WO2019109780A1 - 一种预测路损的方法及装置 - Google Patents

一种预测路损的方法及装置 Download PDF

Info

Publication number
WO2019109780A1
WO2019109780A1 PCT/CN2018/114881 CN2018114881W WO2019109780A1 WO 2019109780 A1 WO2019109780 A1 WO 2019109780A1 CN 2018114881 W CN2018114881 W CN 2018114881W WO 2019109780 A1 WO2019109780 A1 WO 2019109780A1
Authority
WO
WIPO (PCT)
Prior art keywords
path loss
transmitter
feature
learning model
parameter
Prior art date
Application number
PCT/CN2018/114881
Other languages
English (en)
French (fr)
Inventor
苏惠荞
何峰
徐志节
苗加成
李小龙
Original Assignee
华为技术有限公司
Priority date (The priority date 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 date listed.)
Filing date
Publication date
Application filed by 华为技术有限公司 filed Critical 华为技术有限公司
Priority to EP18886137.1A priority Critical patent/EP3687210A4/en
Publication of WO2019109780A1 publication Critical patent/WO2019109780A1/zh

Links

Images

Classifications

    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04W—WIRELESS COMMUNICATION NETWORKS
    • H04W24/00—Supervisory, monitoring or testing arrangements
    • H04W24/08—Testing, supervising or monitoring using real traffic
    • 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
    • H04W16/22—Traffic simulation tools or models
    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04W—WIRELESS COMMUNICATION NETWORKS
    • H04W24/00—Supervisory, monitoring or testing arrangements
    • H04W24/04—Arrangements for maintaining operational condition

Definitions

  • the embodiments of the present invention relate to the field of communications technologies, and in particular, to a method and an apparatus for predicting path loss.
  • the final decision of the wireless network planning and design depends largely on the effect of the simulation coverage prediction, and the effect of the coverage prediction is subject to the propagation model selected by each cell.
  • the accuracy and scope of the propagation model will affect the overall planning and design quality.
  • the embodiment of the present application provides a method and apparatus for predicting path loss, so as to solve the problem that the prediction efficiency caused by propagation correction is low when predicting propagation path loss.
  • the embodiment of the present application provides a method for predicting a path loss, including: acquiring sample data, where the sample data includes a parameter parameter that affects wireless propagation between a transmitter and a cell, a parameter value of a feature parameter, and a cell.
  • the measured path loss between each receiver and the transmitter is rasterized to generate feature images, and a deep learning model is constructed based on the feature image and the measured path loss.
  • the input variables of the deep learning model are feature images and deep learning.
  • the output variable of the model is used to represent the path loss between each receiver and transmitter in the cell; the constructed depth learning model is used to predict the path loss.
  • the sample learning data in the live network can be obtained, and a deep learning model is constructed according to the sample data, and a model mapping relationship between each characteristic parameter affecting wireless propagation between the transmitter and the receiver and the path loss is obtained, and the model is used.
  • the model mapping relationship predicts the path loss without correcting the existing propagation model and improving the efficiency of predicting path loss.
  • the feature parameter is rasterized to generate a feature image, including: creating a raster image centered on the transmitter, wherein the raster image includes at least a grid with different grids at different coordinate positions; traversing each grid in the raster image, using the grid as a pixel, and taking the parameter value of the feature parameter corresponding to the pixel as the value of the pixel; The traversed raster image is used as a feature image.
  • the feature image corresponding to the feature parameter can be generated by the rasterization method to meet the image input requirement of the deep learning.
  • the characteristic parameter includes: a building height, a feature type, an altitude, and a transmitter At least one of a feature of the distance, the horizontal azimuth of the transmitter, the vertical azimuth of the transmitter, the height of the transmitter from the ground, the frequency at which the transmitter transmits the signal, and the transmit power of the transmitter.
  • the actual geographic features and the transmitter's own characteristics can be used as the characteristic parameters affecting the wireless propagation between the transmitter and the receiver, without extracting complex features, and enhancing the adaptability of the deep learning model to the communication scenario.
  • the method further includes filtering the sample data; normalizing the parameter values of the characteristic parameters in the filtered sample data and the measured path loss; and performing rasterization on the normalized characteristic parameters
  • the feature processing generates a feature image and increases the number of feature images by model flipping or model rotation.
  • the sample data can be filtered and cleaned to improve the quality of the sample data, and at the same time, the number of feature images is increased, and the information input capability of the deep learning model can be improved.
  • the output variable of the deep learning model is a feature image corresponding to the path loss or a vector corresponding to the path loss.
  • the output variable of the deep learning model can be represented by the feature image or vector, and the output flexibility of the deep learning model is improved.
  • an apparatus for predicting a path loss including:
  • An acquiring unit configured to acquire sample data, where the sample data includes a characteristic parameter that affects wireless propagation between the transmitter and the cell, a parameter value of the characteristic parameter, and a measured path loss between the receiver and the transmitter in the cell, and the characteristic parameter
  • the parameter value corresponds to the coordinate position
  • the measured path loss corresponds to the coordinate position of the receiver
  • An image generating unit configured to perform rasterization processing on the feature parameter to generate a feature image, where the feature image includes at least one pixel point, and each pixel point corresponds to one parameter value;
  • a model building unit configured to construct a deep learning model according to the feature image and the measured path loss, wherein the input variable of the deep learning model is a feature image, and the output variable of the deep learning model is used to represent a path loss between the transmitter and the cell;
  • a prediction unit for predicting a path loss using the constructed deep learning model is a prediction unit for predicting a path loss using the constructed deep learning model.
  • the apparatus for predicting path loss For a specific implementation manner of the apparatus for predicting the path loss, reference may be made to the implementation manner of the method for predicting path loss provided by the foregoing first aspect or the possible implementation manner of the foregoing first aspect, and details are not described herein again. Therefore, the apparatus for predicting path loss provided by this aspect can achieve the same advantageous effects as the first aspect described above.
  • the embodiment of the present application provides a device for predicting a path loss, where the device for predicting a path loss can implement a function performed by a device for predicting a path loss in the foregoing first aspect, where the function may be implemented by hardware, or The corresponding software implementation is performed by hardware.
  • the hardware or software includes one or more modules corresponding to the above functions.
  • the apparatus for predicting path loss includes a processor and a communication interface configured to support the device for predicting path loss to perform a corresponding function in the above method.
  • the communication interface is used to support communication between the device for predicting path loss and other network elements.
  • the means for predicting path loss can also include a memory for coupling with the processor that holds program instructions and data necessary for the device that predicts the path loss.
  • an embodiment of the present application provides a computer storage medium for storing computer software instructions for use in a device for predicting a path loss, the computer software instructions comprising a program for performing the solution of the above aspects.
  • embodiments of the present application provide a computer program product storing computer software instructions for use in the apparatus for predicting path loss, the computer software instructions comprising a program for performing the solution of the above aspects.
  • the embodiment of the present application provides a device, which is in the form of a product of a chip.
  • the device includes a processor and a memory, and the memory is coupled to the processor, and saves necessary program instructions of the device.
  • the processor is operative to execute program instructions stored in the memory such that the apparatus performs the functions corresponding to the means for predicting path loss in the above method.
  • Figure 1 is a schematic block diagram of an embodiment of the present application
  • FIG. 2 is a schematic diagram of a device for predicting a path loss according to an embodiment of the present application
  • FIG. 3 is a flowchart of a method for predicting path loss according to an embodiment of the present application
  • Figure 3a is a schematic diagram of rasterization provided by an embodiment of the present application.
  • FIG. 3b is a schematic diagram of a mapping relationship between a feature parameter and a path loss according to an embodiment of the present application
  • FIG. 3 is a schematic diagram of another mapping relationship between feature parameters and path loss according to an embodiment of the present disclosure
  • FIG. 3 is a schematic diagram of a composition of a convolutional neural network according to an embodiment of the present application.
  • FIG. 4 is a schematic diagram of another device for predicting path loss according to an embodiment of the present application.
  • FIG. 5 is a schematic structural diagram of another apparatus for predicting path loss according to an embodiment of the present application.
  • the basic principle of the embodiment of the present application is as shown in FIG. 1 , which includes: extracting characteristic parameters affecting wireless propagation between a transmitter and a receiver between cells, and actually measuring a path loss between the transmitter and the receiver (referred to as a measured path loss).
  • the feature parameters are rasterized to generate feature images, and the deep learning model (such as convolutional neural network model) is constructed by using the feature image and the measured path loss.
  • the deep learning model is used to represent the wireless communication between the transmitter and the inter-cell receiver.
  • the relationship between the characteristic parameters and the path loss, and subsequently, when predicting the path loss between the transmitter and the receiver, the characteristic parameters affecting the wireless propagation between the transmitter and the receiver can be input into the deep learning model to predict the transmitter and receive.
  • the path loss between the machines does not need to correct the existing propagation model for path loss prediction, which improves the prediction efficiency.
  • the method for predicting path loss provided by the embodiment of the present application can be used to predict a path loss between a transmitter and a receiver in any of the following wireless communication systems: Universal Mobile Telecommunications System (UMTS), Long Term Evolution (long) The term evolution (LTE) system, the fifth generation mobile communication technology (5-Generation, 5G), etc.
  • the transmitter may be a base station (nodeB, NB), an evolved base station (evolution node B, eNB) ), an access node, a generation node (gNB), a transmission receive point (TRP), a transmission point (TP), or any other access device or an antenna integrated in the access device
  • the array; the receiver may be a receiving device such as a terminal device, or an antenna array integrated in the receiving device, or the like.
  • the method for predicting the path loss may be performed by the device for predicting the path loss shown in FIG. 2, or may be performed by any computer capable of executing the method for predicting the path loss provided by the embodiment of the present application,
  • FIG. 2 is a schematic diagram of a device for predicting a path loss according to an embodiment of the present disclosure.
  • the device may be separately deployed in a network planning system, or may be integrated into a server of the network planning system, and is not limited.
  • the apparatus for predicting path loss may include the predicted path loss apparatus 200 including at least one processor 201, a communication line 202, a memory 203, and at least one communication interface 204.
  • the processor 201 can be a central processing unit (CPU), an application specific integrated circuit (ASIC), or one or more integrated circuits configured to implement the embodiments of the present application.
  • CPU central processing unit
  • ASIC application specific integrated circuit
  • DSPs digital signal processors
  • FPGAs field programmable gate arrays
  • Communication line 202 can include a path for communicating information between the components described above.
  • the memory 203 can be a read-only memory (ROM) or other type of static storage device that can store static information and instructions, a random access memory (RAM) or other type that can store information and instructions.
  • the dynamic storage device can also be an electrically erasable programmable read-only memory (EEPROM), a compact disc read-only memory (CD-ROM) or other optical disc storage, and a disc storage device. (including compact discs, laser discs, optical discs, digital versatile discs, Blu-ray discs, etc.), magnetic disk storage media or other magnetic storage devices, or can be used to carry or store desired program code in the form of instructions or data structures and can be Any other media accessed, but not limited to this.
  • the memory may be stand-alone and connected to the processor via communication line 202. The memory can also be integrated with the processor.
  • the memory 203 is used to store computer execution instructions for executing the solution of the present application, and is controlled by the processor 201 for execution.
  • the processor 201 is configured to execute computer-executed instructions stored in the memory 203 to implement the method of predicting path loss provided by the embodiments of the present application.
  • the computer-executed instructions in the embodiment of the present application may also be referred to as an application code, which is not specifically limited in this embodiment of the present application.
  • the communication interface 204 uses devices such as any transceiver for communicating with other devices or communication networks, such as Ethernet, radio access network (RAN), wireless local area networks (WLAN), etc. .
  • devices such as any transceiver for communicating with other devices or communication networks, such as Ethernet, radio access network (RAN), wireless local area networks (WLAN), etc. .
  • RAN radio access network
  • WLAN wireless local area networks
  • processor 201 may include one or more CPUs, such as CPU0 and CPU1 in FIG.
  • the apparatus 200 for predicting path loss may include a plurality of processors, such as processor 201 and processor 207 in FIG. Each of these processors can be a single core processor or a multi-CPU processor.
  • a processor herein may refer to one or more devices, circuits, and/or processing cores for processing data, such as computer program instructions.
  • the apparatus 200 for predicting path loss may further include an output device 205 and an input device 206.
  • Output device 205 is in communication with processor 201 and can display information in a variety of ways.
  • the output device 205 can be a liquid crystal display (LCD), a light emitting diode (LED) display device, a cathode ray tube (CRT) display device, or a projector.
  • Input device 206 is in communication with processor 201 and can receive user input in a variety of ways.
  • input device 206 can be a mouse, keyboard, touch screen device or sensing device, and the like.
  • the above device 200 for predicting path loss may be a general purpose device or a dedicated device.
  • the device 200 for predicting path loss may be a desktop, a portable computer, a network server, a mobile handset, a tablet, a wireless terminal device, an embedded device, or a device having a similar structure in FIG.
  • the embodiment of the present application does not limit the type of device 200 for predicting path loss.
  • FIG. 3 is a flowchart of a method for predicting a path loss according to an embodiment of the present disclosure. The method is performed by the apparatus shown in FIG. 2. As shown in FIG. 3, the method may include:
  • Step 301 Acquire sample data.
  • the sample data corresponding to the cell, the sample data corresponding to different cells may be the same or different, and the sample data corresponding to each cell may include a parameter parameter that affects wireless transmission between the transmitter and the cell, a parameter value of the feature parameter, and each parameter in the cell.
  • the measured path loss between the receiver and the transmitter may also include other information and is not limited.
  • the parameter value of the feature parameter corresponds to the coordinate position, and the parameter value of the feature parameter corresponding to the different coordinate position may be different; the measured path loss corresponds to the coordinate position of the receiver, and the receiver and the transmitter at different coordinate positions
  • the measured path loss between the machines may be different.
  • the measured path loss may be carried in a measurement report (MR), so that the device for predicting the path loss obtains the measured path loss between the receiver and the transmitter from the MR, and the coordinate position may be Refers to location information expressed in longitude and latitude.
  • the above characteristic parameters affecting radio propagation between the transmitter and the cell may include geographic features of the cell covered by the transmitter and information of the work parameters of the transmitter.
  • the geographic feature may include at least one of a feature of a building height, a feature type, an altitude, and a distance between the transmitter, and the height of the building may refer to a linear distance from the highest point of the building to the ground.
  • the type can refer to the type of things deployed on the surface, such as different types of things such as buildings and communication equipment.
  • the altitude can refer to the height of the highest point of the building from the sea level; the distance from the transmitter can refer to the coordinate position. Straight line distance from the transmitter.
  • the parameter value of each geographic feature may be determined by querying a 3D (3 Dimensions, 3D) map in a geographic information system, where the feature type
  • the value may be a number or a letter or other value symbol, and is not limited. For example, it is assumed that 1 represents a building, and 2 represents a communication device. When a building is deployed at a coordinate position, the coordinate position corresponds to the coordinate position.
  • the value of the feature type is 1.
  • the center point of a certain area may be regarded as a coordinate position, and the height of the highest building in the area is taken as the height of the building corresponding to the coordinate position, and the area is The most similar thing deployed in the coordinate type is the feature type corresponding to the coordinate position, and the height of the highest building in the area is taken as the altitude corresponding to the coordinate position.
  • the work parameter information of the transmitter may include at least one of a horizontal azimuth angle of the transmitter, a vertical azimuth angle of the transmitter, a height of the transmitter from the ground, a frequency of the transmitter transmitting signal, and a transmit power of the transmitter;
  • the horizontal azimuth of the transmitter may refer to the angle between the main lobe direction of the transmitter and the line between the transmitter and the coordinate position.
  • the vertical azimuth angle of the transmitter may refer to the downtilt angle of the transmitter and the position between the transmitter and the coordinate position.
  • the angle of the connection, the height of the transmitter from the ground can refer to the linear distance of the highest point of the transmitter from the surface.
  • the frequency of the signal transmitted by the transmitter can refer to the frequency at which the transmitter sends a signal to a coordinate position, and the transmission power of the transmitter. It can refer to the power of the transmitter to send a signal to a coordinate position.
  • the work parameter information of the transmitter may be obtained from the configuration information of the transmitter.
  • path loss described in the embodiments of the present application may also be replaced by other technical terms for characterizing the severity of the propagation characteristics of the transmitter to the receiver on the propagation of electromagnetic waves, such as level or reference signal received power.
  • Reference Signal Receiving Power RSRP
  • no restrictions RSRP
  • a large amount of sample data may be acquired in step 301, that is, sample data corresponding to multiple cells covered by the transmitter is acquired.
  • Step 302 Rasterizing the feature parameters to generate a feature image, wherein the feature image includes at least one pixel, and each pixel corresponds to one parameter value.
  • the step of rasterizing the feature parameters to generate the feature image may include the following steps (1) to (3):
  • the raster image is a two-dimensional planar image, and the raster image corresponds to a coverage area centered on the transmitter, and the coverage area may be a regular area (such as a square area or a rectangular area), and the size of the coverage area is not Restrictions, such as: rasterization of a 50m*50m or 100m*100m coverage area centered at the transmitter to obtain a raster image.
  • the transmitter-centered coverage area is divided into multiple grids, and the divided image is used as a raster image, wherein each grid has the same size and the length of each grid.
  • the width may be the same, as may be 10m or 20m.
  • each grid in the raster image may be regarded as one pixel, and the coordinate position of the pixel may be the center position of the grid. .
  • the center coordinate of the grid may be used as the coordinate position of the pixel point, and the parameter value of the feature parameter corresponding to the coordinate position is assigned to the pixel point.
  • the process of the above (1) to (3) can be used to generate a feature image, for example, in the embodiment of the present application, the height of the building, the type of the object, the altitude, and the transmitter can be The distance, the horizontal azimuth of the transmitter, the vertical azimuth of the transmitter, the height of the transmitter from the ground, the frequency of the signal transmitted by the transmitter, and the transmit power of the transmitter, respectively, perform the above process to generate 9 features. image.
  • Step 303 Construct a deep learning model according to the feature image and the measured path loss, wherein the input variable of the deep learning model is a feature image, and the output variable of the deep learning model is used to represent a path loss between the transmitter and each receiver in the cell.
  • the deep learning model may be used to represent a functional relationship between the feature parameters and the path loss.
  • the output variable of the deep learning model may be a path loss between the transmitter and each receiver in the cell predicted by the deep learning model.
  • the output variable of the deep learning model may be a feature image corresponding to the path loss, and the feature image corresponding to the path loss corresponds to the path loss at the coordinate position, and the feature image corresponding to the path loss is corresponding to each feature parameter.
  • the feature image obtained by the deep learning of the deep learning model that is, the deep learning model realizes image to image mapping; or, as shown in FIG.
  • the output variable of the deep learning model is a vector corresponding to the path loss
  • the elements in the vector correspond to the path loss at the coordinate position
  • the vector is a vector obtained by deep learning of the feature image corresponding to each feature parameter through the deep learning model, that is, the depth learning model implements image-to-vector mapping.
  • the deep learning model is a convolutional neural network (CNN), and the process of constructing the deep learning model according to the feature image and the measured path loss includes:
  • the error between the path loss outputted by the convolutional neural network and the measured path loss is input back to the convolutional neural network, and the parameters of the convolutional neural network are adjusted;
  • each time the feature image input to the convolutional neural network is a feature image of the feature parameter in the sample data corresponding to the different cell.
  • the convolutional neural network may be a multi-layer neural network, each layer consisting of multiple two-dimensional planes, each plane consisting of multiple independent neurons, such as: the convolutional neural network may include an input layer, a volume Multiple layers such as stacking, pooling layer (such as maximum pooling layer, minimum pooling layer), Inception V3 layer, upsampling layer, and output layer.
  • the description of the convolutional neural network may refer to the prior art, and details are not described herein again.
  • the feature image is input to the convolutional neural network through the input layer, and passes through the convolution layer, the maximum pooling layer, the convolution layer, the maximum pooling layer, the Inception V3 layer, the Inception V3 layer, and the upsampling.
  • Layer, convolutional layer, upsampling layer, and convolutional layer are correlated to predict the path loss between the transmitter and the receiver.
  • the path loss is output through the output layer, the path loss is compared with the measured path loss. The error between the path losses is propagated back to the input layer of the convolutional neural network to adjust the parameters between the layers in the convolutional neural network.
  • Step 304 Predict the path loss by using the constructed deep learning model.
  • using the constructed deep learning model to predict the path loss may refer to: predicting the path loss between the transmitter and the receiver in other cells, or predicting the path loss between the other transmitters and the receiving in other cells, without limitation. .
  • a deep learning model is constructed using the characteristic parameters between the transmitter 1 and the cell 1 to the cell 1000, and the measured path loss of the receiver in the transmitter 1 and the cell 1 to the cell 1000, the deep learning model can predict the transmitter 1 and the path loss between the receivers in the cell 1001 can also predict the path loss between the transmitter 2 and the receivers in other cells.
  • the sample data in the live network is obtained, and the deep learning model is constructed according to the sample data, and various characteristic parameters and path loss factors affecting wireless propagation between the transmitter and the receiver are obtained.
  • the model mapping relationship between the models is used to predict the path loss. In this way, it is not necessary to correct the existing propagation model, and the efficiency of predicting the path loss is improved.
  • the method further includes:
  • the rasterizing the feature parameters to generate the feature image includes:
  • the normalized processed feature parameters are rasterized to generate a feature image, and the number of feature images is increased by model flipping or model rotation, thereby increasing the number of feature images input to the deep learning model, wherein the model is flipped It may be that the pixel points in the feature image are flipped up and down or left and right symmetrically, and the model rotation refers to the rotation of the pixel points in the feature image at a certain angle (for example, an angle of 60 degrees or 80 degrees).
  • the sample data can be filtered by off-group discrimination (such as box plot method), raster information accuracy discrimination, and cell information amount discrimination to remove noise effects.
  • the outlier discrimination can refer to: for a plurality of measured path losses on a certain location area (such as a grid), remove the measured path loss that is too large or too small, and remove any measured path loss or remove it.
  • the average value of the measured path loss is taken as the measured path loss corresponding to the location area; the accuracy of the raster information can be determined: if the value of the measured path loss corresponding to a certain location area is less than a preset threshold, the location area is The setting is an invalid area, where the preset threshold can be set as needed, and is not limited; the cell information quantity discrimination can be: when the number of invalid areas corresponding to a certain cell is greater than or equal to a preset number, the corresponding cell is deleted.
  • the sample data wherein the preset number can be set as needed, and is not limited.
  • Normalizing the parameter values of the characteristic parameters and the measured path loss in the filtered sample data may include: dimensioning the parameter values of the characteristic parameters and the measured value of the measured path loss to eliminate the difference between the different features The influence of dimensions.
  • the max-min method is selected to normalize the parameter values of the characteristic parameters in the filtered sample data and the measured path loss.
  • the device for predicting the path loss includes hardware structures and/or software modules corresponding to the execution of the respective functions in order to implement the above functions.
  • the present application can be implemented in a combination of hardware or hardware and computer software in combination with the algorithmic steps of the various examples described in the embodiments disclosed herein. Whether a function is implemented in hardware or computer software to drive hardware depends on the specific application and design constraints of the solution. A person skilled in the art can use different methods to implement the described functions for each particular application, but such implementation should not be considered to be beyond the scope of the present application.
  • the embodiment of the present application may perform the division of the function module on the device for predicting the path loss according to the foregoing method.
  • each function module may be divided according to each function, or two or more functions may be integrated into one processing module.
  • the above integrated modules can be implemented in the form of hardware or in the form of software functional modules. It should be noted that the division of the module in the embodiment of the present application is schematic, and is only a logical function division, and the actual implementation may have another division manner.
  • FIG. 4 shows still another possible composition diagram of the device for predicting the path loss, and the device for predicting the path loss can be used to execute the solution involved in the above method embodiment.
  • the apparatus for predicting path loss may include an acquisition unit 40, an image generation unit 41, a model construction unit 42, a prediction unit 43, and a filter cleaning unit 44.
  • the acquiring unit 40 is configured to acquire sample data, where the sample data includes a parameter parameter that affects wireless propagation between the transmitter and the cell, a parameter value of the feature parameter, and a measured path loss between the receiver and the transmitter in the cell.
  • the parameter value of the feature parameter corresponds to the coordinate position
  • the measured path loss corresponds to the coordinate position of the receiver; for example, the device for obtaining the predicted path loss by the obtaining unit 40 performs step 301.
  • the image generating unit 41 is configured to perform rasterization processing on the feature parameters to generate a feature image, wherein the feature image includes at least one pixel point, and each pixel point corresponds to one parameter value; for example, the image generating unit 41 is configured to support the predicted path loss.
  • the device performs step 302.
  • the model construction unit 42 is configured to construct a deep learning model according to the feature image and the measured path loss, wherein the input variable of the deep learning model is a feature image, and the output variable of the deep learning model is used to represent a path loss between the transmitter and the cell;
  • the means for supporting the prediction of the path loss by the model construction unit 42 performs step 303.
  • the prediction unit 43 is configured to predict the path loss by using the constructed deep learning model; for example, the apparatus for predicting the path loss by the prediction unit 43 performs step 304.
  • the apparatus may further include: a filter cleaning unit 44, configured to filter the sample data before the model construction unit 42 constructs the deep learning model according to the feature image and the measured path loss, and after filtering The parameter values of the characteristic parameters in the sample data and the measured path loss are normalized;
  • the image generating unit 41 is specifically configured to perform rasterization processing on the normalized feature parameters to generate a feature image, and increase the number of feature images by model inversion or model rotation.
  • the apparatus for predicting path loss provided by the embodiment of the present application is for performing the above method for predicting path loss, and thus the same effect as the method for predicting path loss described above can be achieved.
  • FIG. 5 shows a device in the form of a product of a chip for performing the function of the device for predicting path loss in the above embodiment, as shown in FIG. It may include: a processing module 50 and a communication module 51.
  • the processing module 50 is for controlling management of the actions of the device.
  • the processing module 50 is configured to support the device to perform steps 301-304 and/or other processes for the techniques described herein.
  • the communication module 51 is used to support communication of the device with other network entities.
  • the apparatus can also include a storage module 52 for storing program code and data for the device.
  • the processing module 50 can be a processor or a controller. It is possible to implement or carry out the various illustrative logical blocks, modules and circuits described in connection with the present disclosure.
  • the processor can also be a combination of computing functions, for example, including one or more microprocessor combinations, a combination of a DSP and a microprocessor, and the like.
  • the communication module 51 can be a communication interface, a transceiver circuit, a communication interface, or the like.
  • the storage module 52 can be a memory.
  • the processing module 50 is a processor
  • the communication module 51 is a communication interface
  • the storage module 52 is a memory
  • the device involved in the embodiment of the present application may be the device shown in FIG.
  • the disclosed apparatus and method may be implemented in other manners.
  • the device embodiments described above are merely illustrative.
  • the division of the modules or units is only a logical function division.
  • there may be another division manner for example, multiple units or components may be used.
  • the combination may be integrated into another device, or some features may be ignored or not performed.
  • the mutual coupling or direct coupling or communication connection shown or discussed may be an indirect coupling or communication connection through some interface, device or unit, and may be in an electrical, mechanical or other form.
  • the units described as separate components may or may not be physically separated, and the components displayed as units may be one physical unit or multiple physical units, that is, may be located in one place, or may be distributed to multiple different places. . Some or all of the units may be selected according to actual needs to achieve the purpose of the solution of the embodiment.
  • each functional unit in each embodiment of the present application may be integrated into one processing unit, or each unit may exist physically separately, or two or more units may be integrated into one unit.
  • the above integrated unit can be implemented in the form of hardware or in the form of a software functional unit.
  • the integrated unit if implemented in the form of a software functional unit and sold or used as a standalone product, may be stored in a readable storage medium.
  • the technical solution of the embodiments of the present application may be embodied in the form of a software product in the form of a software product in essence or in the form of a contribution to the prior art, and the software product is stored in a storage medium.
  • a number of instructions are included to cause a device (which may be a microcontroller, chip, etc.) or processor to perform all or part of the steps of the methods described in various embodiments of the present application.
  • the foregoing storage medium includes various media that can store program codes, such as a USB flash drive, a mobile hard disk, a ROM, a RAM, a magnetic disk, or an optical disk.

Landscapes

  • Engineering & Computer Science (AREA)
  • Computer Networks & Wireless Communication (AREA)
  • Signal Processing (AREA)
  • Management, Administration, Business Operations System, And Electronic Commerce (AREA)
  • Image Analysis (AREA)

Abstract

本申请实施例公开了一种预测路损的方法及装置，以解决现有在预测传播路损时，对传播校正导致的预测效率低下的问题。该方法包括：获取样本数据，其中，样本数据包括影响发射机与小区间无线传播的特征参数、特征参数的参数值、以及小区中各接收机与发射机间的实测路损；对特征参数进行栅格化处理生成特征图像；根据特征图像和实测路损构建深度学习模型，其中，深度学习模型的输入变量为特征图像，深度学习模型的输出变量用于表示小区中各接收机与发射机间的路损；利用构建后的深度学习模型预测路损。本申请实施例提供的方法用于预测发射机与接收机间的路损。

Description

一种预测路损的方法及装置
本申请要求于2017年12月05日提交中国专利局、申请号为201711271150.6、申请名称为“一种预测路损的方法及装置”的中国专利申请的优先权，其全部内容通过引用结合在本申请中。
技术领域
本申请实施例涉及通信技术领域，尤其涉及一种预测路损的方法及装置。
背景技术
在现有的网络规划技术中，无线网络规划设计最终确定的方案很大程度上取决于仿真覆盖预测的效果，而覆盖预测的效果又受制于各个小区选择的传播模型(propagation model)，选择的传播模型的精确性以及适用范围，会影响到整体的规划设计质量。
当前，业界通常选择下述三类传播模型进行覆盖预测：确定型传播模型、经验传播模型、基于大数据的分类拟合模型。但是，由于规划场景的多样性、复杂性，这三类传播模型都需要根据实际场景数据进行传播模型校正才能够预测路损，进而根据预测路损预测出网络覆盖情况，预测效率低下。
发明内容
本申请实施例提供一种预测路损的方法及装置，以解决现有在预测传播路损时，对传播校正导致的预测效率低下的问题。
为达到上述目的，本申请实施例采用如下技术方案：
第一方面，本申请实施例提供了一种预测路损的方法，包括：获取样本数据，其中，样本数据包括影响发射机与小区间无线传播的特征参数、特征参数的参数值、以及小区中各接收机与发射机间的实测路损，对特征参数进行栅格化处理生成特征图像，根据特征图像和实测路损构建深度学习模型，其中，深度学习模型的输入变量为特征图像，深度学习模型的输出变量用于表示小区中各接收机与发射机间的路损；利用构建后的深度学习模型预测路损。
基于上述方案，可以通过获取现网中的样本数据，根据该样本数据构建深度学习模型，得到影响发射机与接收机间无线传播的各个特征参数与路损之间的模型映射关系，并利用该模型映射关系预测路损，无需对现有传播模型进行校正，提高了预测路损的效率。
结合第一方面，在第一方面的第一种可能的实现方式中，对特征参数进行栅格化处理生成特征图像，包括：以发射机为中心建立栅格图像，其中，栅格图像包括至少一个栅格，不同栅格位于不同的坐标位置；遍历栅格图像中的每个栅格，将栅格作为一个像素点，将像素点对应的特征参数的参数值作为像素点的取值；将遍历后的栅格图像作为特征图像。
如此，可以通过栅格化方法生成与特征参数对应的特征图像，满足深度学习的图像输入要求。
结合第一方面或者第一方面的第一种可能的实现方式，在第一方面的第二种可能的实现方式中，特征参数包括：建筑物高度、地物类型、海拔高度、与发射机间的距离、发射机的水平方位角、发射机的垂直方位角、发射机距地面的高度、发射机发送信号的频率以及发射机的发射功率等特征中的至少一种特征。
如此，可以将实际的地理特征和发射机自身特征作为影响发射机与接收机间无线传播的特征参数，无需提取复杂特征，增强了深度学习模型对通信场景的适应性。
结合第一方面或者第一方面的任一可能的实现方式，在第一方面的第三种可能的实现方式中，为了提高深度学习模型预测出的路损的精度，在根据特征图像和实测路损构建深度学习模型之前，还包括对样本数据进行过滤处理；对过滤后的样本数据中的特征参数的参数值和实测路损进行归一化处理；对归一化处理后的特征参数进行栅格化处理生成特征图像，并通过模型翻转或者模型旋转增加特征图像的个数。
如此，可以对样本数据进行过滤清洗，提高样本数据的质量，同时，增加特征图像的个数，提高可深度学习模型的信息输入能力。
结合第一方面或者第一方面的任一可能的实现方式，在第一方面的第四种可能的实现方式中，深度学习模型的输出变量为路损对应的特征图像或者路损对应的向量。
如此，可以通过特征图像或者向量来表示深度学习模型的输出变量，提高了深度学习模型的输出灵活性。
第二方面，本申请实施例提供了一种预测路损的装置，包括：
获取单元，用于获取样本数据，其中，样本数据包括影响发射机与小区间无线传播的特征参数、特征参数的参数值、以及小区中的接收机与发射机间的实测路损，特征参数的参数值与坐标位置对应，实测路损与接收机的坐标位置对应；
图像生成单元，用于对特征参数进行栅格化处理生成特征图像，其中，特征图像包括至少一个像素点，每个像素点对应一个参数值；
模型构建单元，用于根据特征图像和实测路损构建深度学习模型，其中，深度学习模型的输入变量为特征图像，深度学习模型的输出变量用于表示发射机与小区间的路损；
预测单元，用于利用构建后的深度学习模型预测路损。
其中，预测路损的装置的具体实现方式可以参考上述第一方面或上述第一方面的可能的实现方式提供的预测路损的方法中的实现方式，在此不再赘述。因此，该方面提供的预测路损的装置可以达到与上述第一方面相同的有益效果。
一方面，本申请实施例提供了一种预测路损的装置，该预测路损的装置可以实现上述第一方面中预测路损的装置所执行的功能，所述功能可以通过硬件实现，也可以通过硬件执行相应的软件实现。所述硬件或软件包括一个或多个上述功能相应的模块。
在一种可能的设计中，该预测路损的装置的结构中包括处理器和通信接口，该处理器被配置为支持该预测路损的装置执行上述方法中相应的功能。该通信接口用于支持该预测路损的装置与其他网元之间的通信。该预测路损的装置还可以包括存储器，该存储器用于与处理器耦合，其保存该预测路损的装置必要的程序指令和数据。
一方面，本申请实施例提供了一种计算机存储介质，用于储存为上述预测路损的装置所用的计算机软件指令，该计算机软件指令包含用于执行上述方面所述方案的程 序。
一方面，本申请实施例提供了一种计算机程序产品，该程序产品储存有上述预测路损的装置所用的计算机软件指令，该计算机软件指令包含用于执行上述方面所述方案的程序。
一方面，本申请实施例提供了一种装置，该装置以芯片的产品形态存在，该装置的结构中包括处理器和存储器，该存储器用于与处理器耦合，保存该装置必要的程序指令和数据，该处理器用于执行存储器中存储的程序指令，使得该装置执行上述方法中与预测路损的装置相应的功能。
附图说明
图1为本申请实施例提供的原理框图；
图2为本申请实施例提供的一种预测路损的装置组成示意图；
图3为本申请实施例提供的一种预测路损的方法流程图；
图3a为本申请实施例提供的栅格化示意图；
图3b为本申请实施例提供的一种特征参数与路损间的映射关系示意图；
图3c为本申请实施例提供的又一种特征参数与路损间的映射关系示意图；
图3d为本申请实施例提供的一种卷积神经网络的组成示意图；
图4为本申请实施例提供的又一种预测路损的装置组成示意图；
图5为本申请实施例提供的再一种预测路损的装置组成示意图。
具体实施方式
本申请实施例的基本原理如图1所示，包括：提取影响发射机与小区间接收机间无线传播的特征参数、以及发射机与接收机间实际测量的路损(简称实测路损)，对特征参数进行栅格化处理生成特征图像，利用特征图像和实测路损构建深度学习模型(如卷积神经网络模型)，该深度学习模型用于表示影响发射机与小区间接收机间无线传播的特征参数与路损之间的关系，后续，当预测发射机与接收机间的路损时，可以将影响发射机与接收机间无线传播的特征参数输入该深度学习模型预测发射机与接收机间的路损，无需通过校正现有传播模型进行路损预测，提高了预测效率。
下面结合附图对本申请实施例的实施方式进行详细描述。
本申请实施例提供的预测路损的方法可以用于预测下述任一无线通信系统中发射机与接收机间的路损：通用移动通信系统(Universal Mobile Telecommunications System，UMTS)、长期演进(long term evolution，LTE)系统、第五代移动通信技术(5-Generation，5G)等，其中，在本申请实施例中，发射机可以为基站(nodeB，NB)、演进型基站(evolution nodeB，eNB)、接入节点、下一代基站(generation nodeB，gNB)、收发点(transmission receive point，TRP)、传输点(transmission point，TP)等任一接入设备或某集成在接入设备中的天线阵列；接收机可以为终端设备等接收设备、或者集成在接收设备中的天线阵列等。该预测路损的方法可以由图2所示的预测路损的装置执行，也可以由任一能够执行本申请实施例提供的预测路损的方法的计算机执行，不予限制。
图2为本申请实施例提供的一种预测路损的装置的组成示意图，该装置可以单独部署在网络规划系统中，也可以集成在网络规划系统的某个服务器中，不予限制。如 图2所示，该预测路损的装置可以包括该预测路损的装置200包括至少一个处理器201，通信线路202，存储器203以及至少一个通信接口204。
处理器201可以是一个中央处理器(central processing unit，CPU)，也可以是特定集成电路(application specific integrated circuit，ASIC)，或者是被配置成实施本申请实施例的一个或多个集成电路，例如：一个或多个微处理器(digital signal processor，DSP)，或，一个或者多个现场可编程门阵列(field programmable gate array，FPGA)。
通信线路202可包括一通路，在上述组件之间传送信息。
存储器203可以是只读存储器(read-only memory，ROM)或可存储静态信息和指令的其他类型的静态存储设备，随机存取存储器(random access memory，RAM)或者可存储信息和指令的其他类型的动态存储设备，也可以是电可擦可编程只读存储器(electrically erasable programmable read-only memory，EEPROM)、只读光盘(compact disc read-only memory，CD-ROM)或其他光盘存储、光碟存储(包括压缩光碟、激光碟、光碟、数字通用光碟、蓝光光碟等)、磁盘存储介质或者其他磁存储设备、或者能够用于携带或存储具有指令或数据结构形式的期望的程序代码并能够由计算机存取的任何其他介质，但不限于此。存储器可以是独立存在，通过通信线路202与处理器相连接。存储器也可以和处理器集成在一起。
其中，存储器203用于存储执行本申请方案的计算机执行指令，并由处理器201来控制执行。处理器201用于执行存储器203中存储的计算机执行指令，从而实现本申请下述实施例提供的预测路损的方法。可选的，本申请实施例中的计算机执行指令也可以称之为应用程序代码，本申请实施例对此不作具体限定。
通信接口204，使用任何收发器一类的装置，用于与其他设备或通信网络通信，如以太网，无线接入网(radio access network，RAN)，无线局域网(wireless local area networks，WLAN)等。
在具体实现中，作为一种实施例，处理器201可以包括一个或多个CPU，例如图2中的CPU0和CPU1。
在具体实现中，作为一种实施例，预测路损的装置200可以包括多个处理器，例如图2中的处理器201和处理器207。这些处理器中的每一个可以是一个单核处理器，也可以是一个multi-CPU处理器。这里的处理器可以指一个或多个设备、电路、和/或用于处理数据(例如计算机程序指令)的处理核。
在具体实现中，作为一种实施例，预测路损的装置200还可以包括输出设备205和输入设备206。输出设备205和处理器201通信，可以以多种方式来显示信息。例如，输出设备205可以是液晶显示器(liquid crystal display，LCD)，发光二级管(light emitting diode，LED)显示设备，阴极射线管(cathode ray tube，CRT)显示设备，或投影仪(projector)等。输入设备206和处理器201通信，可以以多种方式接收用户的输入。例如，输入设备206可以是鼠标、键盘、触摸屏设备或传感设备等。
上述的预测路损的装置200可以是一个通用设备或者是一个专用设备。在具体实现中，预测路损的装置200可以是台式机、便携式电脑、网络服务器、移动手机、平板电脑、无线终端设备、嵌入式设备或有图2中类似结构的设备。本申请实施例不限定预测路损的装置200的类型。
下面结合图2所示装置，对本申请实施例提供的预测路损的方法进行详细描述。图3为本申请实施例提供的一种预测路损的方法流程图，该方法由图2所示装置执行，如图3所示，该方法可以包括：
步骤301：获取样本数据。
其中，上述样本数据与小区对应，不同小区对应的样本数据可以相同或者不同，每个小区对应的样本数据可以包括影响发射机与小区间无线传播的特征参数、特征参数的参数值、小区中各接收机与发射机间的实测路损，还可以包括其他信息，不予限制。在本申请实施例中，特征参数的参数值与坐标位置对应，不同坐标位置对应的特征参数的参数值可以不同；实测路损与接收机的坐标位置对应，不同坐标位置上的接收机与发射机间的实测路损可以不同，该实测路损可以携带在测量报告(measurement report，MR)中，以便预测路损的装置从MR中获取接收机与发射机间的实测路损，坐标位置可以指用经度和纬度表示的位置信息。
上述影响发射机与小区间无线传播的特征参数可以包括发射机所覆盖的小区的地理特征和发射机的工参信息。其中，地理特征可以包括建筑物高度、地物类型、海拔高度、与发射机间的距离等特征中的至少一种特征，建筑物高度可以指建筑物的最高点距离地面的直线距离，地物类型可以指地表上部署的事物的类型，如可以为建筑物、通信设备等不同类型的事物，海拔高度可以指建筑物的最高点距离海平面的高度；与发射机间的距离可以指坐标位置与发射机间的直线距离。在本申请实施例中，可以通过查询地理信息系统(geographic information system)中的三维(3 Dimensions，3D)地图确定每个地理特征的参数值(即地理特征的取值)，其中，地物类型的取值可以为数字或者字母或者其他取值符号，不予限制，如：假设用1代表建筑物，用2代表通信设备，当某坐标位置上部署有建筑物时，与该坐标位置对应的地物类型的取值为1。需要说明的是，在本申请实施例中，可以将某个区域的中心点看做为一个坐标位置，将该区域中最高的建筑物的高度作为该坐标位置对应的建筑物高度，将该区域中部署的个数最多的同类事物作为该坐标位置对应的地物类型，将该区域中最高的建筑物的高度作为该坐标位置对应的海拔高度。
其中，发射机的工参信息可以包括发射机的水平方位角、发射机的垂直方位角、发射机距地面的高度、发射机发送信号的频率以及发射机的发射功率中的至少一种特征；发射机的水平方位角可以指发射机的主瓣方向与该发射机与坐标位置间的连线的夹角，发射机的垂直方位角可以指发射机的下倾角与该发射机与坐标位置间的连线的夹角，发射机距离地面的高度可以指发射机的最高点距离地表的直线距离，发射机发送信号的频率可以指发射机向某坐标位置发送信号的频率，发射机的发射功率可以指发射机向某坐标位置发送信号的功率。在本申请实施例中，可以从发射机的配置信息中获取发射机的工参信息。
需要说明的是，本申请实施例所述的路损还可以替换为其他用于表征发射机到接收机的传播特征对电磁波的传播影响严重程度的技术术语，如：电平或者参考信号接收功率Reference Signal Receiving Power，RSRP)，不予限制。此外，为了提高后续深度学习模型构建的准确性，步骤301中可以获取大量的样本数据，即获取发射机覆盖的多个小区对应的样本数据。
步骤302：对特征参数进行栅格化处理生成特征图像，其中，特征图像包括至少一个像素点，每个像素点对应一个参数值。
其中，对特征参数进行栅格化处理生成特征图像可以包括下述(1)～(3)所示步骤：
(1)以发射机为中心建立栅格图像。
其中，栅格图像为二维平面图像，该栅格图像对应以发射机为中心的覆盖区域，该覆盖区域可以为一规则区域(如正方形区域或者矩形区域)，该覆盖区域的大小尺寸不予限制，如：可以对发射机为中心的50m*50m或者100m*100m的覆盖区域进行栅格化处理得到栅格图像。例如，如图3a所示，将以发射机为中心的覆盖区域划分为多个栅格，将划分后的图像作为栅格图像，其中，每个栅格的大小相同，每个栅格的长度和宽度可以相同，如可以为10m或者20m，在本申请实施例中，可以将栅格图像中的每个栅格看做为一个像素点，该像素点的坐标位置可以为栅格的中心位置。
(2)遍历栅格图像中的每个栅格，将栅格作为一个像素点，将像素点对应的特征参数的参数值作为像素点的取值。
其中，可以将栅格的中心坐标作为像素点的坐标位置，将该坐标位置对应的特征参数的参数值赋予该像素点。
(3)将遍历后的栅格图像作为特征图像。
需要说明的是，当某个像素点不存在与其对应的特征参数的参数值时，可以将该像素点赋予0。此外，对于每类特征参数均可以采用上述(1)～(3)的过程生成特征图像，如：在本申请实施例中，可以将建筑物高度、地物类型、海拔高度、与发射机间的距离、发射机的水平方位角、发射机的垂直方位角、发射机距地面的高度、发射机发送信号的频率以及发射机的发射功率这9个特征参数，分别执行上述过程生成9个特征图像。
步骤303：根据特征图像和实测路损构建深度学习模型，其中，深度学习模型的输入变量为特征图像，深度学习模型的输出变量用于表示发射机与小区中各个接收机间的路损。
其中，上述深度学习模型可以用于表示特征参数与路损间的函数关系，上述深度学习模型的输出变量可以为深度学习模型预测出来的发射机与小区中各个接收机间的路损。如图3b所示，深度学习模型的输出变量可以为路损对应的特征图像，该路损对应的特征图像与坐标位置上的路损对应，该路损对应的特征图像是各个特征参数对应的特征图像经过深度学习模型的深度学习后得到的特征图像，即该深度学习模型实现了图像到图像的映射；或者，如图3c所示，深度学习模型的输出变量为路损对应的向量，该向量中的元素与坐标位置上的路损相对应，该向量是各个特征参数对应的特征图像经过深度学习模型的深度学习后得到的向量，即该深度学习模型实现了图像到向量的映射。
可选的，上述深度学习模型为卷积神经网络(convolutional neural networks，CNN)，根据特征图像和实测路损构建深度学习模型的过程包括：
将特征参数对应的特征图像输入到卷积神经网络，输出发射机与接收机间的路损；
将卷积神经网络输出的路损与实测路损间的误差返向输入到卷积神经网络，调整 卷积神经网络的参数；
重复上述过程，直至卷积神经网络输出的路损与实测路损间的误差收敛，至此，卷积神经网络构建完成。需要说明的是，在构建深度学习模型的过程中，每次输入到卷积神经网络的特征图像为不同小区对应的样本数据中的特征参数的特征图像。
其中，上述卷积神经网络可以为一个多层的神经网络，每层由多个二维平面组成，每个平面由多个独立神经元组成，如：该卷积神经网络可以包括输入层、卷积层、池化层(如最大池化层、最小池化层)、Inception V3层、上采样层、输出层等多个层。具体的，卷积神经网络的描述可参照现有技术，在此不再赘述。
例如，如图3d所示，特征图像通过输入层输入到卷积神经网络之后，经过卷积层、最大池化层、卷积层、最大池化层、Inception V3层、Inception V3层、上采样层、卷积层、上采样层、卷积层的相关处理，预测出发射机与接收机间的路损，该路损经过输出层输出后，与实测路损进行比较，该路损与实测路损间的误差通过反向传播至卷积神经网络的输入层，用于调整卷积神经网络中各层间的参数。
步骤304：利用构建后的深度学习模型预测路损。
其中，利用构建后的深度学习模型预测路损可以指：预测上述发射机与其他小区中的接收机间的路损，或者预测其他发射机与其他小区中的接收间的路损，不予限制。
例如，假设利用发射机1与小区1～小区1000间的特征参数、以及发射机1与小区1～小区1000中的接收机的实测路损构建出深度学习模型，该深度学习模型可以预测发射机1与小区1001中接收机间的路损，还可以预测发射机2与其他小区中的接收机间的路损。
与现有技术相比，在图3所示方案中，获取现网中的样本数据，根据该样本数据构建深度学习模型，得到影响发射机与接收机间无线传播的各个特征参数与路损之间的模型映射关系，利用该模型映射关系预测路损。如此，无需对现有传播模型进行校正，提高了预测路损的效率。
进一步的，在图3所示方案中，为了提高预测出的路损的精度，在根据特征图像和实测路损构建深度学习模型之前，所述方法还包括：
对样本数据进行过滤处理；
对过滤后的样本数据中的特征参数的参数值和实测路损进行归一化处理；
所述对特征参数进行栅格化处理生成特征图像，包括：
对归一化处理后的特征参数进行栅格化处理生成特征图像，并通过模型翻转或者模型旋转增加特征图像的个数，进而提高输入到深度学习模型的特征图像的个数，其中，模型翻转可以指将特征图像中的像素点进行上下或者左右对称翻转，模型旋转指将特征图像中的像素点进行一定角度(如60度或者80度等角度)的旋转。
其中，可以通过离群点判别(如箱型图法)、栅格信息准确性判别、小区信息量判别等方法对样本数据进行过滤处理，以清除噪声影响。离群点判别可以指：对于某个位置区域(如一个栅格)上的多个实测路损，去除取值偏大或者偏小的实测路损，将去除后的任一实测路损或者去除后的实测路损的平均值作为该位置区域对应的实测路损；栅格信息准确性判别可以指：若某个位置区域对应的实测路损的取值小于预设阈值，则将该位置区域设置为无效区域，其中，预设阈值可以根据需要进行设置，不 予限制；小区信息量判别可以指：当某个小区对应的无效区域的个数大于等于预设个数时，删除该小区对应的样本数据，其中，预设个数可以根据需要进行设置，不予限制。
对过滤后的样本数据中的特征参数的参数值和实测路损进行归一化处理可以包括：对特征参数的参数值和实测路损的取值单位进行量纲处理，以消除不同特征之间量纲的影响。可选的，选择max-min方法对对过滤后的样本数据中的特征参数的参数值和实测路损进行归一化处理。
上述主要从各个节点之间交互的角度对本申请实施例提供的方案进行了介绍。可以理解的是，预测路损的装置为了实现上述功能，其包含了执行各个功能相应的硬件结构和/或软件模块。本领域技术人员应该很容易意识到，结合本文中所公开的实施例描述的各示例的算法步骤，本申请能够以硬件或硬件和计算机软件的结合形式来实现。某个功能究竟以硬件还是计算机软件驱动硬件的方式来执行，取决于技术方案的特定应用和设计约束条件。专业技术人员可以对每个特定的应用来使用不同方法来实现所描述的功能，但是这种实现不应认为超出本申请的范围。
本申请实施例可以根据上述方法示例对预测路损的装置进行功能模块的划分，例如，可以对应各个功能划分各个功能模块，也可以将两个或两个以上的功能集成在一个处理模块中。上述集成的模块既可以采用硬件的形式实现，也可以采用软件功能模块的形式实现。需要说明的是，本申请实施例中对模块的划分是示意性的，仅仅为一种逻辑功能划分，实际实现时可以有另外的划分方式。
在采用对应各个功能划分各个功能模块的情况下，图4示出了预测路损的装置的又一种可能的组成示意图，该预测路损的装置可以用于执行上述方法实施例涉及的方案。如图4所示，该预测路损的装置可以包括：获取单元40、图像生成单元41、模型构建单元42、预测单元43，过滤清洗单元44。
其中，获取单元40，用于获取样本数据，其中，样本数据包括影响发射机与小区间无线传播的特征参数、特征参数的参数值、以及小区中的接收机与发射机间的实测路损，特征参数的参数值与坐标位置对应，实测路损与接收机的坐标位置对应；如：获取单元40用于支持预测路损的装置执行步骤301。
图像生成单元41，用于对特征参数进行栅格化处理生成特征图像，其中，特征图像包括至少一个像素点，每个像素点对应一个参数值；如：图像生成单元41用于支持预测路损的装置执行步骤302。
模型构建单元42，用于根据特征图像和实测路损构建深度学习模型，其中，深度学习模型的输入变量为特征图像，深度学习模型的输出变量用于表示发射机与小区间的路损；如：模型构建单元42用于支持预测路损的装置执行步骤303。
预测单元43，用于利用构建后的深度学习模型预测路损；如：预测单元43用于支持预测路损的装置执行步骤304。
进一步的，如图4所示，该装置还可以包括：过滤清洗单元44，用于在模型构建单元42根据特征图像和实测路损构建深度学习模型之前，对样本数据进行过滤处理，对过滤后的样本数据中的特征参数的参数值和实测路损进行归一化处理；
图像生成单元41，具体用于对归一化处理后的特征参数进行栅格化处理生成特征 图像，并通过模型翻转或者模型旋转增加特征图像的个数。
需要说明的是，上述方法实施例涉及的各步骤的所有相关内容均可以援引到对应功能模块的功能描述，在此不再赘述。本申请实施例提供的预测路损的装置，用于执行上述预测路损的方法，因此可以达到与上述预测路损的方法相同的效果。
在采用集成的单元的情况下，图5示出了一种装置，该装置以芯片的产品形态存在，用于执行上述实施例中预测路损的装置的功能，如图5所示，该装置可以包括：处理模块50和通信模块51。
处理模块50用于对装置的动作进行控制管理，例如，处理模块50用于支持该装置执行步骤301～步骤304和/或用于本文所描述的技术的其它过程。通信模块51用于支持装置与其他网络实体的通信。装置还可以包括存储模块52，用于存储装置的程序代码和数据。
其中，处理模块50可以是处理器或控制器。其可以实现或执行结合本申请公开内容所描述的各种示例性的逻辑方框，模块和电路。处理器也可以是实现计算功能的组合，例如包含一个或多个微处理器组合，DSP和微处理器的组合等等。通信模块51可以是通信接口、收发电路或通信接口等。存储模块52可以是存储器。
当处理模块50为处理器，通信模块51为通信接口，存储模块52为存储器时，本申请实施例所涉及的装置可以为图2所示的装置。
通过以上的实施方式的描述，所属领域的技术人员可以清楚地了解到，为描述的方便和简洁，仅以上述各功能模块的划分进行举例说明，实际应用中，可以根据需要而将上述功能分配由不同的功能模块完成，即将装置的内部结构划分成不同的功能模块，以完成以上描述的全部或者部分功能。
在本申请所提供的几个实施例中，应该理解到，所揭露的装置和方法，可以通过其它的方式实现。例如，以上所描述的装置实施例仅仅是示意性的，例如，所述模块或单元的划分，仅仅为一种逻辑功能划分，实际实现时可以有另外的划分方式，例如多个单元或组件可以结合或者可以集成到另一个装置，或一些特征可以忽略，或不执行。另一点，所显示或讨论的相互之间的耦合或直接耦合或通信连接可以是通过一些接口，装置或单元的间接耦合或通信连接，可以是电性，机械或其它的形式。
所述作为分离部件说明的单元可以是或者也可以不是物理上分开的，作为单元显示的部件可以是一个物理单元或多个物理单元，即可以位于一个地方，或者也可以分布到多个不同地方。可以根据实际的需要选择其中的部分或者全部单元来实现本实施例方案的目的。
另外，在本申请各个实施例中的各功能单元可以集成在一个处理单元中，也可以是各个单元单独物理存在，也可以两个或两个以上单元集成在一个单元中。上述集成的单元既可以采用硬件的形式实现，也可以采用软件功能单元的形式实现。
所述集成的单元如果以软件功能单元的形式实现并作为独立的产品销售或使用时，可以存储在一个可读取存储介质中。基于这样的理解，本申请实施例的技术方案本质上或者说对现有技术做出贡献的部分或者该技术方案的全部或部分可以以软件产品的形式体现出来，该软件产品存储在一个存储介质中，包括若干指令用以使得一个设备(可以是单片机，芯片等)或处理器执行本申请各个实施例所述方法的全部或部分步 骤。而前述的存储介质包括：U盘、移动硬盘、ROM、RAM、磁碟或者光盘等各种可以存储程序代码的介质。
以上所述，仅为本申请的具体实施方式，但本申请的保护范围并不局限于此，任何在本申请揭露的技术范围内的变化或替换，都应涵盖在本申请的保护范围之内。因此，本申请的保护范围应以所述权利要求的保护范围为准。

Claims (12)

  1. 一种预测路损的方法，其特征在于，所述方法包括：
    获取样本数据，其中，所述样本数据包括影响发射机与小区间无线传播的特征参数、所述特征参数的参数值、以及所述小区中各接收机与所述发射机间的实测路损，所述特征参数的参数值与坐标位置对应，所述实测路损与所述接收机的坐标位置对应；
    对所述特征参数进行栅格化处理生成特征图像，其中，所述特征图像包括至少一个像素点，每个所述像素点对应一个参数值；
    根据所述特征图像和所述实测路损构建深度学习模型，其中，所述深度学习模型的输入变量为所述特征图像，所述深度学习模型的输出变量用于表示所述小区中各接收机与所述发射机间的路损；
    利用构建后的深度学习模型预测路损。
  2. 根据权利要求1所述的方法，其特征在于，对所述特征参数进行栅格化处理生成特征图像，包括：
    以所述发射机为中心建立栅格图像，其中，所述栅格图像包括至少一个栅格，不同栅格对应不同的坐标位置；
    遍历所述栅格图像中的每个栅格，将所述栅格作为一个像素点，将所述像素点对应的所述特征参数的参数值作为所述像素点的取值；
    将遍历后的栅格图像作为所述特征图像。
  3. 根据权利要求1和2所述的方法，其特征在于，
    所述特征参数包括：建筑物高度、地物类型、海拔高度、与所述发射机间的距离、所述发射机的水平方位角、所述发射机的垂直方位角、所述发射机距地面的高度、所述发射机发送信号的频率以及所述发射机的发射功率中的至少一种特征。
  4. 根据权利要求1-3任一项所述的方法，其特征在于，在根据所述特征图像和所述实测路损构建深度学习模型之前，所述方法还包括：
    对所述样本数据进行过滤处理；
    对过滤后的样本数据中的特征参数的参数值和实测路损进行归一化处理；
    所述对特征参数进行栅格化处理生成特征图像，包括：
    对所述归一化处理后的特征参数进行栅格化处理生成特征图像，并通过模型翻转或者模型旋转增加特征图像的个数。
  5. 根据权利要求1-4任一项所述的方法，其特征在于，
    所述深度学习模型的输出变量为所述路损对应的特征图像或者所述路损对应的向量。
  6. 一种预测路损的装置，其特征在于，所述装置包括：
    获取单元，用于获取样本数据，其中，所述样本数据包括影响发射机与小区间无线传播的特征参数、所述特征参数的参数值、以及所述小区中的接收机与所述发射机间的实测路损，所述特征参数的参数值与坐标位置对应，所述实测路损与所述接收机的坐标位置对应；
    图像生成单元，用于对所述特征参数进行栅格化处理生成特征图像，其中，所述特征图像包括至少一个像素点，每个所述像素点对应一个参数值；
    模型构建单元，用于根据所述特征图像和所述实测路损构建深度学习模型，其中，所述深度学习模型的输入变量为所述特征图像，所述深度学习模型的输出变量用于表示所述发射机与所述小区间的路损；
    预测单元，用于利用构建后的深度学习模型预测路损。
  7. 根据权利要求6所述的装置，其特征在于，
    所述图像生成单元，具体用于以所述发射机为中心建立栅格图像，其中，所述栅格图像包括至少一个栅格，不同栅格对应不同的坐标位置；
    遍历所述栅格图像中的每个栅格，将所述栅格作为一个像素点，将所述像素点对应的所述特征参数的参数值作为所述像素点的取值；
    将遍历后的栅格图像作为所述特征图像。
  8. 根据权利要求6和7所述的装置，其特征在于，
    所述特征参数包括：建筑物高度、地物类型、海拔高度、与所述发射机间的距离、所述发射机的水平方位角、所述发射机的垂直方位角、所述发射机距地面的高度、所述发射机发送信号的频率以及所述发射机的发射功率中的至少一种特征。
  9. 根据权利要求6-8任一项所述的装置，其特征在于，所述装置还包括：
    过滤清洗单元，用于在所述模型构建单元根据所述特征图像和所述实测路损构建深度学习模型之前，对所述样本数据进行过滤处理，对过滤后的样本数据中的特征参数的参数值和实测路损进行归一化处理；
    所述图像生成单元，具体用于对所述归一化处理后的特征参数进行栅格化处理生成特征图像，并通过模型翻转或者模型旋转增加特征图像的个数。
  10. 根据权利要求6-9任一项所述的装置，其特征在于，
    所述深度学习模型的输出变量为所述路损对应的特征图像或者所述路损对应的向量。
  11. 一种预测路损的装置，包括：至少一个处理器，以及存储器；其特征在于，
    所述存储器用于存储计算机程序，使得所述计算机程序被所述至少一个处理器执行时实现如权利要求1-5中任一项所述的预测路损的方法。
  12. 一种计算机存储介质，其上存储有计算机程序，其特征在于，所述程序被处理器执行时实现如权利要求1-5中任一项所述的预测路损的方法。
PCT/CN2018/114881 2017-12-05 2018-11-09 一种预测路损的方法及装置 WO2019109780A1 (zh)

Priority Applications (1)

Application Number Priority Date Filing Date Title
EP18886137.1A EP3687210A4 (en) 2017-12-05 2018-11-09 METHOD AND DEVICE FOR PREDICTING PATH LOSS

Applications Claiming Priority (2)

Application Number Priority Date Filing Date Title
CN201711271150.6A CN109874146B (zh) 2017-12-05 2017-12-05 一种预测路损的方法及装置
CN201711271150.6 2017-12-05

Publications (1)

Publication Number Publication Date
WO2019109780A1 true WO2019109780A1 (zh) 2019-06-13

Family

ID=66750812

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/CN2018/114881 WO2019109780A1 (zh) 2017-12-05 2018-11-09 一种预测路损的方法及装置

Country Status (3)

Country Link
EP (1) EP3687210A4 (zh)
CN (1) CN109874146B (zh)
WO (1) WO2019109780A1 (zh)

Families Citing this family (9)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN110990500B (zh) * 2019-03-29 2023-12-15 天维讯达(湖南)科技有限公司 传播路径模型地图建立方法及路径损耗确定方法
CN111147163A (zh) * 2019-12-17 2020-05-12 南京航空航天大学 一种基于dnn神经网络的无线通信链路损耗预测方法
CN110912627B (zh) * 2019-12-24 2021-05-14 电子科技大学 一种基于数据驱动的小区接收功率预测方法
US11528620B2 (en) * 2020-08-14 2022-12-13 Samsung Electronics Co., Ltd. Generating and calibrating signal strength prediction in a wireless network
CN114615693B (zh) * 2021-11-18 2024-04-16 广州亚信技术有限公司 网络容量预测方法、装置、电子设备及计算机存储介质
JPWO2023170762A1 (zh) * 2022-03-07 2023-09-14
GB202209921D0 (en) * 2022-07-06 2022-08-17 Samsung Electronics Co Ltd Artificial intelligence and machine learning capability indication
CN115297536B (zh) * 2022-09-26 2022-12-16 网络通信与安全紫金山实验室 功率控制方法、装置及存储介质
WO2024122689A1 (ko) * 2022-12-09 2024-06-13 삼성전자 주식회사 무선 통신 시스템에서 다중 업링크 캐리어 접속을 지원하는 방법 및 장치

Citations (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN101998411A (zh) * 2010-12-10 2011-03-30 上海百林通信软件有限公司 一种基于扫频及传播模型覆盖预测的频率优化方法
WO2012075805A1 (zh) * 2010-12-06 2012-06-14 中兴通讯股份有限公司 利用路测数据进行自动小区规划优化的方法及装置
CN102869020A (zh) * 2011-07-08 2013-01-09 中国移动通信集团湖南有限公司 一种无线网络优化的方法及装置

Family Cites Families (7)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN103379510B (zh) * 2012-04-23 2018-09-25 中兴通讯股份有限公司 一种利用mr数据进行小区规划的方法及装置
WO2014011191A1 (en) * 2012-07-13 2014-01-16 Adaptive Spectrum And Signal Alignment, Inc. Method and system for performance estimation of a communication link
US8768337B2 (en) * 2012-10-26 2014-07-01 International Business Machines Corporation Base station power control in a mobile network
US20150156647A1 (en) * 2013-12-03 2015-06-04 Motorola Mobility Llc Methods and Devices for Path-Loss Estimation
WO2016001473A1 (en) * 2014-06-30 2016-01-07 Nokia Technologies Oy System and method for radio and other parameter estimation based on maps
CN105430664B (zh) * 2015-10-30 2019-05-28 上海华为技术有限公司 一种基于分类拟合预测传播路损的方法和装置
US9913146B2 (en) * 2016-01-06 2018-03-06 Southern Methodist University Using geographical features to reduce in-field propagation experimentation

Patent Citations (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
WO2012075805A1 (zh) * 2010-12-06 2012-06-14 中兴通讯股份有限公司 利用路测数据进行自动小区规划优化的方法及装置
CN101998411A (zh) * 2010-12-10 2011-03-30 上海百林通信软件有限公司 一种基于扫频及传播模型覆盖预测的频率优化方法
CN102869020A (zh) * 2011-07-08 2013-01-09 中国移动通信集团湖南有限公司 一种无线网络优化的方法及装置

Also Published As

Publication number Publication date
CN109874146A (zh) 2019-06-11
EP3687210A4 (en) 2020-11-04
CN109874146B (zh) 2020-12-25
EP3687210A1 (en) 2020-07-29

Similar Documents

Publication Publication Date Title
WO2019109780A1 (zh) 一种预测路损的方法及装置
US11388610B2 (en) Detecting radio coverage problems
CN110731096B (zh) 使用深度神经网络预测电信网络中的接收信号强度
US11954879B2 (en) Methods, systems and apparatus to optimize pipeline execution
CN105430664B (zh) 一种基于分类拟合预测传播路损的方法和装置
ES2743234T3 (es) Aparato y método para utilización en comunicación por radio
WO2020034616A1 (zh) 一种信号覆盖效果图的获取方法及网络设备
CN110786046A (zh) 使用深度学习优化蜂窝网络
EP3990939B1 (en) Method and apparatus for localization
US11012978B2 (en) Uplink scheduling-free transmission method and device with groups of PDMA resource pools
Yan et al. Improved hop‐based localisation algorithm for irregular networks
CN110139359A (zh) 一种干扰源定位处理方法及装置
CN111050387B (zh) 基于能效估计的基站休眠方法、装置、电子设备及介质
Huang et al. Estimating distances via connectivity in wireless sensor networks
KR102210693B1 (ko) 경계 비트맵을 이용하는 이미지 분류 개선 기법
US20240054721A1 (en) Image rendering based on light baking
EP4115641A1 (en) Method, electronic device and non-transitory computer-readable storage medium for determining indoor radio transmitter distribution
WO2022242475A1 (zh) 一种无线信道多径信息的确定方法、装置及相关设备
Kim et al. Achievable rate estimation of IEEE 802.11 ad visual big-data uplink access in cloud-enabled surveillance applications
TWI491192B (zh) 無線電子裝置及其無線傳輸方法
WO2024050803A1 (zh) 感知方法、装置和系统
US20230354251A1 (en) Method for increasing accuracy of position estimation, system, and storage medium
WO2024174176A1 (zh) 一种通信方法及装置
CN118574179A (zh) 逻辑小区的切换控制方法、装置、计算机设备和存储介质
CN116456349A (zh) 信道分配方法及相关装置

Legal Events

Date Code Title Description
121 Ep: the epo has been informed by wipo that ep was designated in this application

Ref document number: 18886137

Country of ref document: EP

Kind code of ref document: A1

ENP Entry into the national phase

Ref document number: 2018886137

Country of ref document: EP

Effective date: 20200424

NENP Non-entry into the national phase

Ref country code: DE