Power-adjustable wireless communication module

Abstract

Disclosed is a power-adjustable wireless communication module. It first obtains the signal strength values of a received wireless signal at multiple predetermined time points within a predetermined time period. Next, it extracts the local signal strength features of the signal strength values at the multiple predetermined time points to obtain a sequence of local signal strength time-series feature vectors. Next, it extracts the topological relationships between the signal strength features of the sequence of local signal strength time-series feature vectors to obtain a topological feature matrix of relationship between signal strength features. Next, it fuses the sequence of local signal strength time-series feature vectors with the topological feature matrix of relationship between signal strength features to obtain a topological global signal strength time-series feature matrix. Finally, based on the topological global signal strength time-series feature matrix, it determines whether the power value of the communication module at the current time point needs to be increased. In this way, dynamic adjustment of the power value of the wireless communication module can be achieved.

Description

Technical Field

[0001] The present application relates to the field of wireless communications, and more specifically, to a power-adjustable wireless communication module. Background Art

[0002] A wireless communication module is an electronic device that can realize wireless data transmission. It usually consists of an antenna, a radio frequency circuit, a baseband processor, and an interface circuit.

[0003] The power value of the wireless communication module is one of the important factors affecting the quality of wireless communication. Too low a power value will lead to reduced communication efficiency or even cause communication interruption, while too high a power value will greatly generate unnecessary energy loss.

[0004] Therefore, a power-adjustable wireless communication module is desired. Summary of the Invention

[0005] In view of this, the present application proposes a power-adjustable wireless communication module, which can use deep learning-based artificial intelligence technology to process and analyze the signal strength value of the received wireless signal within a predetermined time, so as to extract the time series change characteristics of the signal strength, that is, the fluctuation characteristics of the received wireless signal in the time dimension, and thereby realize dynamic adjustment of the power value of the wireless communication module.

[0006] According to one aspect of the present application, a power-adjustable wireless communication module is provided, comprising:

[0007] a signal strength value acquiring unit, configured to acquire signal strength values of a received wireless signal at a plurality of predetermined time points within a predetermined time period;

[0008] a local feature extraction unit, configured to extract signal strength local features of the signal strength values at the plurality of predetermined time points to obtain a sequence of signal strength local time series feature vectors;

[0009] A topological relationship extraction unit, configured to extract the topological relationship between signal strength features between the sequences of the signal strength local time series feature vectors to obtain a topological feature matrix of the relationship between signal strength features;

[0010] a fusion unit, configured to fuse the sequence of the signal strength local time series feature vectors and the relationship topology feature matrix between the signal strength features to obtain a relationship topology global signal strength time series feature matrix; and

[0011] The power control analysis unit is used to determine whether the power value of the communication module at the current time point needs to be increased based on the relationship topology global signal strength time series characteristic matrix.

[0012] According to an embodiment of the present application, it first obtains the signal strength values of the received wireless signal at multiple predetermined time points within a predetermined time period, then extracts the signal strength local features of the signal strength values at the multiple predetermined time points to obtain a sequence of signal strength local time series feature vectors, then extracts the topological relationship between the signal strength features of the sequence of the signal strength local time series feature vectors to obtain a signal strength feature relationship topological feature matrix, then fuses the sequence of the signal strength local time series feature vectors and the signal strength feature relationship topological feature matrix to obtain a relationship topology global signal strength time series feature matrix, and finally, based on the relationship topology global signal strength time series feature matrix, determines whether the power value of the communication module at the current time point needs to be increased. In this way, dynamic adjustment of the power value of the wireless communication module can be achieved.

[0013] Further features and aspects of the present application will become apparent from the following detailed description of exemplary embodiments with reference to the attached drawings. BRIEF DESCRIPTION OF THE DRAWINGS

[0014] The accompanying drawings, which are incorporated in and constitute a part of the specification, illustrate exemplary embodiments, features, and aspects of the application and, together with the description, serve to explain the principles of the application.

[0015] Figure 1 A block diagram of a power-adjustable wireless communication module according to an embodiment of the present application is shown.

[0016] Figure 2 A block diagram of the local feature extraction unit in the power-adjustable wireless communication module according to an embodiment of the present application is shown.

[0017] Figure 3 A block diagram of the data pre-processing subunit in the power-adjustable wireless communication module according to an embodiment of the present application is shown.

[0018] Figure 4 A block diagram of the topology relationship extraction unit in the power-adjustable wireless communication module according to an embodiment of the present application is shown.

[0019] Figure 5 A block diagram of the power control analysis unit in the power adjustable wireless communication module according to an embodiment of the present application is shown.

[0020] Figure 6 The flowchart of the power-adjustable wireless communication control method according to an embodiment of the present application is shown.

[0021] Figure 7 A schematic diagram illustrating the architecture of a power-adjustable wireless communication control method according to an embodiment of the present application is shown.

[0022] Figure 8A diagram showing an application scenario of a power-adjustable wireless communication module according to an embodiment of the present application is shown. DETAILED DESCRIPTION

[0023] The following will clearly and completely describe the technical solutions in the embodiments of the present application with reference to the accompanying drawings. Obviously, the embodiments described are only some embodiments of the present application, not all embodiments. Based on the embodiments of the present application, all other embodiments obtained by ordinary technicians in this field without making creative efforts also fall within the scope of protection of this application.

[0024] As used in this application and the claims, unless the context clearly indicates otherwise, the words "a," "an," "an," and / or "the" are not intended to refer to the singular but may include the plural. Generally speaking, the terms "comprises" and "include" only indicate the inclusion of the steps and elements specifically identified, and these steps and elements do not constitute an exclusive list. A method or apparatus may also include other steps or elements.

[0025] Various exemplary embodiments, features, and aspects of the present application will be described in detail below with reference to the accompanying drawings. The same reference numerals in the accompanying drawings represent elements with the same or similar functions. Although various aspects of the embodiments are shown in the accompanying drawings, the drawings are not necessarily drawn to scale unless otherwise indicated.

[0026] In addition, numerous specific details are provided in the detailed description below to better illustrate the present application. Those skilled in the art will appreciate that the present application can be practiced without certain specific details. In some instances, methods, means, components, and circuits well known to those skilled in the art are not described in detail in order to highlight the main purpose of the present application.

[0027] In response to the above technical problems, the technical concept of this application is to use artificial intelligence technology based on deep learning to process and analyze the signal strength value of the received wireless signal within a predetermined time, so as to extract the time series change characteristics of the signal strength, that is, the fluctuation characteristics of the received wireless signal in the time dimension, and thereby realize dynamic adjustment of the power value of the wireless communication module.

[0028] Based on this, Figure 1 FIG. 1 is a block diagram of a power adjustable wireless communication module according to an embodiment of the present application. Figure 1As shown, the power adjustable wireless communication module 100 according to the embodiment of the present application includes: a signal strength value acquisition unit 110, used to obtain the signal strength values of the received wireless signal at multiple predetermined time points within a predetermined time period; a local feature extraction unit 120, used to extract the signal strength local features of the signal strength values of the multiple predetermined time points to obtain a sequence of signal strength local time series feature vectors; a topological relationship extraction unit 130, used to extract the topological relationship between the signal strength features of the sequence of the signal strength local time series feature vectors to obtain a signal strength feature relationship topological feature matrix; a fusion unit 140, used to fuse the sequence of the signal strength local time series feature vectors and the signal strength feature relationship topological feature matrix to obtain a relationship topology global signal strength time series feature matrix; and a power control analysis unit 150, used to determine whether the power value of the communication module at the current time point needs to be increased based on the relationship topology global signal strength time series feature matrix.

[0029] It should be understood that the signal strength value acquisition unit 110 is responsible for measuring and recording the strength of the received wireless signal. The local feature extraction unit 120 generates a vector sequence representing the signal strength by extracting features, such as statistical features or frequency domain features, from the signal strength values. The topological relationship extraction unit 130 constructs a feature matrix representing the topological relationships between different signal strength feature vector sequences by analyzing the relationships between them. The fusion unit 140 fuses the local feature vector sequences with the topological feature matrix to obtain a global feature matrix that comprehensively represents the temporal characteristics of the signal strength. The power control analysis unit 150 determines whether the current power of the communication module is appropriate by analyzing the global feature matrix and, if necessary, makes corresponding adjustments. These units work together to adjust and optimize the power of the wireless communication module by acquiring signal strength values, extracting features, analyzing topological relationships, and controlling power.

[0030] Specifically, in the technical solution of this application, the signal strength values of the received wireless signal at multiple predetermined time points within a predetermined time period are first obtained. It should be understood that signal strength is one of the important indicators reflecting signal transmission quality. By analyzing and processing the signal strength, information about the current communication environment and transmission quality can be obtained, allowing power adjustment to be performed as needed. Specifically, these signal strength values at different time points can reflect and reveal the changing trends and fluctuations of signal strength.

[0031] Accordingly, if Figure 2As shown, the local feature extraction unit 120 includes: a data preprocessing subunit 121, used to perform data preprocessing on the signal strength values of the multiple predetermined time points to obtain a sequence of signal strength local time series input vectors; and a feature extraction subunit 122, used to perform feature extraction on the sequence of signal strength local time series input vectors using a deep learning network model to obtain a sequence of signal strength local time series feature vectors.

[0032] It should be understood that in the data preprocessing subunit 121, data preprocessing may include operations such as denoising, filtering, and normalization, which are intended to eliminate noise and interference so that the input vector is more accurate and reliable. In the feature extraction subunit 122, the deep learning network model can be a convolutional neural network (CNN), a recurrent neural network (RNN) or a variant model, which extracts a feature vector sequence with discriminative and characterizing capabilities by learning the feature representation of the input vector. The combination of the data preprocessing subunit and the feature extraction subunit allows the signal strength value to be preprocessed and feature extracted to obtain a more informative signal strength local time series feature vector sequence. These feature vector sequences will be used in subsequent steps such as topological relationship extraction and fusion to further analyze and optimize the power control of the communication module.

[0033] Specifically, if Figure 3As shown, the data preprocessing subunit 121 includes: a vectorization secondary subunit 1211, which is used to arrange the signal strength values of the multiple predetermined time points into a signal strength time series input vector according to the time dimension; and a vector segmentation secondary subunit 1212, which is used to vector segment the signal strength time series input vector to obtain a sequence of the signal strength local time series input vectors. It should be understood that vector segmentation is an important step in the data preprocessing subunit. Its function is to segment the signal strength time series input vector to obtain a sequence of signal strength local time series input vectors. The segmented local time series input vector sequence can provide more detailed and specific information for subsequent feature extraction and analysis. By segmenting the vector, the original signal strength time series input vector can be decomposed into multiple smaller local vectors, each local vector representing the signal strength within a specific time period. The benefits of doing so are as follows: 1. Providing finer-grained features: The segmented local vector can capture more subtle changes in the time dimension, allowing the feature extraction subunit to better analyze local time series features. 2. Adapting to Dynamic Changes: Wireless signal strength can vary significantly over time. By splitting the vector, it can better adapt to dynamic changes in signal strength and provide more accurate and real-time feature representation. 3. Improving Model Robustness: Splitting the vector reduces data redundancy and the interference of irrelevant information on feature extraction, thereby improving model robustness and performance. In summary, the role of vector splitting in data preprocessing is to split the signal strength time series input vector into a sequence of local time series input vectors, providing a more detailed, dynamic, and accurate feature representation, laying the foundation for subsequent feature extraction and analysis steps.

[0034] Then, the signal strength local features of the signal strength values at the plurality of predetermined time points are extracted to obtain a sequence of signal strength local time series feature vectors, that is, the variation pattern and characteristic distribution of the signal strength within the local time range are captured.

[0035] In a specific example of the present application, the encoding process of extracting the signal strength local features of the signal strength values of the multiple predetermined time points to obtain a sequence of signal strength local time series feature vectors includes: first arranging the signal strength values of the multiple predetermined time points into a signal strength time series input vector according to the time dimension; then, performing vector segmentation on the signal strength time series input vector to obtain a sequence of signal strength local time series input vectors; and then passing the sequence of signal strength local time series input vectors through a signal strength time series feature extractor based on a one-dimensional convolutional layer to obtain a sequence of signal strength local time series feature vectors.

[0036] Accordingly, the deep learning network model is a one-dimensional convolutional layer-based signal strength temporal feature extractor, wherein the one-dimensional convolutional layer-based signal strength temporal feature extractor includes an input layer, a one-dimensional convolutional layer, an activation layer, a pooling layer, and an output layer. Specifically, the feature extraction subunit 122 is configured to pass the sequence of signal strength local temporal series input vectors through the one-dimensional convolutional layer-based signal strength temporal feature extractor to obtain a sequence of signal strength local temporal series feature vectors.

[0037] It's worth noting that the one-dimensional convolutional layer is a commonly used neural network layer in deep learning. It is used to process data with temporal structure, such as time series and signal sequences. It plays an important role in feature extraction and pattern recognition tasks. A one-dimensional convolutional layer extracts local features in the temporal dimension of the input data through convolution. It performs a convolution operation on the input sequence by sliding a fixed-size convolution kernel (also called a filter) over it to produce an output sequence. The convolution kernel can be viewed as a set of learnable weight parameters that extract local features at different locations in the input sequence. Specifically, the main components of a one-dimensional convolutional layer include: 1. Input layer: This layer receives the input sequence, such as a sequence of local temporal input vectors of signal strength. 2. One-dimensional convolutional layer: This layer consists of multiple convolution kernels, each of which performs a convolution operation on the input sequence to extract local features. Different convolution kernels can capture different characteristic patterns. 3. Activation layer: This layer performs a nonlinear transformation on the output of the convolutional layer, introducing nonlinear relationships and increasing the network's expressive power. Common activation functions include ReLU, Sigmoid, and Tanh. 4. Pooling layer: Downsample the output of the convolution layer to reduce the dimension of the features, extract the main features and reduce the amount of data. Common pooling operations include maximum pooling and average pooling. 5. Output layer: The feature vector sequence obtained after convolution, activation and pooling operations is used as the output. The one-dimensional convolution layer has the following advantages in signal strength time series feature extraction: 1. Local feature extraction: The one-dimensional convolution layer can capture the local features of the input sequence through convolution operations, such as the signal strength change pattern at different time points, thereby extracting discriminative feature representations. 2. Parameter sharing: In the one-dimensional convolution layer, each convolution kernel shares weight parameters on the entire input sequence, reducing the number of parameters of the model and improving the efficiency and generalization ability of the model. 3. Translation invariance: The one-dimensional convolution layer is invariant to the translation of the input sequence, that is, no matter how many time steps the input sequence is translated, the convolution kernel can still extract the same features. In summary, the one-dimensional convolutional layer is an important neural network layer for processing time series data. It extracts local features through convolution operations and provides useful feature representations for subsequent feature extraction and analysis.

[0038] Next, the topological relationships between signal strength features in the sequences of local signal strength time series feature vectors are extracted to obtain a signal strength feature relationship topological feature matrix. In other words, the correlation and mutual influence between signal strength variation features in different time periods are captured to understand the complex time series characteristics of signal strength.

[0039] In a specific example of the present application, the encoding process of extracting the topological relationship between signal strength features between the sequence of signal strength local time series feature vectors to obtain a signal strength feature relationship topological feature matrix includes: first calculating the correlation between any two signal strength local time series feature vectors in the sequence of signal strength local time series feature vectors to obtain a signal strength feature relationship matrix; then passing the signal strength feature relationship matrix through a relational topological feature extractor based on a convolutional neural network model to obtain a signal strength feature relationship topological feature matrix.

[0040] Accordingly, if Figure 4 As shown, the topological relationship extraction unit 130 includes: a correlation calculation subunit 131, used to calculate the correlation between any two signal strength local time series feature vectors in the sequence of the signal strength local time series feature vectors to obtain a signal strength feature relationship matrix; and a relationship topology feature extraction subunit 132, used to pass the signal strength feature relationship matrix through a relationship topology feature extractor based on a convolutional neural network model to obtain the signal strength feature relationship topology feature matrix.

[0041] It should be understood that the role of the correlation calculation subunit 131 is to calculate the correlation between any two feature vectors in the signal strength local time series feature vector sequence, thereby obtaining a relationship matrix between signal strength features. By calculating the correlation, the similarity or correlation between different feature vectors can be measured, providing a basis for subsequent relationship topology feature extraction. The role of the relationship topology feature extraction subunit 132 is to extract the relationship topology feature matrix between signal strength features from the relationship matrix between signal strength features through a relationship topology feature extractor based on a convolutional neural network model. This subunit uses a convolutional neural network model to process and analyze the relationship matrix and extract relationship topology features. These topological relationship features can provide structural information between signal strengths, help understand and analyze the relationship between signal strengths, and thus provide a more comprehensive and accurate information basis for power control and analysis.

[0042] It should be understood that a convolutional neural network (CNN) is a deep learning model that excels at processing data with a grid structure and can automatically learn and extract features from the data. The core component of a CNN is the convolutional layer, which extracts local features from the input data through a convolution operation. The convolution operation slides a learnable filter (also known as a convolution kernel) over the input data and calculates the dot product between the filter and the input data to produce an output feature map. By operating multiple convolution kernels in parallel, the convolution layer is able to extract different feature patterns. Convolutional neural networks also include other important components, such as activation functions and pooling layers. Activation functions introduce nonlinear relationships and increase the network's expressive power. Common activation functions include ReLU, Sigmoid, and Tanh. Pooling layers are used to downsample the output of the convolutional layer, reducing the feature dimensionality, extracting key features, and reducing the amount of data. Convolutional neural networks typically consist of multiple alternating convolutional layers, activation functions, and pooling layers, ultimately mapping the extracted features to output categories via fully connected layers. During training, convolutional neural networks update their parameters using a backpropagation algorithm, enabling the network to automatically learn features from the input data and achieve excellent performance in tasks such as classification, object detection, and image segmentation. In short, convolutional neural networks are deep learning models specifically designed for processing grid-structured data. They extract local features through convolution operations, introduce nonlinear relationships through activation functions, and perform downsampling through pooling layers to learn and represent the features of the input data.

[0043] Specifically, the correlation calculation subunit 131 is configured to calculate the correlation between any two signal strength local time series feature vectors in the sequence of signal strength local time series feature vectors using the following correlation formula; wherein the correlation formula is:

[0044]

[0045] Among them, V1 is the previous signal strength local time series feature vector of any two signal strength local time series feature vectors in the sequence of the signal strength local time series feature vectors, V2 is the next signal strength local time series feature vector of any two signal strength local time series feature vectors in the sequence of the signal strength local time series feature vectors, and are two different linear transformations, is the correlation between the previous signal strength local time series feature vector and the next signal strength local time series feature vector, (·) TRepresents a transpose operation.

[0046] Furthermore, the sequence of the signal strength local time series feature vectors and the relationship topology feature matrix between the signal strength features are passed through a graph neural network model to obtain a relationship topology global signal strength time series feature matrix. Here, the graph neural network model can effectively process graph structure data and extract the complex relationships between the nodes in the graph. Here, the sequence of signal strength local time series feature vectors can be regarded as node information in the graph data structure, while the relationship topology feature matrix between the signal strength features describes the topological relationship between the nodes, thereby utilizing the encoding ability of the graph neural network model to learn a more comprehensive feature expression of signal strength.

[0047] Correspondingly, the fusion unit 140 is used to: pass the sequence of the signal strength local time series feature vectors and the relationship topology feature matrix between the signal strength features through a graph neural network model to obtain the relationship topology global signal strength time series feature matrix.

[0048] It's worth noting that graph neural networks (GNNs) are a type of deep learning model designed to process graph-structured data. Unlike traditional convolutional neural networks (CNNs) and recurrent neural networks (RNNs), which primarily process grid-structured or sequential data, GNNs are specifically designed to process graph data with nodes and edges. GNNs model and analyze graph data by learning the connectivity and features between nodes. They can exchange and update node feature information with that of their neighbors through message passing and aggregation. This allows each node to incorporate information from its neighbors and gradually propagate and update its own feature representation. The core operations of GNNs include node update and graph aggregation. In the node update phase, each node updates based on its own features and those of its neighbors, using a method similar to convolution. In the graph aggregation phase, node features are aggregated to obtain a global feature representation of the entire graph. This allows GNNs to simultaneously consider both local node information and the overall structure of a node, better capturing the characteristics and relationships of graph data. In the fusion unit 140, a graph neural network model is used to input the sequence of local timing feature vectors of signal strength and the relational topological feature matrix between signal strength features, and by learning the connection relationship and feature propagation between nodes, a relational topological global signal strength timing feature matrix is obtained. This process can combine local timing features with global topological relationships, extract more comprehensive and accurate signal strength timing features, and provide richer information for subsequent analysis and application. In short, a graph neural network is a deep learning model for processing graph structured data, which can capture the features and relationships of graph data by learning the connection relationship and feature propagation between nodes. In the fusion unit, the graph neural network model is used to fuse the sequence of local timing feature vectors of signal strength and the relational topological feature matrix between signal strength features to obtain a relational topological global signal strength timing feature matrix.

[0049] Then, the relational topology global signal strength time series feature matrix is passed through a classifier to obtain a classification result, and the classification result is used to indicate whether the power value of the communication module at the current time point needs to be increased.

[0050] Accordingly, if Figure 5 As shown, the power control analysis unit 150 includes: a feature distribution correction subunit 151, used to perform feature distribution correction on the relationship topology global signal strength timing feature matrix to obtain a corrected relationship topology global signal strength timing feature matrix; and a classification subunit 152, used to pass the corrected relationship topology global signal strength timing feature matrix through a classifier to obtain a classification result, and the classification result is used to indicate whether the power value of the communication module at the current time point needs to be increased.

[0051] It should be understood that the power control and analysis unit 150 includes a feature distribution correction subunit 151 and a classification subunit 152. The feature distribution correction subunit 151 is configured to correct the feature distribution of the relational topology global signal strength time series feature matrix to obtain a corrected relational topology global signal strength time series feature matrix. This subunit may employ statistical methods or data processing techniques to adjust the feature distribution to improve the performance of subsequent classification tasks. The corrected feature matrix serves as input to the next classification subunit. The classification subunit 152 processes the corrected relational topology global signal strength time series feature matrix through a classifier to obtain a classification result. This classification result indicates whether the power value of the communication module at the current time point needs to be increased. The classifier in the classification subunit can be various machine learning algorithms, such as support vector machines (SVMs), decision trees, random forests, neural networks, etc. The classification subunit learns a classification model based on the input feature matrix and then uses this model to classify the features at each time point to determine whether the power of the communication module needs to be increased. By combining the feature distribution correction subunit with the classification subunit, the power control analysis unit can correct and classify the temporal characteristics of the global signal strength of the relational topology, thereby determining whether the power of the communication module needs to be increased at the current point in time. This analysis helps adjust the power level of the communication module based on the real-time signal strength characteristics, thereby improving communication quality and efficiency.

[0052] In the technical solution of the present application, each of the local time series feature vectors of the signal strength expresses the temporal correlation characteristics of the signal strength value in the local time domain determined by vector segmentation in the global time domain. Thus, after the sequence of the local time series feature vectors of the signal strength and the relationship topological feature matrix between the signal strength features are passed through the graph neural network model, the feature vectors corresponding to the local time series feature vectors of the relationship topology global signal strength time series feature matrix, such as the row feature vector, are used to express the topological correlation characteristics of the temporal correlation characteristics of the signal strength values in each local time domain under the temporal correlation correlation topology between the local time domains in the global time domain. That is, on the basis of the classification and regression of the relationship topology global signal strength time series feature matrix through the classifier, each row feature vector of the relationship topology global signal strength time series feature matrix conforms to the interpolation topological correlation mixture of the temporal correlation characteristics of the signal strength local time series feature vector for the classification and regression target.

[0053] Therefore, in order to improve the feature topology association enhancement expression effect of each row feature vector of the relational topology global signal strength timing feature matrix based on the expression consistency of the timing correlation characteristics of the signal strength local timing feature vector, for each group of row feature vectors and the signal strength local timing feature vector, the row feature vector is optimized based on the signal strength local timing feature vector.

[0054] Accordingly, in one example, the feature distribution correction subunit 151 is configured to perform feature distribution correction on the relational topology global signal strength time series feature matrix using the following correction formula to obtain the corrected relational topology global signal strength time series feature matrix; wherein the correction formula is:

[0055]

[0056]

[0057] Among them, V1 is the local time series feature vector of the signal strength, V2 is the row feature vector obtained by expanding the relationship topology global signal strength time series feature matrix, v 1max -1 and v 2max -1 Respectively represent the reciprocal of the global maximum value of each row feature vector V2 obtained after the local signal strength time series feature vector V1 and the relationship topology global signal strength time series feature matrix are expanded, I is a unit vector, and V2 ⊙-1 It represents the reciprocal of the position-by-position eigenvalue of each row eigenvector V2 obtained after expanding the global signal strength time series feature matrix of the relationship topology, ⊙ represents the position-by-position point multiplication, represents vector subtraction, represents vector addition, and V′2 is each row feature vector obtained by expanding the modified relational topology global signal strength time series feature matrix.

[0058] Specifically, for the interpolation topological correlation mixing of regression targets in the feature extraction process, based on the idea of interpolation regularization, by unmixing the feature mapping of outlier features, the high-dimensional feature manifold is restored to the manifold geometry based on weak enhancement based on the inductive bias, and a consistent feature enhancement mapping of interpolation samples and interpolation predictions based on feature extraction is realized, so as to obtain the feature enhancement effect based on topological correlation while maintaining the feature consistency of each row feature vector of the relational topology global signal strength time series feature matrix relative to the signal strength local time series feature vector, thereby improving the feature topological correlation enhancement expression effect based on the expression consistency between each group of corresponding signal strength local time series feature vectors and the row feature vector of the relational topology global signal strength time series feature matrix, so as to improve the accuracy of the classification results obtained by the classifier of the relational topology global signal strength time series feature matrix.

[0059] Furthermore, the classification subunit 152 is further used to: expand the corrected relational topology global signal strength time series feature matrix into a corrected classification feature vector according to a row vector or a column vector; use the fully connected layer of the classifier to fully connect encode the corrected classification feature vector to obtain an encoded classification feature vector; and input the encoded classification feature vector into the Softmax classification function of the classifier to obtain the classification result.

[0060] That is, in the technical solution of the present application, the label of the classifier includes that the power value of the communication module at the current time point needs to be increased (first label), and that the power value of the communication module at the current time point does not need to be increased (second label), wherein the classifier determines to which classification label the revised relational topology global signal strength time series feature matrix belongs through a soft maximum function. It is worth noting that the first label p1 and the second label p2 here do not contain artificially set concepts. In fact, during the training process, the computer model does not have the concept of "whether the power value of the communication module at the current time point needs to be increased". It only has two classification labels and the probability of the output feature under these two classification labels, that is, the sum of p1 and p2 is one. Therefore, the classification result of whether the power value of the communication module at the current time point needs to be increased is actually converted into a two-class class probability distribution that conforms to natural laws through the classification label. In essence, what is used is the physical meaning of the natural probability distribution of the label, rather than the linguistic text meaning of "whether the power value of the communication module at the current time point needs to be increased".

[0061] It should be understood that the role of a classifier is to use given categories and known training data to learn classification rules and classifiers, and then classify (or predict) unknown data. Logistic regression and SVM are commonly used to solve binary classification problems. For multi-class classification problems, logistic regression or SVM can also be used, but multiple binary classifications are required to form a multi-class classification. However, this is prone to errors and is inefficient. A commonly used multi-classification method is the Softmax classification function.

[0062] In summary, the power-adjustable wireless communication module 100 according to the embodiment of the present application is described, which can implement dynamic adjustment of the power value of the wireless communication module.

[0063] As described above, the power-adjustable wireless communication module 100 according to the embodiments of the present application can be implemented in various terminal devices, such as a server equipped with a power-adjustable wireless communication control algorithm. In one example, the power-adjustable wireless communication module 100 can be integrated into the terminal device as a software module and / or a hardware module. For example, the power-adjustable wireless communication module 100 can be a software module in the operating system of the terminal device, or can be an application developed for the terminal device; of course, the power-adjustable wireless communication module 100 can also be one of the many hardware modules of the terminal device.

[0064] Alternatively, in another example, the power-adjustable wireless communication module 100 and the terminal device may be separate devices, and the power-adjustable wireless communication module 100 may be connected to the terminal device via a wired and / or wireless network and transmit interactive information in accordance with an agreed data format.

[0065] Figure 6 The flowchart of the power-adjustable wireless communication control method according to an embodiment of the present application is shown. Figure 7 Schematic diagram showing the system architecture of the power adjustable wireless communication control method according to an embodiment of the present application. Figure 6 and Figure 7As shown, according to the power-adjustable wireless communication control method of an embodiment of the present application, it includes: S110, obtaining the signal strength values of the received wireless signal at multiple predetermined time points within a predetermined time period; S120, extracting the signal strength local characteristics of the signal strength values at the multiple predetermined time points to obtain a sequence of signal strength local time series feature vectors; S130, extracting the topological relationship between the signal strength characteristics of the sequence of the signal strength local time series feature vectors to obtain a signal strength feature relationship topological feature matrix; S140, fusing the sequence of the signal strength local time series feature vectors and the signal strength feature relationship topological feature matrix to obtain a relationship topology global signal strength time series feature matrix; and, S150, based on the relationship topology global signal strength time series feature matrix, determining whether the power value of the communication module at the current time point needs to be increased.

[0066] Here, those skilled in the art will appreciate that the specific operations of each step in the above power adjustable wireless communication control method have been described in detail above. Figures 1 to 5 The power adjustable wireless communication module has been described in detail, and therefore, its repeated description will be omitted.

[0067] Figure 8 FIG. 1 shows an application scenario diagram of a power adjustable wireless communication module according to an embodiment of the present application. Figure 8 As shown, in this application scenario, first, the signal strength values of the received wireless signal at multiple predetermined time points within a predetermined time period are obtained (for example, Figure 8 Then, the signal strength values at the plurality of predetermined time points are input to a server (eg, Figure 8 In S) shown in , the server can use the power adjustable wireless communication control algorithm to process the signal strength values of the multiple predetermined time points to obtain a classification result indicating whether the power value of the communication module at the current time point needs to be increased.

[0068] The flow charts and block diagrams in the accompanying drawings show the possible architecture, functions and operations of the system, method and computer program product according to multiple embodiments of the present application. In this regard, each box in the flow chart or block diagram can represent a part of a module, program segment or instruction, and the part of the module, program segment or instruction includes one or more executable instructions for realizing the logical function of the specification. In some alternative implementations, the functions marked in the box can also occur in a sequence different from that marked in the accompanying drawings. For example, two continuous boxes can actually be executed substantially in parallel, and they can sometimes be executed in the opposite order, depending on the functions involved. It should also be noted that each box in the block diagram and / or flow chart, and the combination of the boxes in the block diagram and / or flow chart can be implemented by a special hardware-based system that performs the function or action of the specification, or can be implemented by a combination of special hardware and computer instructions.

[0069] The embodiments of the present application have been described above. The above description is illustrative and not exhaustive, and is not limited to the disclosed embodiments. Many modifications and variations will be apparent to those skilled in the art without departing from the scope and spirit of the described embodiments. The terminology used herein is selected to best explain the principles of the embodiments, their practical applications, or improvements to the technology in the market, or to enable other persons skilled in the art to understand the embodiments disclosed herein.

Claims

1. A power adjustable wireless communication module, characterized in that: include: a signal strength value acquiring unit, configured to acquire signal strength values of a received wireless signal at a plurality of predetermined time points within a predetermined time period; a local feature extraction unit, configured to extract signal strength local features of the signal strength values at the plurality of predetermined time points to obtain a sequence of signal strength local time series feature vectors; A topological relationship extraction unit, configured to extract the topological relationship between signal strength features between the sequences of the signal strength local time series feature vectors to obtain a topological feature matrix of the relationship between signal strength features; a fusion unit, configured to fuse the sequence of the signal strength local time series feature vectors and the relationship topology feature matrix between the signal strength features to obtain a relationship topology global signal strength time series feature matrix; as well as A power control analysis unit, configured to determine whether a power value of the communication module at a current time point needs to be increased based on the relationship topology global signal strength time series characteristic matrix; Wherein, the local feature extraction unit includes: a data preprocessing subunit, configured to perform data preprocessing on the signal strength values at the plurality of predetermined time points to obtain a sequence of signal strength local time series input vectors; and a feature extraction subunit, configured to perform feature extraction on the sequence of the signal strength local time series input vectors using a deep learning network model to obtain a sequence of the signal strength local time series feature vectors; Wherein, the data preprocessing subunit includes: a vectorization secondary subunit, configured to arrange the signal strength values of the plurality of predetermined time points into a signal strength time series input vector according to a time dimension; and The vector segmentation secondary subunit is used to perform vector segmentation on the signal strength time series input vector to obtain a sequence of the signal strength local time series input vectors.

2. The power adjustable wireless communication module according to claim 1, characterized in that: The deep learning network model is a signal strength temporal feature extractor based on a one-dimensional convolutional layer; The signal strength temporal feature extractor based on the one-dimensional convolution layer includes an input layer, a one-dimensional convolution layer, an activation layer, a pooling layer and an output layer.

3. The power adjustable wireless communication module according to claim 2, characterized in that: The feature extraction subunit is used to: The sequence of the signal strength local time series input vectors is passed through the signal strength time series feature extractor based on the one-dimensional convolution layer to obtain the sequence of the signal strength local time series feature vectors.

4. The power adjustable wireless communication module according to claim 3, characterized in that: The topological relationship extraction unit includes: a correlation calculation subunit, configured to calculate the correlation between any two signal strength local time series feature vectors in the sequence of signal strength local time series feature vectors to obtain a signal strength feature relationship matrix; and The relationship topology feature extraction subunit is used to pass the relationship matrix between the signal strength features through a relationship topology feature extractor based on a convolutional neural network model to obtain the relationship topology feature matrix between the signal strength features.

5. The power adjustable wireless communication module according to claim 4, characterized in that: The correlation calculation subunit is used to: The correlation between any two signal strength local time series feature vectors in the sequence of signal strength local time series feature vectors is calculated using the following correlation formula; Wherein, the correlation formula is: in, is the previous signal strength local time series feature vector of any two signal strength local time series feature vectors in the sequence of signal strength local time series feature vectors, is the next signal strength local time series feature vector between any two signal strength local time series feature vectors in the sequence of signal strength local time series feature vectors, and are two different linear transformations, is the correlation between the previous signal strength local time series feature vector and the next signal strength local time series feature vector, Represents a transpose operation.

6. The power adjustable wireless communication module according to claim 5, characterized in that: The fusion unit is used to: The sequence of the signal strength local time series feature vectors and the relationship topology feature matrix between the signal strength features are passed through a graph neural network model to obtain the relationship topology global signal strength time series feature matrix.

7. The power adjustable wireless communication module according to claim 6, characterized in that: The power control analysis unit includes: a feature distribution correction subunit, configured to perform feature distribution correction on the relational topology global signal strength time series feature matrix to obtain a corrected relational topology global signal strength time series feature matrix; and The classification subunit is used to pass the modified relational topology global signal strength time series feature matrix through a classifier to obtain a classification result, and the classification result is used to indicate whether the power value of the communication module at the current time point needs to be increased.

Images