A method, apparatus, and device for predicting received power.

Abstract

This invention discloses a method, apparatus, and device for predicting received power. The method includes: acquiring information related to the actual geographical locations of the signal transmitter and receiver in the prediction scenario; converting the information related to the actual geographical locations into corresponding feature information, and then inputting it into a received power prediction model trained based on antenna gain features for processing to obtain the predicted received power; and correcting the predicted received power to obtain the target predicted received power. Through this method, this invention achieves a more accurate and reasonable neural network model for propagation prediction, reducing the computation time of the trained model.

Description

Technical Field

[0001] This invention relates to the field of communication technology, and more specifically to a method, apparatus, and device for predicting received power. Background Technology

[0002] Using AI (Artificial Intelligence) models to predict field strength is a hot topic in the future development of propagation models. Data sets will not be limited to traditional model-corrected CW (continuous wave) data, but will also include MR (Measuring Report) data and 3D ray simulation data. This expanded dataset forms the basis for obtaining more reasonable propagation models and allows for the training of more complex AI models. As data sources expand, inconsistencies in various data conditions will become more prominent. How to train the network and standardize diverse data is crucial for obtaining and validating reasonable models.

[0003] Existing technologies use CW drive test data for propagation model correction, and the antenna pattern is an omnidirectional and symmetrical radiation pattern. This has a limited scope of application and a small amount of data, making it unsuitable for complex models like AI. Summary of the Invention

[0004] In view of the above problems, embodiments of the present invention are proposed to provide a method, apparatus and device for predicting received power that overcomes or at least partially solves the above problems.

[0005] According to one aspect of the present invention, a method for predicting received power is provided, comprising:

[0006] Obtain information related to the actual geographical locations of the signal transmitter and receiver in the predicted scenario;

[0007] After converting the actual geographical location information into corresponding feature information, it is input into the received power prediction model trained based on antenna gain features for processing to obtain the predicted received power.

[0008] The predicted received power is corrected to obtain the target predicted received power.

[0009] According to another aspect of the present invention, a receiving power prediction device is provided, comprising:

[0010] The acquisition module is used to acquire information related to the actual geographical location of the signal transmitter and receiver in the predicted scenario.

[0011] The processing module is used to convert the actual geographical location information into corresponding feature information, input it into the received power prediction model trained based on antenna gain features for processing, and obtain the predicted received power; the predicted received power is then corrected to obtain the target predicted received power.

[0012] According to another aspect of the present invention, a computing device is provided, comprising: a processor, a memory, a communication interface, and a communication bus, wherein the processor, the memory, and the communication interface communicate with each other through the communication bus;

[0013] The memory is used to store at least one executable instruction that causes the processor to perform the operation corresponding to the above-described method for predicting received power.

[0014] According to another aspect of the present invention, a computer storage medium is provided, the storage medium storing at least one executable instruction that causes a processor to perform an operation corresponding to the above-described method for predicting received power.

[0015] According to the solution provided in the above embodiments of the present invention, by obtaining the actual geographical location information of the signal transmitter and the signal receiver in the prediction scenario; converting the actual geographical location information into corresponding feature information, and inputting it into the received power prediction model trained based on antenna gain characteristics for processing, the predicted received power is obtained; the predicted received power is corrected to obtain the target predicted received power; a more accurate received power prediction model can be trained, meeting the requirements of signal propagation. The computation time of the model during received power prediction is reduced.

[0016] The above description is merely an overview of the technical solutions of the embodiments of the present invention. In order to better understand the technical means of the embodiments of the present invention and to implement them in accordance with the contents of the specification, and to make the above and other objects, features and advantages of the embodiments of the present invention more apparent and understandable, specific implementation methods of the embodiments of the present invention are described below. Attached Figure Description

[0017] Various other advantages and benefits will become apparent to those skilled in the art upon reading the following detailed description of preferred embodiments. The accompanying drawings are for illustrative purposes only and are not intended to limit the embodiments of the invention. Furthermore, the same reference numerals denote the same parts throughout the drawings. In the drawings:

[0018] Figure 1 A flowchart of the received power prediction method provided in an embodiment of the present invention is shown;

[0019] Figure 2 A flowchart illustrating the training process of the received power prediction model provided in an embodiment of the present invention is shown.

[0020] Figure 3 The diagram shows the training network structure of the received power prediction model provided in an embodiment of the present invention.

[0021] Figure 4A scatter plot showing the relationship between the received power and distance of the omnidirectional antenna provided in an embodiment of the present invention is shown.

[0022] Figure 5 A scatter plot showing the relationship between the received power and antenna gain of the omnidirectional antenna provided in an embodiment of the present invention is shown.

[0023] Figure 6 The diagram shows the mean value of the relationship between the received power and distance of the omnidirectional antenna provided in the embodiment of the present invention;

[0024] Figure 7 The diagram shows the mean value of the relationship between the received power and antenna gain of the omnidirectional antenna provided in the embodiment of the present invention.

[0025] Figure 8 A scatter plot showing the relationship between the received power and distance of the directional antenna provided in an embodiment of the present invention is shown.

[0026] Figure 9 A scatter plot showing the relationship between the received power and antenna gain of the directional antenna provided in an embodiment of the present invention is shown.

[0027] Figure 10 A mean value graph showing the relationship between the received power and distance of the directional antenna provided in an embodiment of the present invention is shown.

[0028] Figure 11 The diagram shows the mean value of the relationship between the received power and antenna gain of the directional antenna provided in the embodiment of the present invention.

[0029] Figure 12 A schematic diagram of the receiving power prediction device provided in an embodiment of the present invention is shown;

[0030] Figure 13 A schematic diagram of the structure of a computing device provided in an embodiment of the present invention is shown. Detailed Implementation

[0031] Exemplary embodiments of the invention will now be described in more detail with reference to the accompanying drawings. While exemplary embodiments of the invention are shown in the drawings, it should be understood that the invention may be implemented in various forms and should not be limited to the embodiments set forth herein. Rather, these embodiments are provided so that this invention will be thorough and complete, and will fully convey the scope of the invention to those skilled in the art.

[0032] Figure 1 A flowchart of the received power prediction method provided by an embodiment of the present invention is shown. Figure 1 As shown, the method includes the following steps:

[0033] Step 11: Obtain information related to the actual geographical locations of the signal transmitter and receiver in the predicted scenario;

[0034] Step 12: After converting the actual geographical location information into corresponding feature information, input it into the received power prediction model trained based on the antenna gain features for processing to obtain the predicted received power.

[0035] Step 13: Correct the predicted received power to obtain the target predicted received power.

[0036] In the above embodiments of the present invention, the actual geographical location information of the signal transmitter and receiver in the predicted scenario is obtained; after converting the actual geographical location information into corresponding feature information, it is input into the received power prediction model trained based on antenna gain characteristics for processing to obtain the predicted received power; the predicted received power is then corrected to obtain the target predicted received power. This allows for the training of a more accurate received power prediction model, meeting the requirements of signal propagation and reducing the computation time of the model during received power prediction.

[0037] In an optional embodiment of the present invention, step 11, predicting the geographic location-related information of the signal transmitter and the signal receiver in the scenario includes at least one of the following:

[0038] The distance between the signal transmitter and the signal receiver;

[0039] Antenna gain between the signal transmitter and the signal receiver;

[0040] The relative height of the transmitting and receiving antennas between the signal transmitting end and the signal receiving end;

[0041] The type of grid feature closest to the signal receiver on the line connecting the signal transmitter and the signal receiver;

[0042] The signal transmitter and receiver belong to indoor or outdoor user information;

[0043] Information on occlusion of nearby ground features;

[0044] The indoor or outdoor distance between the signal transmitter and the signal receiver;

[0045] The distance lengths of various land cover types along the line connecting the signal transmitter and receiver.

[0046] In this embodiment, the geographic location information of the signal transmitter and receiver in the prediction scenario is not limited to the information mentioned above, but may also include other geographic location-related information; wherein, converting each type of information into corresponding feature information may include:

[0047] 1) The feature information corresponding to the distance between the signal transmitter and the signal receiver includes: log(d), where d is the distance between the signal transmitter and the signal receiver;

[0048] Specifically, d represents the Euclidean distance between the signal transmitter and the signal receiver. The distance feature value is log(d), the data dimension is 1x1, and the data type is float.

[0049] 2) The characteristic information corresponding to the antenna gain between the signal transmitter and the signal receiver includes: the gain G(θ,β) corresponding to the horizontal and vertical angle reading of the antenna pattern, where θ is the gain corresponding to the horizontal angle reading of the antenna pattern and β is the gain corresponding to the vertical angle reading of the antenna pattern.

[0050] Specifically, the antenna gain characteristic value is G(θ, β), the data dimension is 1X1, and the type is float.

[0051] 3) The characteristic information corresponding to the relative height of the transmitting and receiving antennas between the signal transmitting end and the signal receiving end includes: log(1+H), where H=H Tx -H Rx H TR The absolute altitude of the transmitting end is: sea level at the transmitting end + terrain elevation at the transmitting end + building height at the transmitting end; H Rx Absolute height of the receiving end: sea level at the signal receiving end + terrain elevation at the signal receiving end + building height at the signal receiving end;

[0052] Specifically, the building height includes the antenna mounting height, and the relative height feature value is log(1+H), the data dimension is 1X1, and the type is float.

[0053] 4) On the distance line connecting the signal transmitter and the signal receiver, the feature information corresponding to the nearest preset number of raster land cover types to the signal receiver includes: the land cover type code corresponding to the preset number of raster land cover types;

[0054] Specifically, the preset number of raster land cover types closest to the signal receiver are encoded according to the land cover type encoding table for their respective locations. Starting from the receiver, the vector groups from closest to furthest are used sequentially as feature values ​​for the nearest land cover types and input into an integer vector in the neural network. The land cover type encoding uses a one-hot orthogonal vector encoding method. The nearest land cover type refers to the m raster land cover types closest to the receiver on the Euclidean distance line connecting the transmitter and receiver, where m is a positive integer. The feature value for the nearest land cover type is: `near_vector`, the vector length refers to the number of land cover types, the data dimension is 1 x (m * n), and the type is: `int`, where m is the number of nearest land cover types and n is the number of vector lengths.

[0055] The following example, using a map accuracy of 5 meters, illustrates how to obtain the feature values ​​of nearby ground features. First, after the nearby grid cells are encoded according to the local ground feature type encoding table (as shown in Table 1), five nearby grid cells are used as input, with the receiving end as the starting point, under the condition of a map accuracy of 5 meters. That is, five one-hot vectors of length 20 are input. Considering the geographical information within a 25-meter range of the receiving end, if the distance between the sending and receiving ends is less than 25 meters, zero vectors are used to fill in the gap. This ensures that the neural network is fixed. Finally, the example input feature vector is obtained, as shown in Table 2.

[0056]

[0057]

[0058] Table 1 - Land Feature Types and Land Feature Codes of a Certain City

[0059]

[0060]

[0061] Table 2 - Example Input Feature Vectors

[0062] 5) The characteristic information corresponding to indoor or outdoor user information at the signal transmitting end and signal receiving end includes: a first identifier and / or a second identifier, wherein the user with the first identifier is an indoor user and the user with the second identifier is an outdoor user;

[0063] Specifically, the indoor or outdoor user feature value is in_out_door, where 0 represents the outdoor user feature value and 1 represents the indoor user feature value. The data dimension is 1X1 and the type is int. Furthermore, when collecting MR data for training, or when testing CW data and 3D ray simulation data for model calibration, 0 is still used as the outdoor user feature value input and 1 is used as the indoor user feature value input.

[0064] 6) The feature information corresponding to the occlusion information of nearby ground features includes: the first value when there is occlusion at the signal receiver and the signal transmitter; and the second value when there is no occlusion at the signal receiver and the signal transmitter.

[0065] Specifically, since the occlusion of nearby ground features has a significant impact on signal propagation, it is also considered as a feature affecting signal propagation. The feature value for the occlusion of nearby ground features is "block," where 1 indicates that there is an occlusion feature value at both the signal receiver and the signal transmitter, and 0 indicates that there is no occlusion feature value at either the signal receiver or the signal transmitter. The data dimension is 1x1, and the type is int.

[0066] 7) The feature information corresponding to the indoor or outdoor distance between the signal transmitter and the signal receiver includes: (D0=log(1+∑d0), D1=log(1+∑d1)), where d0 represents the unobstructed distance, d1 represents the obstructed distance, D0 represents the feature information corresponding to the outdoor distance between the signal transmitter and the signal receiver, and D1 represents the feature information corresponding to the indoor distance between the signal transmitter and the signal receiver;

[0067] Specifically, since building obstruction causes signal attenuation, the indoor or outdoor distance between the signal transmitter and receiver is also considered as a feature affecting signal propagation. The feature value of the indoor or outdoor distance between the signal transmitter and receiver is in_out_door_distance, where D0 = log(1 + ∑d0) represents the outdoor distance feature value between the signal transmitter and receiver, and D1 = log(1 + ∑d1) represents the indoor distance feature value between the signal transmitter and receiver. The data dimension is 1x2, and the type is float.

[0068] 8) The feature information corresponding to the distance lengths of various land cover types along the distance line connecting the signal transmitter and receiver includes: (log(1+D1), log(1+D2), ..., log(1+D... m Where D1 represents the numerical value corresponding to the distance of the first type of land cover to the line connecting the signal transmitter and the signal receiver, D2 represents the numerical value corresponding to the distance of the second type of land cover to the line connecting the signal transmitter and the signal receiver, ..., D m This represents the numerical value corresponding to the distance of the m-th land feature type to the line connecting the signal transmitter and the signal receiver.

[0069] Specifically, the distance line connecting the signal transmitter and receiver passes through various land cover types. The length of each land cover type is statistically analyzed and used as the feature information corresponding to the distance length of each land cover type on the distance line between the signal transmitter and receiver. If a land cover type not listed in the local land cover type encoding table exists on the distance line between the signal transmitter and receiver, then the feature information of this land cover type and its corresponding length is set to 0. The distance length feature value for each land cover type on the distance line between the signal transmitter and receiver is: clutter_distance, where log(1+D1) represents the distance length feature value of the first type of land cover on the distance line between the signal transmitter and receiver, log(1+D2) represents the distance length feature value of the second type of land cover on the distance line between the signal transmitter and receiver, and log(1+D2) represents the distance length feature value of the second type of land cover on the distance line between the signal transmitter and receiver. m) represents the distance length feature value of the m-th land cover type on the distance line connecting the signal transmitter and the signal receiver, where m is the m-th land cover type, the data dimension is 1Xm, and the type is float.

[0070] In this embodiment, since the logarithmic function is a monotonically increasing function within its domain, taking the logarithm does not change the relative relationship of the eigenvalues. Therefore, the eigenvalues ​​can be logarithmed, which reduces the absolute value of the eigenvalues ​​and transforms multiplication into addition, simplifying the calculation. Furthermore, to avoid the logarithm being zero, the variable whose logarithm is taken is uniformly incremented by 1 to ensure that the eigenvalue is greater than or equal to 0.

[0071] In another optional embodiment of the present invention, in step 12, the received power prediction model is trained through the following process:

[0072] Step 121: Obtain training set data for training the received power prediction model in the training scenario. The training set data includes training data for omnidirectional antennas and training data for directional antennas.

[0073] Step 122: Obtain the gain features corresponding to the training data of the omnidirectional antenna, and input the gain features corresponding to the training data of the omnidirectional antenna into the received power prediction model for training processing to obtain the first training model;

[0074] Step 123: Obtain the gain features corresponding to the training data of the directional antenna, and input the gain features corresponding to the training data of the directional antenna into the first training model for training processing to obtain the second training model;

[0075] Step 124: Obtain the randomized sequence of the gain features corresponding to the training data of the omnidirectional antenna and the gain features corresponding to the training data of the directional antenna, and input the randomized sequence into the second training model to obtain the received power prediction model.

[0076] like Figure 2 As shown, in this embodiment, the specific steps are as follows:

[0077] The first step is to obtain the training set data for training the received power prediction model in the training scenario.

[0078] The second step is to input the gain features corresponding to the training data of the omnidirectional antenna in the training set into the received power prediction model for training processing to obtain the first training model. Here, the training of the omnidirectional antenna is based on the gain features related to distance d.

[0079] The third step is to input the gain features corresponding to the training data of the directional antenna in the training set into the first training model to obtain the second training model. The training of the directional antenna is based on the gain features related to the antenna gain G(θ,β).

[0080] The fourth step is to obtain the randomized sequences of the gain features corresponding to the training data of the omnidirectional antenna and the gain features corresponding to the training data of the directional antenna, and input the randomized sequences into the second training model for overall training.

[0081] Fifth step: when the randomized sequence is fully converged by the second training model, the training is completed and the received power prediction model is obtained. If it is not fully converged, the training is restarted until it is fully converged. The received power prediction model has better overall performance and can fully reflect the control characteristics of antenna gain and distance.

[0082] In the above embodiments, when inputting training set data into the model for training, the training set data needs to be standardized. Specifically, the following four factors need to be considered:

[0083] (a) Factors affecting transmission power

[0084] Transmit power is not used as input to the neural network corresponding to the received power prediction model, but transmit power does affect received power. Given CW transmit power p1, received power r1, 3D ray simulation transmit power p2, received power r2, and MR measured cell transmit power p3, received power r3, the transmit power needs to be adjusted to a standard, and the received power adjusted accordingly, such as:

[0085] When using CW transmission power as the benchmark, the adjustment amount for r2 in the 3D ray simulation data is p1-p2, r2 = r2 + (p1-p2); similarly, the adjustment amount for r3 in the MR data is p1-p3, r3 = r3 + (p1-p3). Power must be unified first, and the receiving power adjusted before proper training can be performed.

[0086] (ii) Antenna and distance factors

[0087] Antenna gain, as input to a neural network, needs to be varied to train for effective results. However, CW testing uses an omnidirectional antenna, whose gain can be considered constant, making it impossible to train for effective antenna gain. To train for effective antenna gain, three-dimensional ray-tracing simulation results of both directional and omnidirectional antennas are needed as input. The simulation results of the directional antenna are used as the training set, with various combinations of antenna angles: horizontal angles of 0, 90, 180, and 270 degrees; and vertical angles of 0, 5, 10, 15, and 20 degrees.

[0088] (III) Multiple height factors

[0089] The transmitter height is used as input to the neural network. The height needs to vary to train the neural network effectively. However, a single-station CW test only has one height value, and even with joint calibration from three stations, there are only three values. Therefore, it's impossible to train a suitable height-based model. Thus, we rely on 3D ray simulation results, using simulations with heights of 15, 20, 25, 30, 35, 40, 45, and 50 meters as the input to the training set.

[0090] The antenna height of a typical cell in a city is concentrated between 25 meters and 40 meters, showing a certain degree of diversity. Therefore, using the height of each cell as input for MR training can also be effective.

[0091] (iv) Indoor / Outdoor Factors

[0092] Both the CW and 3D ray tracing propagation models are for outdoor scenes, and only the MR dataset can distinguish between indoor and outdoor environments. When using MR data as the training set, indoor / outdoor identifiers should be extracted as input.

[0093] Specifically, the data is organized into the order shown in Table 3.

[0094]

[0095]

[0096] Table 3

[0097] In another optional embodiment of the present invention, in step 12, the received power prediction model is:

[0098] R = f1(log(d)) + f2(G(θ,β)) + f n (log(d),G(θ,β),log(1+H),near_vector,in_out_door,block,D0,D1, clutter_distance)

[0099] Where R is the predicted received power, f1(log(d)) is the range characteristic value, f2(G(θ,β)) is the antenna gain characteristic value, and f n(log(d)), G(θ,β), log(1+H) are the relative height feature values, near_vector is the feature value of the nearby ground feature type, in_out_door is the feature value of the indoor or outdoor user, block is the feature value of the occlusion of the nearby ground feature type, D0 and D1 are the feature values ​​of the indoor or outdoor distance between the signal transmitter and the signal receiver, clutte_distance is the distance length feature value of various ground feature types on the distance line connecting the signal transmitter and the signal receiver, d is the distance between the signal transmitter and the signal receiver, G is the gain corresponding to the horizontal and vertical angle reading of the antenna pattern, θ is the gain corresponding to the horizontal angle reading of the antenna pattern, β is the gain corresponding to the vertical angle reading of the antenna pattern, and H = H Tx -H Rx H Tx The absolute altitude of the transmitting end is: sea level at the transmitting end + terrain elevation at the transmitting end + building height at the transmitting end; H Rx Absolute height of the receiving end: sea level at the signal receiving end + terrain height at the signal receiving end + building height at the signal receiving end.

[0100] like Figure 3 As shown, the received power prediction model is a function of the multilayer perceptron in the figure:

[0101] R=f(log(d),G(θ,β),log(1+H),near_vector,in_out_door,block,D0,D1,clutter_distance)

[0102] Among them, the predicted received power R is a function of the above 8 characteristics.

[0103] In this diagram, the predicted received power is used as the training objective. Labeled supervised learning is employed, with the feature information corresponding to the training set data used as input to the multilayer perceptron in the neural network. The output is the received power of the signal receiver in the training scenario. After obtaining the received power of the signal receiver in the training scenario, the labels of the test data are then... The input is the MMSE loss function, used for supervised learning of the received power at the signal receiver in the training scenario. This allows the multilayer perceptron in the neural network to learn the impact of the distance and antenna gain information of the actual geographical locations of the signal transmitter and receiver on the actual received power field strength, including the labels of the test data. This represents the actual received power of the signal receiver in the training scenario.

[0104] In practical applications, it is expected that the multilayer perceptron can embody control functions. These control functions are implemented through steps 121 to 124 as described above, and the model is expected to decompose into the following form:

[0105] R = f1(log(d)) + f2(G(θ,β)) + f n (log(d),G(θ,β),log(1+H),near_vector,in_out_door,block,D0,D1, clutter_distance)

[0106] Furthermore, the output received power responds accordingly to changes in distance and antenna gain, specifically as follows:

[0107] R=f1(log(d+Δ d ))+f2(G(θ,β)+Δ G )+f n (log(d+Δ d ),G(θ,β)+Δ G ,log(1+H),near_vector,in_out_door,block,D0,D1, clutter_distance)

[0108] In addition to the output received power response mentioned above, there will also be a characteristic response, specifically: R = f1(log(d)) + Δ f1,d +f2(G(θ,β))+Δ f2,G +f n (log(d+Δ d ),G(θ,β)+Δ G ,log(1+H),near_vector,in_out_door,block,D0,D1, clutter_distance)

[0109] Where, Δ f1,d and Δ f2,G It is the received power that responds accordingly to changes in distance and antenna gain.

[0110] In yet another optional embodiment of the invention, the dataset of the omnidirectional antenna is used to train f1(log(d+Δ)). d The parameters in ));

[0111] The dataset for directional antennas is used to train the parameters in f2(G(θ,β));

[0112] Where, f1(log(d+Δ) df2(G(θ,β)) is a distance-related eigenvalue parameter, f2(G(θ,β)) is a antenna gain-related eigenvalue parameter, d is the distance between the signal transmitter and the signal receiver, and Δ is the distance between the transmitter and receiver. d Let G be the varying distance between the signal transmitter and the signal receiver, G be the gain corresponding to the horizontal and vertical angle readings of the antenna pattern, θ be the gain corresponding to the horizontal angle readings of the antenna pattern, and β be the gain corresponding to the vertical angle readings of the antenna pattern.

[0113] In this embodiment, the CW data is a test dataset under omnidirectional antenna conditions, and the antenna gain can be approximated as constant; the three-dimensional ray simulation dataset of the omnidirectional antenna can also be approximated as constant antenna gain. The dataset of the omnidirectional antenna shows the correlation between received power and distance well; while the three-dimensional ray simulation dataset of the directional antenna can show the correlation between received power and antenna gain.

[0114] like Figures 4-11 The data feature description shown intuitively demonstrates that the omnidirectional antenna dataset is suitable for training f1(log(d+Δ)). d The parameters in f1(log(d+Δ)) are suitable for training, while the dataset for directional antennas is suitable for training the parameters in f2(G(θ,β)). However, directly mixing the two for training is unlikely to achieve the desired results. For example, a higher antenna gain at a greater distance results in a higher field strength, while a lower antenna gain at a closer distance results in a lower field strength. This phenomenon will affect f1(log(d+Δ)). d The parameter training of f2(G(θ,β)) can cause confusion; the same confusion can also occur during parameter training of f2(G(θ,β)). However, the training method for the received power prediction model described in steps 121 to 124 above can avoid this confusion.

[0115] In another optional embodiment of the present invention, step 13 may include:

[0116] Step 131: Obtain the correction amount of the predicted received power based on the actual transmit power of the signal transmitter in the predicted scenario and the transmit power of the signal transmitter in the training scenario.

[0117] Step 132: The predicted received power is superimposed with the correction amount to obtain the target predicted received power.

[0118] In this embodiment, during actual prediction, the predicted received power of the hidden layer of the multilayer perceptron in the neural network model is superimposed with a correction amount, which can make the predicted received power of the target more accurate. The configuration of the hidden layer of the multilayer perceptron is not complex; the network width can be set to no more than 4 times the input dimension, the depth can be no more than 5 layers, and there are no restrictions on the number of neurons.

[0119] In yet another optional embodiment of the invention, step 1311 may include:

[0120] Step 1311: The difference between the actual transmission power of the signal transmitter in the predicted scenario and the transmission power of the signal transmitter in the training scenario is used as the correction amount for the predicted received power.

[0121] In this embodiment, since the transmit power is not used as the input to the neural network, and the influence of power changes linearly, a correction amount needs to be superimposed when actually predicting the received power. The correction amount is the difference between the actual transmit power of the signal transmitter in the prediction scenario and the transmit power of the signal transmitter in the training scenario.

[0122] In the above embodiments of the present invention, antenna gain is used for model training, which can obtain a more accurate and reasonable received power prediction model for predicting received power.

[0123] Figure 12 A schematic diagram of the receiving power prediction device 120 provided in an embodiment of the present invention is shown. Figure 12 As shown, the device includes:

[0124] The acquisition module 121 is used to acquire information related to the actual geographical location of the signal transmitter and the signal receiver in the predicted scenario;

[0125] Processing module 122 is used to convert the actual geographical location information into corresponding feature information, input it into the received power prediction model trained based on antenna gain features for processing, and obtain the predicted received power; and correct the predicted received power to obtain the target predicted received power.

[0126] Optionally, the geographic location-related information of the signal transmitter and receiver in the predicted scenario includes at least one of the following:

[0127] The distance between the signal transmitter and the signal receiver;

[0128] Antenna gain between the signal transmitter and the signal receiver;

[0129] The relative height of the transmitting and receiving antennas between the signal transmitting end and the signal receiving end;

[0130] The type of grid feature closest to the signal receiver on the line connecting the signal transmitter and the signal receiver;

[0131] The signal transmitter and receiver belong to indoor or outdoor user information;

[0132] Information on occlusion of nearby ground features;

[0133] The indoor or outdoor distance between the signal transmitter and the signal receiver;

[0134] The distance lengths of various land cover types along the line connecting the signal transmitter and receiver.

[0135] Optionally, the received power prediction model obtained by training based on antenna gain characteristics is trained through the following process:

[0136] Acquire training set data for training the received power prediction model in the training scenario, wherein the training set data includes training data for omnidirectional antennas and training data for directional antennas;

[0137] The gain features corresponding to the training data of the omnidirectional antenna are obtained, and the gain features corresponding to the training data of the omnidirectional antenna are input into the received power prediction model for training processing to obtain the first training model.

[0138] Obtain the gain features corresponding to the training data of the directional antenna, and input the gain features corresponding to the training data of the directional antenna into the first training model for training processing to obtain the second training model;

[0139] Obtain randomized sequences of the gain features corresponding to the training data of the omnidirectional antenna and the gain features corresponding to the training data of the directional antenna, and input the randomized sequences into the second training model to obtain the received power prediction model.

[0140] Optionally, the received power prediction model is:

[0141] R = f1(log(d)) + f2(G(θ,β)) + f n (log(d),G(θ,β),log(1+H),near_vector,in_out_door,block,D0,D1, clutter_distance)

[0142] Where R is the predicted received power, f1(log(d)) is the range characteristic value, f2(G(θ,β)) is the antenna gain characteristic value, and f n(log(d)), G(θ,β), log(1+H) are the relative height feature values, near_vector is the feature value of the nearby ground feature type, in_out_door is the feature value of the indoor or outdoor user, block is the feature value of the occlusion of the nearby ground feature type, D0 and D1 are the feature values ​​of the indoor or outdoor distance between the signal transmitter and the signal receiver, clutte_distance is the distance length feature value of various ground feature types on the distance line connecting the signal transmitter and the signal receiver, d is the distance between the signal transmitter and the signal receiver, G is the gain corresponding to the horizontal and vertical angle reading of the antenna pattern, θ is the gain corresponding to the horizontal angle reading of the antenna pattern, β is the gain corresponding to the vertical angle reading of the antenna pattern, and H = H Tx -H Rx H Tx The absolute altitude of the transmitting end is: sea level at the transmitting end + terrain elevation at the transmitting end + building height at the transmitting end; H Rx Absolute height of the receiving end: sea level at the signal receiving end + terrain height at the signal receiving end + building height at the signal receiving end.

[0143] Optionally, the dataset of omnidirectional antennas is used to train f1(log(d+Δ)). d The parameters in ));

[0144] The dataset for directional antennas is used to train the parameters in f2(G(θ,β));

[0145] Where, f1(log(d+Δ) d f2(G(θ,β)) is a distance-related eigenvalue parameter, f2(G(θ,β)) is a antenna gain-related eigenvalue parameter, d is the distance between the signal transmitter and the signal receiver, and Δ is the distance between the transmitter and receiver. d Let G be the varying distance between the signal transmitter and the signal receiver, G be the gain corresponding to the horizontal and vertical angle readings of the antenna pattern, θ be the gain corresponding to the horizontal angle readings of the antenna pattern, and β be the gain corresponding to the vertical angle readings of the antenna pattern.

[0146] Optionally, the processing module 122 is further configured to obtain a correction amount for the predicted received power based on the actual transmit power of the signal transmitter in the prediction scenario and the transmit power of the signal transmitter in the training scenario.

[0147] The predicted received power is superimposed with the correction amount to obtain the target predicted received power.

[0148] Optionally, the processing module 122 is further configured to use the difference between the actual transmission power of the signal transmitter in the prediction scenario and the transmission power of the signal transmitter in the training scenario as the correction amount for the predicted received power.

[0149] It should be noted that this embodiment is a device embodiment corresponding to the above method embodiment. All implementation methods in the above method embodiment are applicable to this device embodiment and can achieve the same technical effect.

[0150] This invention provides a non-volatile computer storage medium storing at least one executable instruction that can execute the received power prediction method in any of the above method embodiments.

[0151] Figure 13 The diagram shows a structural schematic of a computing device provided in an embodiment of the present invention. The specific embodiments of the present invention do not limit the specific implementation of the computing device.

[0152] like Figure 13 As shown, the computing device may include a processor, a communications interface, memory, and a communications bus.

[0153] The processor, communication interface, and memory communicate with each other via a communication bus. The communication interface is used to communicate with other network elements, such as clients or other servers. The processor executes programs, specifically the steps described in the embodiment of the method for predicting the received power of the device.

[0154] Specifically, the program may include program code, which includes computer operation instructions.

[0155] The processor may be a central processing unit (CPU), an application-specific integrated circuit (ASIC), or one or more integrated circuits configured to implement embodiments of the present invention. The computing device includes one or more processors, which may be processors of the same type, such as one or more CPUs; or processors of different types, such as one or more CPUs and one or more ASICs.

[0156] Memory is used to store programs. Memory may include high-speed RAM, and may also include non-volatile memory, such as at least one disk drive.

[0157] Specifically, the program can be used to cause the processor to execute the received power prediction method in any of the above method embodiments. The specific implementation of each step in the program can be found in the corresponding descriptions of the steps and units in the above-described received power prediction method embodiments, and will not be repeated here. Those skilled in the art will clearly understand that, for the sake of convenience and brevity, the specific working process of the devices and modules described above can be referred to the corresponding process descriptions in the foregoing method embodiments, and will not be repeated here.

[0158] The algorithms or displays provided herein are not inherently related to any particular computer, virtual system, or other device. Various general-purpose systems can also be used in conjunction with the teachings herein. The required structure for constructing such systems is apparent from the above description. Furthermore, the embodiments of the present invention are not directed to any particular programming language. It should be understood that the embodiments of the present invention described herein can be implemented using various programming languages, and the above description of specific languages ​​is for the purpose of disclosing the best mode of implementation of the embodiments of the present invention.

[0159] Numerous specific details are set forth in the specification provided herein. However, it will be understood that embodiments of the invention may be practiced without these specific details. In some instances, well-known methods, structures, and techniques have not been shown in detail so as not to obscure the understanding of this specification.

[0160] Similarly, it should be understood that, in order to streamline the embodiments of the invention and aid in understanding one or more of the various inventive aspects, features of the embodiments of the invention are sometimes grouped together in a single embodiment, figure, or description thereof in the above description of exemplary embodiments of the invention. However, this disclosure should not be construed as reflecting an intention that the claimed embodiments of the invention require more features than are expressly recited in each claim. Rather, as reflected in the following claims, inventive aspects lie in fewer than all features of a single foregoing disclosed embodiment. Therefore, the claims following the detailed description are hereby expressly incorporated into that detailed description, wherein each claim itself is a separate embodiment of the invention.

[0161] Those skilled in the art will understand that modules in the device of the embodiments can be adaptively changed and placed in one or more devices different from that embodiment. Modules, units, or components in the embodiments can be combined into a single module, unit, or component, and further, they can be divided into multiple sub-modules, sub-units, or sub-components. Except where at least some of such features and / or processes or units are mutually exclusive, any combination can be used to combine all features disclosed in this specification (including the accompanying claims, abstract, and drawings) and all processes or units of any method or device so disclosed. Unless expressly stated otherwise, each feature disclosed in this specification (including the accompanying claims, abstract, and drawings) may be replaced by an alternative feature that serves the same, equivalent, or similar purpose.

[0162] Furthermore, those skilled in the art will understand that although some embodiments herein include certain features included in other embodiments but not others, combinations of features from different embodiments are intended to be within the scope of the invention and form different embodiments. For example, in the following claims, any of the claimed embodiments can be used in any combination.

[0163] The various component embodiments of the present invention can be implemented in hardware, or as software modules running on one or more processors, or a combination thereof. Those skilled in the art will understand that microprocessors or digital signal processors (DSPs) can be used in practice to implement some or all of the functions of some or all of the components according to the embodiments of the present invention. The embodiments of the present invention can also be implemented as device or apparatus programs (e.g., computer programs and computer program products) for performing part or all of the methods described herein. Such programs implementing the embodiments of the present invention can be stored on a computer-readable medium, or can be in the form of one or more signals. Such signals can be downloaded from an Internet website, provided on a carrier signal, or provided in any other form.

[0164] It should be noted that the above embodiments are illustrative of the present invention and not restrictive of the invention, and that those skilled in the art can devise alternative embodiments without departing from the scope of the appended claims. In the claims, any reference signs placed between parentheses should not be construed as limiting the claims. The word "comprising" does not exclude the presence of elements or steps not listed in the claims. The word "a" or "an" preceding an element does not exclude the presence of a plurality of such elements. Embodiments of the present invention can be implemented by means of hardware comprising several different elements and by means of a suitably programmed computer. In the unit claims enumerating several means, several of these means may be embodied by the same item of hardware. The use of the words first, second, and third, etc., does not indicate any order. These words can be interpreted as names. The steps in the above embodiments, unless otherwise specified, should not be construed as limiting the order of execution.

Claims

1. A method for predicting received power, characterized in that, The method includes: Obtain information related to the actual geographical locations of the signal transmitter and receiver in the predicted scenario; After converting the actual geographical location information into corresponding feature information, it is input into the received power prediction model trained based on antenna gain features for processing to obtain the predicted received power. The predicted received power is corrected to obtain the target predicted received power; The received power prediction model obtained by training based on antenna gain characteristics is trained through the following process: Acquire training set data for training the received power prediction model in the training scenario, wherein the training set data includes training data for omnidirectional antennas and training data for directional antennas; The gain features corresponding to the training data of the omnidirectional antenna are obtained, and the gain features corresponding to the training data of the omnidirectional antenna are input into the received power prediction model for training processing to obtain the first training model. Obtain the gain features corresponding to the training data of the directional antenna, and input the gain features corresponding to the training data of the directional antenna into the first training model for training processing to obtain the second training model; Obtain randomized sequences of the gain features corresponding to the training data of the omnidirectional antenna and the gain features corresponding to the training data of the directional antenna, and input the randomized sequences into the second training model to obtain the received power prediction model.

2. The method for predicting received power according to claim 1, characterized in that, The geographic location information of the signal transmitter and receiver in the prediction scenario includes at least one of the following: The distance between the signal transmitter and the signal receiver; Antenna gain between the signal transmitter and the signal receiver; The relative height of the transmitting and receiving antennas between the signal transmitting end and the signal receiving end; The type of grid feature closest to the signal receiver on the line connecting the signal transmitter and the signal receiver; The signal transmitter and receiver belong to indoor or outdoor user information; Information on occlusion of nearby ground features; The indoor or outdoor distance between the signal transmitter and the signal receiver; The distance lengths of various land cover types along the line connecting the signal transmitter and receiver.

3. The method for predicting received power according to claim 1, characterized in that, The received power prediction model is as follows: ， Where R is the predicted received power. For distance feature values, This is the characteristic value of antenna gain. Here, the values ​​are: relative height feature value, near_vector, near ground feature type feature value, in_out_door, indoor or outdoor user feature value, block, occlusion feature value for near ground feature type, D0 and D1, indoor or outdoor distance feature values ​​between the signal transmitter and receiver, clutte_distance, distance length feature values ​​for various ground feature types along the distance line connecting the signal transmitter and receiver, d, distance between the signal transmitter and receiver, and G, gain corresponding to the horizontal and vertical angle readings of the antenna pattern. To read the gain corresponding to the antenna pattern at a horizontal angle. To read the gain corresponding to the antenna pattern at a vertical angle, H , Absolute altitude of the transmitting end: sea level at the signal transmitting end + terrain elevation at the signal transmitting end + building height at the signal transmitting end; Absolute height of the receiving end: sea level at the signal receiving end + terrain height at the signal receiving end + building height at the signal receiving end.

4. The method for predicting received power according to claim 3, characterized in that, The dataset of omnidirectional antennas is used for training. Parameters in; The dataset of directional antennas is used for training. Parameters in; in, For distance-related feature parameters, Here, is an eigenvalue parameter related to antenna gain, and d is the distance between the signal transmitter and the signal receiver. Let G be the varying distance between the signal transmitter and receiver, and G be the gain corresponding to the horizontal and vertical angles used to read the antenna pattern. To read the gain corresponding to the antenna pattern at a horizontal angle. The gain corresponding to the antenna pattern is read at the vertical angle.

5. The method for predicting received power according to claim 1, characterized in that, Correcting the predicted received power to obtain the target predicted received power includes: The correction amount for the predicted received power is obtained based on the actual transmit power of the signal transmitter in the predicted scenario and the transmit power of the signal transmitter in the training scenario. The predicted received power is superimposed with the correction amount to obtain the target predicted received power.

6. The method for predicting received power according to claim 5, characterized in that, Based on the actual transmit power of the signal transmitter in the predicted scenario and the transmit power of the signal transmitter in the training scenario, the correction amount for the predicted received power is obtained, including: The difference between the actual transmit power of the signal transmitter in the predicted scenario and the transmit power of the signal transmitter in the training scenario is used as the correction amount for the predicted received power.

7. A device for predicting received power, characterized in that, The device includes: The acquisition module is used to acquire information related to the actual geographical location of the signal transmitter and receiver in the predicted scenario. The processing module is used to convert the actual geographical location information into corresponding feature information, input it into the received power prediction model trained based on antenna gain features for processing, and obtain the predicted received power; then, the predicted received power is corrected to obtain the target predicted received power. The received power prediction model obtained by training based on antenna gain characteristics is trained through the following process: Acquire training set data for training the received power prediction model in the training scenario, wherein the training set data includes training data for omnidirectional antennas and training data for directional antennas; The gain features corresponding to the training data of the omnidirectional antenna are obtained, and the gain features corresponding to the training data of the omnidirectional antenna are input into the received power prediction model for training processing to obtain the first training model. Obtain the gain features corresponding to the training data of the directional antenna, and input the gain features corresponding to the training data of the directional antenna into the first training model for training processing to obtain the second training model; Obtain randomized sequences of the gain features corresponding to the training data of the omnidirectional antenna and the gain features corresponding to the training data of the directional antenna, and input the randomized sequences into the second training model to obtain the received power prediction model.

8. A computing device, comprising: The processor, memory, communication interface, and communication bus are provided, wherein the processor, memory, and communication interface communicate with each other via the communication bus. The memory is used to store at least one executable instruction that causes the processor to perform an operation corresponding to the method for predicting received power as described in any one of claims 1-6.

9. A computer storage medium storing at least one executable instruction that causes a processor to perform an operation corresponding to the method for predicting received power as described in any one of claims 1-6.

Images