Internet of vehicles channel path loss prediction method, system, device and storage medium

Abstract

The application discloses a vehicle-to-everything channel path loss prediction method, system, device and storage medium, wherein the method comprises the following steps: acquiring an RGB image, a depth image and laser radar point cloud data in front of a vehicle during driving of a plurality of vehicles; inputting the acquired data into a trained channel path loss prediction model to output a channel path loss prediction result; wherein the trained channel path loss prediction model is used for feature extraction on the laser radar point cloud data to obtain point cloud data features; feature extraction is performed on the depth image to obtain depth image features; feature extraction is performed on the RGB image to obtain RGB image features; feature fusion is performed on the obtained point cloud data features, depth image features and RGB image features to obtain fused features; and the fused features are predicted to output a channel path loss prediction value.

Description

Technical Field

[0001] This invention relates to the field of wireless communication technology, and in particular to a method, system, device, and storage medium for predicting channel path loss in vehicle-to-everything (V2X) networks. Background Technology

[0002] The statements in this section merely refer to the background art related to this invention and do not necessarily constitute prior art.

[0003] For urban intersection scenarios, path loss prediction in vehicle-to-everything (V2X) communication is a critical technology, as its accuracy directly impacts the reliability and performance of vehicle-to-vehicle communication. In highly dynamic V2X scenarios, vehicle communication paths are influenced by various factors, such as building density, traffic flow, and road structure. These factors all increase the uncertainty of path loss to varying degrees, posing challenges to V2X path loss prediction.

[0004] Existing traditional models have some shortcomings in solving the path loss prediction problem in such complex and dynamic traffic scenarios. These shortcomings manifest in their inability to accurately capture the nonlinearity, non-stationarity, and dynamics of vehicular communication paths in urban environments. Accurate path loss prediction is crucial for the performance optimization and security of vehicular communication systems; therefore, a more flexible path loss prediction method is needed. Summary of the Invention

[0005] To address the shortcomings of existing technologies, this invention provides a method, system, device, and storage medium for predicting path loss in vehicle-to-everything (V2X) channels. The algorithm of this invention is concise and easy to use, and can accurately predict the path loss distribution in complex and dynamic traffic scenarios at urban intersections.

[0006] On the one hand, a method for predicting path loss in vehicle-to-everything (V2X) channels is provided;

[0007] Methods for predicting path loss in vehicle-to-everything (V2X) communication channels include:

[0008] Acquire RGB images, depth images, and LiDAR point cloud data of the front of several vehicles during their operation;

[0009] The acquired data is input into the trained channel path loss prediction model, and the channel path loss prediction result is output.

[0010] The trained channel path loss prediction model is used to extract features from lidar point cloud data to obtain point cloud data features; extract features from depth images to obtain depth image features; extract features from RGB images to obtain RGB image features; fuse the obtained point cloud data features, depth image features, and RGB image features to obtain fused features; and predict the fused features to output the predicted channel path loss value.

[0011] On the other hand, a vehicle-to-everything (V2X) channel path loss prediction system is provided.

[0012] The vehicle-to-everything (V2X) channel path loss prediction system includes:

[0013] The acquisition module is configured to acquire RGB images, depth images, and LiDAR point cloud data in front of several vehicles during their operation.

[0014] The output module is configured to input the acquired data into the trained channel path loss prediction model and output the channel path loss prediction result.

[0015] The trained channel path loss prediction model is used to extract features from lidar point cloud data to obtain point cloud data features; extract features from depth images to obtain depth image features; extract features from RGB images to obtain RGB image features; fuse the obtained point cloud data features, depth image features, and RGB image features to obtain fused features; and predict the fused features to output the predicted channel path loss value.

[0016] Furthermore, an electronic device is also provided, including:

[0017] Memory, used for non-transitory storage of computer-readable instructions; and

[0018] Processor, for executing the computer-readable instructions,

[0019] When the computer-readable instructions are executed by the processor, they perform the method described in the first aspect above.

[0020] In another aspect, a storage medium is also provided for non-transitory storage of computer-readable instructions, wherein when the non-transitory computer-readable instructions are executed by a computer, the instructions of the method described in the first aspect are executed.

[0021] In another aspect, a computer program product is also provided, including a computer program that, when run on one or more processors, is used to implement the method described in the first aspect above.

[0022] The above technical solution has the following advantages or beneficial effects:

[0023] To address the problem of path loss prediction in complex and dynamic traffic scenarios at urban intersections, this invention proposes a method for path loss prediction that combines perception information acquired by vehicle-side sensing devices with deep learning. The algorithm is concise, easy to use, and can accurately predict the path loss distribution in complex and dynamic traffic scenarios at urban intersections. Attached Figure Description

[0024] The accompanying drawings, which form part of this invention, are used to provide a further understanding of the invention. The illustrative embodiments of the invention and their descriptions are used to explain the invention and do not constitute an improper limitation of the invention.

[0025] Figure 1 This is a flowchart of the method in Example 1;

[0026] Figure 2 This is the PointNet++ network structure of Example 1. Detailed Implementation

[0027] It should be noted that the following detailed descriptions are exemplary and intended to provide further illustration of the invention. Unless otherwise specified, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this invention pertains.

[0028] With the rapid development of artificial intelligence technology, deep learning-based path loss prediction methods offer new possibilities for addressing these challenges. By collecting sensory information from vehicle-mounted perception devices, including RGB images, depth images, and LiDAR point cloud data, a perception dataset for path loss prediction is constructed. Using this large-scale dataset for training, the deep learning model can automatically learn and extract environmental features in urban intersection scenarios, mapping environmental information to path loss information and dynamically adjusting the prediction results based on real-time environmental conditions. This method can better adapt to the diversity and dynamism of urban traffic node communication environments, improving the accuracy and stability of path loss prediction, thus providing important reference for the design and optimization of vehicle-to-everything (V2X) communication systems.

[0029] Example 1

[0030] This embodiment provides a method for predicting path loss in vehicle-to-everything (V2X) communication channels;

[0031] Methods for predicting path loss in vehicle-to-everything (V2X) communication channels include:

[0032] S101: Acquire RGB images, depth images, and LiDAR point cloud data of the front of several vehicles during their operation;

[0033] S102: Input the acquired data into the trained channel path loss prediction model and output the channel path loss prediction result;

[0034] The trained channel path loss prediction model is used to extract features from lidar point cloud data to obtain point cloud data features; extract features from depth images to obtain depth image features; extract features from RGB images to obtain RGB image features; fuse the obtained point cloud data features, depth image features, and RGB image features to obtain fused features; and predict the fused features to output the predicted channel path loss value.

[0035] Furthermore, the training process of the trained channel path loss prediction model includes:

[0036] Construct a training set, which consists of RGB images, depth images, and lidar point cloud data with known channel path loss prediction values;

[0037] The training set is input into the channel path loss prediction model to train the model. Training is stopped when the total loss function value of the channel path loss prediction model no longer decreases, and the trained channel path loss prediction model is obtained.

[0038] Furthermore, the construction of the training set includes:

[0039] The raw data consists of the effective path loss values ​​obtained at various receivers throughout the scene, with each vehicle acting as the transmitter; all of these are floating-point numbers. Each vehicle is equipped with an antenna, an RGB image camera, and a LiDAR.

[0040] The raw data is transformed into a probability distribution with a step size of 2.5 dB and a range of 0 to 250 dB.

[0041] The path loss distribution of all transmitters is integrated. If there is data for n vehicles, the processed data is a path loss distribution data of 100 rows and n columns, which is taken as the true value.

[0042] The transmitter is an omnidirectional single antenna mounted on the roof of the vehicle. The receiver is an omnidirectional single antenna laid on the ground. The receivers are evenly laid out in the scene to be predicted at 2m intervals. The transmitter emits electromagnetic waves, and the receiver measures the degree of power attenuation, which is the path loss value.

[0043] Furthermore, the total loss function is specifically expressed as follows:

[0044]

[0045] Where MSE represents the mean squared error loss function, and n is the number of path loss values. y represents the predicted path loss value, and y represents the true path loss value.

[0046] Further, in step S101: acquire RGB images, depth images, and lidar point cloud data of the front of several vehicles during their operation;

[0047] The RGB images are captured by a camera mounted at the front of the vehicle.

[0048] The depth images are acquired using a depth camera, which is mounted in front of the vehicle.

[0049] The lidar point cloud data is collected by lidar, which is installed at the front of the vehicle.

[0050] Further, step S101: acquiring RGB images, depth images, and LiDAR point cloud data of the area in front of several vehicles during their operation; also includes:

[0051] If all the RGB images captured by the vehicles are stitched together, and there are n cameras, and the size of each captured image is H×W, then the RGB image data is stitched together in the channel dimension as (3n, H, W).

[0052] The depth images collected by all vehicles are stitched together. If there are n depth cameras, the depth map data is stitched together in the channel dimension as (1n, H, W).

[0053] The lidar point cloud data collected from all vehicles are stitched together. If there are n lidars, the lidar point cloud is stitched together in terms of the number of point clouds as (1, 3, n*x), where x is the number of lidar point clouds sampled from each lidar.

[0054] Furthermore, such as Figure 1 As shown, the trained channel path loss prediction model includes:

[0055] First feature extraction layer and second feature extraction layer;

[0056] The input end of the first feature extraction layer is used to input LiDAR point cloud data;

[0057] The input of the second feature extraction layer is used to input the merged depth image and RGB image; the merged depth image and RGB image are merged in the channel dimension; the depth image is used as the fourth dimension information of the RGB image;

[0058] The outputs of the first and second feature extraction layers are both connected to the input of the feature fusion layer; the output of the feature fusion layer is connected to the input of the classification layer, and the output of the classification layer is used to output the prediction result.

[0059] Furthermore, the first feature extraction layer is implemented using a Pointnet++ network; the second feature extraction layer is implemented using a ResNet50 network; the feature fusion layer is implemented using an attention mechanism layer; and the classification layer is implemented using a multilayer perceptron (MLP).

[0060] Furthermore, such as Figure 2 As shown, the Pointnet++ network includes:

[0061] The first multi-scale grouping set abstraction module, the second multi-scale grouping set abstraction module, and the point cloud feature direct abstraction module are connected in sequence.

[0062] Furthermore, the internal structure of the first multi-scale grouping set abstraction module and the second multi-scale grouping set abstraction module is the same. The first multi-scale grouping set abstraction module includes:

[0063] The first convolutional module, the second convolutional module, and the third convolutional module are connected in sequence.

[0064] Furthermore, the point cloud feature direct abstraction module includes:

[0065] The fourth, fifth, and sixth convolutional modules are connected in sequence.

[0066] Furthermore, the internal structures of the first, second, third, fourth, fifth, sixth, seventh, eighth, and ninth convolutional modules are identical; the first convolutional module includes:

[0067] The convolutional layer, batch normalization layer, activation function layer, and max pooling layer are connected in sequence.

[0068] Furthermore, the first multi-scale grouping set abstraction module is used to abstract the point cloud data. First, by defining different radii and the number of sampling points, the input point cloud is grouped and sampled at multiple scales to capture local features. For each sampling point, feature extraction and nonlinear transformation are performed using convolutional layers and batch normalization layers to capture local feature information. Finally, the local features are fused through pooling operations to obtain a feature map.

[0069] Furthermore, the second multi-scale grouping set abstraction module is used to abstract the point cloud data. First, by defining different radii and the number of sampling points, the input point cloud is grouped and sampled at multiple scales to capture local features. For each sampling point, convolutional layers and batch normalization layers are used for feature extraction and nonlinear transformation to capture local feature information. Finally, the local features are fused through pooling operations to obtain a feature map.

[0070] Furthermore, the point cloud feature direct abstraction module is used to realize the hierarchical abstraction of point clouds. First, the point cloud is grouped and downsampled to obtain new sampling points and their features; then, convolutional layers and batch normalization layers are used for feature extraction, and the ReLU activation function is applied; finally, the features are integrated through pooling operations to obtain an abstract feature representation.

[0071] Furthermore, the Pointnet++ network is used for feature extraction from lidar point cloud data.

[0072] Furthermore, the Resnet50 network includes:

[0073] The convolutional blocks conv1, conv2_x, conv3_x, conv4_x, conv5_x, average pooling layer, fully connected layer and activation function layer are connected in sequence.

[0074] The first convolutional block conv1 is a 7×7 convolutional layer;

[0075] The second convolutional block conv2_x includes three convolutional sub-blocks connected in sequence, each of which includes a 1×1 convolutional layer, a 3×3 convolutional layer, and a 1×1 convolutional layer connected in sequence.

[0076] The third convolutional block conv3_x includes: four convolutional sub-blocks connected in sequence, each convolutional sub-block including a 1×1 convolutional layer, a 3×3 convolutional layer and a 1×1 convolutional layer connected in sequence;

[0077] The fourth convolutional block conv4_x includes: six convolutional sub-blocks connected in sequence, each convolutional sub-block including a 1×1 convolutional layer, a 3×3 convolutional layer and a 1×1 convolutional layer connected in sequence;

[0078] The fifth convolutional block conv5_x includes three convolutional sub-blocks connected in sequence, each of which includes a 1×1 convolutional layer, a 3×3 convolutional layer, and a 1×1 convolutional layer connected in sequence.

[0079] Furthermore, the ResNet50 network is used to extract features from the merged depth image and RGB image.

[0080] The ResNet50 architecture consists of an input layer, convolutional layers, and pooling layers. The first convolutional block, conv1, uses a 7×7 kernel with a stride of 2, padding of 3, and 64 output channels. This is followed by a batch normalization layer and a ReLU activation function. The pooling layer uses a 3×3 core window and a pooling operation with a stride of 2 to reduce the size of the feature map. The model also includes four stages of residual blocks: a second convolutional block conv2_x, a third convolutional block conv3_x, a fourth convolutional block conv4_x, and a fifth convolutional block conv5_x, each containing a different number of residual blocks. Each residual block uses a Bottleneck-type structure, consisting of a series of 1×1, 3×3, and 1×1 convolutional kernels. Each step connects to a batch normalization layer and a ReLU activation function. Finally, a global average pooling layer pools the feature map to a size of (1×1).

[0081] The feature fusion layer concatenates image features and point cloud features along the channel dimension, or uses a multi-head attention mechanism to fuse feature maps and point cloud features.

[0082] The classification layer consists of convolutional layers and fully connected layers. The classification layer uses convolutional layers to extract features from the fused features, and then the fully connected layers convert the extracted features into a one-dimensional tensor. The convolutional layers extract data features step by step, and the fully connected layers further process and output the final result, which is a tensor of size 100×n.

[0083] This invention utilizes RGB images, depth images, and LiDAR point cloud data acquired by vehicle-mounted sensing devices as input to an artificial neural network. Different convolutional neural networks are used to extract features from these three types of sensing information. Subsequently, feature fusion is performed on different modal information from all known devices in the scene at the same time to obtain more comprehensive information. Finally, the fused feature information is decoded and input into a multilayer perceptron to obtain the predicted path loss distribution.

[0084] Example 2

[0085] This embodiment provides a vehicle-to-everything (V2X) channel path loss prediction system;

[0086] The vehicle-to-everything (V2X) channel path loss prediction system includes:

[0087] The acquisition module is configured to acquire RGB images, depth images, and LiDAR point cloud data in front of several vehicles during their operation.

[0088] The output module is configured to input the acquired data into the trained channel path loss prediction model and output the channel path loss prediction result.

[0089] The trained channel path loss prediction model is used to extract features from lidar point cloud data to obtain point cloud data features; extract features from depth images to obtain depth image features; extract features from RGB images to obtain RGB image features; fuse the obtained point cloud data features, depth image features, and RGB image features to obtain fused features; and predict the fused features to output the predicted channel path loss value.

[0090] It should be noted that the above-mentioned acquisition module and output module correspond to steps S101 to S102 in Embodiment 1. The examples and application scenarios implemented by the above modules and corresponding steps are the same, but are not limited to the content disclosed in Embodiment 1. It should be noted that the above modules, as part of the system, can be executed in a computer system such as a set of computer-executable instructions.

[0091] The descriptions of each embodiment in the above embodiments have different focuses. For parts not described in detail in a certain embodiment, please refer to the relevant descriptions in other embodiments.

[0092] The proposed system can be implemented in other ways. For example, the system embodiments described above are merely illustrative, and the division of modules described above is only a logical functional division. In actual implementation, there may be other division methods. For example, multiple modules may be combined or integrated into another system, or some features may be ignored or not executed.

[0093] Example 3

[0094] This embodiment also provides an electronic device, including: one or more processors, one or more memories, and one or more computer programs; wherein, the processor is connected to the memory, and the one or more computer programs are stored in the memory. When the electronic device is running, the processor executes the one or more computer programs stored in the memory to cause the electronic device to perform the method described in Embodiment 1.

[0095] It should be understood that in this embodiment, the processor can be a central processing unit (CPU), or it can be other general-purpose processors, digital signal processors (DSPs), application-specific integrated circuits (ASICs), programmable gate arrays (FPGAs), or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. The general-purpose processor can be a microprocessor or any conventional processor, etc.

[0096] Memory may include read-only memory and random access memory, and provides instructions and data to the processor. A portion of memory may also include non-volatile random access memory. For example, memory may also store information about the device type.

[0097] In the implementation process, each step of the above method can be completed by the integrated logic circuits in the processor hardware or by software instructions.

[0098] The method in Embodiment 1 can be directly implemented by a hardware processor, or implemented by a combination of hardware and software modules within the processor. The software modules can reside in readily available storage media in the art, such as random access memory, flash memory, read-only memory, programmable read-only memory, electrically erasable programmable memory, or registers. This storage medium is located in memory; the processor reads information from the memory and, in conjunction with its hardware, completes the steps of the above method. To avoid repetition, a detailed description is not provided here.

[0099] Those skilled in the art will recognize that the units and algorithm steps described in connection with the various examples of this embodiment can be implemented in electronic hardware or a combination of computer software and electronic hardware. Whether these functions are implemented in hardware or software depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementation should not be considered beyond the scope of this invention.

[0100] Example 4

[0101] This embodiment also provides a computer-readable storage medium for storing computer instructions, which, when executed by a processor, complete the method described in Embodiment 1.

[0102] The above description is merely a preferred embodiment of the present invention and is not intended to limit the invention. Various modifications and variations can be made to the present invention by those skilled in the art. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the scope of protection of the present invention.

Claims

1. A method for predicting path loss in vehicle-to-everything (V2X) communication channels, characterized by: include: Acquire RGB images, depth images, and LiDAR point cloud data of the area in front of several vehicles during their operation, including: If all the RGB images captured by the vehicles are stitched together, and there are n cameras with an image size of H×W, then the RGB image data will be stitched together in the channel dimension as (3n, H, W). The depth images collected by all vehicles are stitched together. If there are n depth cameras, the depth map data is stitched together in the channel dimension as (1n, H, W). The lidar point cloud data collected from all vehicles are stitched together. If there are n lidars, the lidar point cloud is stitched together in terms of the number of point clouds as (1, 3, n×x), where x is the number of lidar point clouds sampled from each lidar. The acquired data is input into the trained channel path loss prediction model, and the channel path loss prediction result is output. The trained channel path loss prediction model is used to extract features from lidar point cloud data to obtain point cloud data features; extract features from depth images to obtain depth image features; extract features from RGB images to obtain RGB image features; fuse the obtained point cloud data features, depth image features, and RGB image features to obtain fused features; and predict the fused features to output the predicted channel path loss value. The trained channel path loss prediction model includes: A first feature extraction layer and a second feature extraction layer; the input of the first feature extraction layer is used to input LiDAR point cloud data; the input of the second feature extraction layer is used to input the merged depth image and RGB image; the merged depth image and RGB image are merged in the channel dimension; the depth image is used as the fourth dimension information of the RGB image; the outputs of the first feature extraction layer and the second feature extraction layer are both connected to the input of the feature fusion layer; the output of the feature fusion layer is connected to the input of the classification layer, and the output of the classification layer is used to output the prediction result.

2. The method for predicting path loss in vehicle-to-everything (V2X) channels as described in claim 1, characterized in that, The training process for the trained channel path loss prediction model includes: Construct a training set, which consists of RGB images, depth images, and lidar point cloud data with known channel path loss prediction values; The training set is input into the channel path loss prediction model to train the model. Training is stopped when the total loss function value of the channel path loss prediction model no longer decreases, and the trained channel path loss prediction model is obtained.

3. The method for predicting path loss in vehicle-to-everything (V2X) channels as described in claim 2, characterized in that, The construction of the training set includes: The raw data consists of the effective path loss values ​​obtained by each vehicle as a transmitter at various receivers throughout the entire scenario, all of which are floating-point numbers; The raw data is transformed into a probability distribution with a step size of 2.5 dB and a range of 0 to 250 dB. The path loss distribution of all transmitters is integrated. If there is data for n vehicles, the processed data is the path loss distribution data in m rows and n columns, which is taken as the true value.

4. The method for predicting path loss in vehicle-to-everything (V2X) channels as described in claim 1, characterized in that, Acquire RGB images, depth images, and LiDAR point cloud data of the front of several vehicles during their operation; The RGB images are captured by a camera mounted at the front of the vehicle. The depth images are acquired using a depth camera, which is mounted in front of the vehicle. The lidar point cloud data is collected by lidar, which is installed at the front of the vehicle.

5. The method for predicting path loss in vehicle-to-everything (V2X) channels as described in claim 1, characterized in that, The first feature extraction layer is implemented using a Pointnet++ network; the second feature extraction layer is implemented using a ResNet50 network; the feature fusion layer is implemented using an attention mechanism layer; and the classification layer is implemented using a multilayer perceptron (MLP). The Pointnet++ network includes: a first multi-scale grouping set abstraction module, a second multi-scale grouping set abstraction module, and a point cloud feature direct abstraction module connected in sequence. The first multi-scale grouping abstraction module first groups and samples the input point cloud at multiple scales by defining different radii and the number of sampling points to capture local features; for each sampling point, it uses convolutional layers and batch normalization layers to extract features and perform nonlinear transformations to capture local feature information; finally, it fuses the local features through pooling operations to obtain a feature map. The second multi-scale grouping abstraction module first groups and samples the input point cloud at multiple scales by defining different radii and the number of sampling points to capture local features; for each sampling point, it uses convolutional layers and batch normalization layers to extract features and perform nonlinear transformations to capture local feature information; finally, it fuses the local features through pooling operations to obtain a feature map. The point cloud feature direct abstraction module is used to realize the hierarchical abstraction of point clouds. First, the point cloud is grouped and downsampled to obtain new sampling points and their features. Then, convolutional layers and batch normalization layers are used to extract features and apply the ReLU activation function. Finally, the features are integrated through pooling operations to obtain an abstract feature representation.

6. A vehicle-to-everything (V2X) channel path loss prediction system, characterized in that, include: The acquisition module is configured to acquire RGB images, depth images, and LiDAR point cloud data of the front of several vehicles during their operation, including: If all the RGB images captured by the vehicles are stitched together, and there are n cameras with an image size of H×W, then the RGB image data will be stitched together in the channel dimension as (3n, H, W). The depth images collected by all vehicles are stitched together. If there are n depth cameras, the depth map data is stitched together in the channel dimension as (1n, H, W). The lidar point cloud data collected from all vehicles are stitched together. If there are n lidars, the lidar point cloud is stitched together in terms of the number of point clouds as (1, 3, n×x), where x is the number of lidar point clouds sampled from each lidar. The output module is configured to input the acquired data into the trained channel path loss prediction model and output the channel path loss prediction result. The trained channel path loss prediction model is used to extract features from lidar point cloud data to obtain point cloud data features; extract features from depth images to obtain depth image features; extract features from RGB images to obtain RGB image features; fuse the obtained point cloud data features, depth image features, and RGB image features to obtain fused features; and predict the fused features to output the predicted channel path loss value. The trained channel path loss prediction model includes: A first feature extraction layer and a second feature extraction layer; the input of the first feature extraction layer is used to input LiDAR point cloud data; the input of the second feature extraction layer is used to input the merged depth image and RGB image; the merged depth image and RGB image are merged in the channel dimension; the depth image is used as the fourth dimension information of the RGB image; the outputs of the first feature extraction layer and the second feature extraction layer are both connected to the input of the feature fusion layer; the output of the feature fusion layer is connected to the input of the classification layer, and the output of the classification layer is used to output the prediction result.

7. An electronic device, characterized in that it comprises: Memory is used to store computer-readable instructions in a non-transitory manner. as well as Processor, for executing the computer-readable instructions, When the computer-readable instructions are executed by the processor, they perform the method described in any one of claims 1-5.

8. A storage medium characterized by being non-transitory. The system stores computer-readable instructions, wherein, when the non-transitory computer-readable instructions are executed by a computer, the instructions of the method according to any one of claims 1-5 are executed.

Images