networks using particle and Kalman filtering with comparison | |
| CN115226112B (en) | Network planning method, device, equipment and storage medium based on machine learning | |
| JP2020112521A (en) | Position detection system, calculation device, program and position detection method | |
| US11567158B2 (en) | Method of beacon-based positioning system | |
| CN112822626B (en) | Terminal positioning method and device, storage medium and electronic equipment | |
| KR20190007310A (en) | Method and apparatus for measuring location | |
| CN114390534A (en) | Beam pattern determination method, beam pattern determination device, electronic equipment and storage medium | |
| CN115980804B (en) | Indoor positioning method and device | |
| US7203504B2 (en) | Method of building a locating service for a wireless network environment | |
| KR101355363B1 (en) | High-Speed Radio Propagation Analysis Apparatus and Method | |
| CN111464945B (en) | Positioning method and system of terminal equipment | |
| Lewandowski et al. | Support vector machines for non-linear radio fingerprint recognition in real-life industrial environments | |
| CN117671303B (en) | Stereoscopic ranging hidden danger early warning method and system based on twin network | |
| CN116148788B (en) | Radar signal simulation method, device, equipment and storage medium | |
| US20240046146A1 (en) | Method for teaching an electronic computing device, a computer program product, a computer-readable storage medium as well as an electronic computing device |
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 |