Narrowband internet of things device battery availability prediction method and apparatus

Abstract

The application discloses a kind of narrowband internet of things equipment battery usability prediction method and device, method includes: constructing input feature data;Input feature data includes first input feature data and second input feature data;First input feature data is constructed according to the equipment running data of obtained narrowband internet of things equipment and battery basic data, and second input feature is calculated and generated by battery basic data;According to input feature data and the corresponding output result obtained by training deep neural network, determine the usability of battery, to notify whether to replace battery.Compared with shallow neural network, the application utilizes deep neural network, learns first input feature data such as narrowband internet of things equipment running data and battery basic data and the second input feature data calculated and generated, to predict the usability of battery, greatly improve the accuracy and prediction accuracy of prediction, improve the on-line rate of narrowband internet of things equipment.

Description

Technical Field

[0001] This invention relates to the field of mobile communication technology, and more specifically to a method and apparatus for predicting battery availability in narrowband Internet of Things (IoT) devices. Background Technology

[0002] NB-IoT (Narrow Band Internet of Things) devices are generally passive devices, meaning they are not directly powered by a power source. Battery availability is a crucial indicator for NB-IoT devices; if the battery fails, the NB-IoT device will be disconnected from the network and unable to provide service. Predicting battery life and replacing it in time before it fails is essential to ensure normal communication for NB-IoT devices.

[0003] Current technologies for predicting battery failure probabilities primarily rely on shallow machine learning algorithms such as logistic regression, random forests, and linear regression, or on manually setting feature weights. This leads to the inability to replace batteries on-site only after they fail, causing NB-IoT devices to go offline and become unusable. Even in the context of big data, shallow machine learning algorithms struggle to achieve higher accuracy, significantly impacting the on-network availability of NB-IoT devices. Summary of the Invention

[0004] In view of the above problems, the present invention is proposed to provide a method and apparatus for predicting battery availability of narrowband Internet of Things (IoT) devices that overcomes or at least partially solves the above problems.

[0005] According to one aspect of the present invention, a method for predicting battery availability of narrowband Internet of Things (IoT) devices is provided, comprising:

[0006] Construct input feature data; the input feature data includes first input feature data and second input feature data; the first input feature data is constructed based on the acquired device operation data of the narrowband IoT device and the basic battery data, and the second input feature data is calculated and generated from the basic battery data;

[0007] Based on the input feature data and the trained deep neural network, the corresponding output result is obtained;

[0008] The battery availability is determined based on the output results to inform whether the battery needs to be replaced.

[0009] According to another aspect of the present invention, a narrowband Internet of Things (IoT) device battery availability prediction apparatus is provided, comprising:

[0010] The module is suitable for constructing input feature data; the input feature data includes first input feature data and second input feature data; the first input feature data is constructed based on the acquired device operation data of the narrowband IoT device and the basic battery data, and the second input feature data is calculated and generated from the basic battery data;

[0011] The prediction module is suitable for obtaining the corresponding output results based on the input feature data and the trained deep neural network;

[0012] The determination module is adapted to determine battery availability based on the output results, so as to notify whether the battery should be replaced.

[0013] According to another aspect of the present invention, an electronic device is provided, comprising: a processor, a memory, a communication interface, and a communication bus, wherein the processor, the memory, and the communication interface communicate with each other via the communication bus;

[0014] The memory is used to store at least one executable instruction, which causes the processor to perform the operation corresponding to the above-described narrowband IoT device battery availability prediction method.

[0015] According to another aspect of the present invention, a computer storage medium is provided, the storage medium storing at least one executable instruction that causes a processor to perform an operation corresponding to the above-described narrowband Internet of Things device battery availability prediction method.

[0016] According to the method and apparatus for predicting battery availability of narrowband IoT devices of the present invention, compared with shallow neural networks, the present invention utilizes deep neural networks to predict battery availability by learning from first input feature data such as operating data and basic battery data of narrowband IoT devices and second input feature data generated by calculation, which greatly improves the accuracy and precision of prediction and increases the on-network rate of narrowband IoT devices.

[0017] The above description is merely an overview of the technical solution of the present invention. In order to better understand the technical means of the present invention and to implement it in accordance with the contents of the specification, and in order to make the above and other objects, features and advantages of the present invention more apparent and understandable, specific embodiments of the present invention are described below. Attached Figure Description

[0018] Various other advantages and benefits will become apparent to those skilled in the art upon reading the following detailed description of preferred embodiments. The accompanying drawings are for illustrative purposes only and are not intended to limit the invention. Furthermore, the same reference numerals denote the same parts throughout the drawings. In the drawings:

[0019] Figure 1A flowchart of a method for predicting battery availability of narrowband Internet of Things (IoT) devices according to an embodiment of the present invention is shown;

[0020] Figure 2 A schematic diagram of the deep neural network framework structure is shown;

[0021] Figure 3 A functional block diagram of a narrowband Internet of Things (IoT) device battery availability prediction apparatus according to an embodiment of the present invention is shown.

[0022] Figure 4 A schematic diagram of the structure of an electronic device according to an embodiment of the present invention is shown. Detailed Implementation

[0023] Exemplary embodiments of the present disclosure will now be described in more detail with reference to the accompanying drawings. While exemplary embodiments of the present disclosure are shown in the drawings, it should be understood that the present disclosure may be implemented in various forms and should not be limited to the embodiments set forth herein. Rather, these embodiments are provided so that this disclosure will be thorough and complete, and will fully convey the scope of the disclosure to those skilled in the art.

[0024] Figure 1 A flowchart illustrating a method for predicting battery availability in narrowband Internet of Things (IoT) devices according to an embodiment of the present invention is shown. Figure 1 As shown, the method for predicting battery availability for narrowband IoT devices specifically includes the following steps:

[0025] Step S101: Construct input feature data.

[0026] Input feature data is used to predict the battery availability of NB-IoT devices. The input feature data includes a first input feature data and a second input feature data. The first input feature data is constructed based on the acquired device operation data and basic battery data of the narrowband IoT device, while the second input feature data is calculated and generated from the basic battery data. Specifically, data such as NB-IoT device operation data, service characteristic data, battery voltage, and battery replacement records for a specified time period, such as the past year, can be obtained from the NB-IoT IoT management center and NB-IoT maintenance center. New data is then calculated and generated based on this directly obtainable data. The first input feature data is constructed based on various device operation data and basic battery data of the directly acquired NB-IoT device, including the following data: T3412 (TAU cycle timer), T3324 (PSM ActiveTimer, Power Save Mode activation timer), eDRX (Extended idle-mode Discontinuous Reception cycle), PTW (Paging Time Window), average daily data interaction count, average single data interaction amount, RSRP (Reference Signal Receiving Power), SINR (Signal-to-Interference and Noise Ratio), end-to-end latency, battery voltage variation, terminal brand, battery brand, battery capacity, and 13 other features, constructing a multi-dimensional first input feature data. The second input feature data includes, for example, the NB-IoT device's battery usage time and the NB-IoT device's previous battery usage time. Wherein, the battery usage time of the NB-IoT device = current time - start time of battery usage of the NB-IoT device; the battery usage time of the previous NB-IoT device = end time of battery usage of the previous NB-IoT device - start time of battery usage of the previous NB-IoT device.

[0027] The first and second input feature data together constitute the input feature data, resulting in multi-dimensional input feature data. In the example above, the input feature data contains 15 dimensions of feature data. In specific implementations, the specific dimensions can be determined according to the implementation situation, and no limitation is made here.

[0028] Step S102: Based on the input feature data and the trained deep neural network, obtain the corresponding output result.

[0029] Deep neural networks are pre-trained offline using TensorFlow models. A deep neural network specifically consists of an input layer, hidden layers, and an output layer. The input layer's data dimension corresponds to the feature dimension of the input data, and the output layer's output characterizes whether the battery is prone to damage. Numerous neurons in the input layer receive a large amount of nonlinear input information. Information in the output layer is transmitted, analyzed, and weighed within the neuron connections to form the output result. Hidden layers are the various layers composed of numerous neurons and connections between the input and output layers. There can be multiple hidden layers, the number determined based on the implementation. The number of neurons in the hidden layers is variable, but a higher number of hidden layers results in more significant nonlinearity and thus greater robustness.

[0030] like Figure 2 As shown, the deep neural network consists of one input layer, eight hidden layers (fully connected), and one output layer. The input layer contains 15 neurons (with the same feature dimension as the input feature data), and the output layer contains one neuron, which is the predicted value. The first and second hidden layers contain 128 neurons, the third and fourth hidden layers contain 64 neurons, the fifth and sixth hidden layers contain 32 neurons, the seventh and eighth hidden layers contain 16 neurons, and so on. The activation function used in the hidden layers is ReLU (Rectified Linear Unit), and the activation function used in the output layer is sigmoid (S-shaped function), mapping the variable to the range of 0-1. The output result indicates whether the battery has a tendency to be damaged.

[0031] The training process of a deep neural network specifically includes: constructing input feature sample data and labeled data. The input feature sample data is constructed according to the description in step S101, and will not be repeated here. The labeled data is set based on whether the narrowband IoT device's battery has been replaced. For example, positive sample labeled data uses 1 to indicate that the NB-IoT device's battery is damaged and has been replaced, while negative sample labeled data uses 0 to indicate that the NB-IoT device's battery has not been replaced, etc. The specific values ​​of the labeled data are set according to the implementation situation and are not limited here.

[0032] After obtaining the input feature sample data and labeled data, the data is preprocessed. The input feature sample data and labeled data are divided into training input sample data, training labeled data, test input sample data, and test labeled data according to a preset ratio. For example, a 7:3 ratio can be used to divide the data into training data and test data. Training data includes training input sample data and training labeled data; test data includes test input sample data and test labeled data. The training data is used to train the deep neural network. After training, the test data is used to verify the accuracy of the training results. Specifically, the function `train_test_split` can be used, such as `train_X, test_X, train_y, test_y = train_test_split(X, y, train_size = 0.7, random_state = 0)`. Here, `train_X` represents the training input sample data, `train_y` represents the training labeled data, `test_X` represents the test input sample data, and `test_y` represents the test labeled data. `train_size = 0.7` means that the training data accounts for 70% and the test data accounts for 30%. `random_state` controls the random state; the specific settings can be adjusted according to the implementation situation. The above is just an example.

[0033] After obtaining the above data, the first specified function of the TensorFlow model, such as `model.fit`, is used to train the deep neural network using the training input sample data, training labeled data, test input sample data, and test labeled data until a specified training epoch threshold is reached. This completes the training of the deep neural network and determines its weights. For example, `model.fit(train_X, train_y, epochs=150, batch_size=10)`, where `epochs` represents the training epochs, specifically specified according to the implementation. As the number of training epochs increases, the training error gradually decreases, and the TensorFlow model gradually converges. Training is complete when the TensorFlow model's convergence change is small and reaches the specified threshold. Specifically, the result can be determined based on the training epochs set during training or the expected threshold of the loss function; no specific limitation is made here. Through training, the deep neural network can autonomously learn the weights. After training, the trained neural network weights can be exported. During prediction, there is no need to retrain the deep neural network; the trained neural network weights can be used directly, saving time and making prediction faster and more efficient.

[0034] During prediction, the second specified function of the TensorFlow model is used to predict the input feature data and obtain the corresponding output result. Specifically, the `predict` function in `keras.model` can be called. The `predict` function is used for the actual prediction, taking the input feature data and generating the output result. As shown in the following code:

[0035] dataframe1 = pandas.read_csv("XX.csv", header = None) --- Saves the input feature data as a CSV file. By reading the CSV file, you can obtain the input feature data contained therein.

[0036] dataset1 = dataframe1.values

[0037] X1 = dataset1[i:, j:k] -- Get the data from row i to the last row, and from column j to column k, which is the input feature data used for prediction; the values ​​of i, j, and k are set according to the implementation.

[0038] predictions = model.predict(X1) -- Prediction complete and output results.

[0039] Predictions indicate whether the battery is prone to damage. For example, predictions=1 means that the battery of the NB-IoT device is predicted to be damaged soon, and predictions=0 means that the battery of the NB-IoT device will not be damaged in the near future.

[0040] Step S103: Determine the battery availability based on the output results to notify whether the battery needs to be replaced.

[0041] Based on the output results, filter out those that indicate a tendency for battery damage, such as those with predictions=1, and identify the corresponding NB-IoT narrowband Internet of Things devices so that the battery can be replaced.

[0042] When notifying, specific methods can include generating a replacement work order online and sending the replacement site to maintenance personnel to notify them to replace the battery on-site, thereby improving the network uptime of NB-IoT narrowband Internet of Things devices.

[0043] According to the battery availability prediction method for narrowband IoT devices provided by the present invention, compared with shallow neural networks, the present invention utilizes deep neural networks to predict battery availability by learning from first input feature data such as operating data and basic battery data of narrowband IoT devices and second input feature data generated by calculation. This greatly improves the accuracy and precision of prediction, and increases the network connectivity rate of narrowband IoT devices.

[0044] Figure 3 A functional block diagram of a narrowband Internet of Things (IoT) device battery availability prediction apparatus according to an embodiment of the present invention is shown. Figure 3 As shown, the narrowband IoT device battery availability prediction device includes the following modules:

[0045] The construction module 310 is suitable for constructing input feature data; the input feature data includes first input feature data and second input feature data; the first input feature data is constructed based on the acquired device operation data of the narrowband IoT device and the basic battery data, and the second input feature data is calculated and generated from the basic battery data;

[0046] The prediction module 320 is adapted to obtain the corresponding output result based on the input feature data and the trained deep neural network;

[0047] The determination module 330 is adapted to determine the availability of the battery based on the output results, so as to notify whether the battery should be replaced.

[0048] Alternatively, deep neural networks can be trained offline using Tensorflow models.

[0049] Optionally, the deep neural network specifically includes: an input layer, a hidden layer, and an output layer; the data dimension of the input layer is the feature dimension of the input feature data; the output result of the output layer is used to characterize whether the device battery is prone to damage.

[0050] Optionally, the training process of a deep neural network is as follows:

[0051] Construct input feature sample data and labeled data; the labeled data is set according to whether the narrowband IoT device has had its battery replaced.

[0052] The input feature sample data and labeled data are divided into training input sample data, training labeled data, test input sample data and test labeled data according to a preset ratio;

[0053] The deep neural network is trained using the first specified function of the Tensorflow model on the training input sample data, training labeled data, test input sample data, and test labeled data until a specified training epoch threshold is reached, thereby completing the training of the deep neural network and determining the weight values ​​of the deep neural network.

[0054] Optionally, the prediction module 320 is further adapted to:

[0055] The second specified function of the Tensorflow model is used to predict the input feature data and obtain the corresponding output results.

[0056] Optionally, the determining module 330 is further adapted to:

[0057] The system filters out outputs indicating a tendency for device batteries to fail, identifies the corresponding narrowband IoT devices, and notifies them to replace the device batteries.

[0058] Optionally, the determining module 330 is further adapted to:

[0059] Generate a replacement work order to notify the device battery to be replaced.

[0060] The descriptions of the above modules refer to the corresponding descriptions in the method embodiments, and will not be repeated here.

[0061] This application also provides a non-volatile computer storage medium storing at least one executable instruction that can execute the narrowband IoT device battery availability prediction method in any of the above method embodiments.

[0062] Figure 4 The diagram illustrates the structure of an electronic device according to an embodiment of the present invention. The specific embodiments of the present invention do not limit the specific implementation of the electronic device.

[0063] like Figure 4 As shown, the electronic device may include: a processor 402, a communications interface 404, a memory 406, and a communications bus 408.

[0064] in:

[0065] The processor 402, communication interface 404, and memory 406 communicate with each other via communication bus 408.

[0066] Communication interface 404 is used to communicate with other network elements such as clients or other servers.

[0067] The processor 402 is used to execute program 410, which can specifically execute the relevant steps in the above embodiment of the narrowband IoT device battery availability prediction method.

[0068] Specifically, program 410 may include program code that includes computer operation instructions.

[0069] Processor 402 may be a central processing unit (CPU), an application-specific integrated circuit (ASIC), or one or more integrated circuits configured to implement embodiments of the present invention. The electronic device includes one or more processors, which may be processors of the same type, such as one or more CPUs; or they may be processors of different types, such as one or more CPUs and one or more ASICs.

[0070] Memory 406 is used to store program 410. Memory 406 may include high-speed RAM memory, and may also include non-volatile memory, such as at least one disk storage device.

[0071] Specifically, program 410 can be used to cause processor 402 to execute the narrowband IoT device battery availability prediction method in any of the above method embodiments. The specific implementation of each step in program 410 can be found in the corresponding descriptions of the steps and units in the above narrowband IoT device battery availability prediction embodiments, and will not be repeated here. Those skilled in the art will clearly understand that, for the sake of convenience and brevity, the specific working process of the devices and modules described above can be referred to the corresponding process descriptions in the foregoing method embodiments, and will not be repeated here.

[0072] The algorithms and displays provided herein are not inherently related to any particular computer, virtual system, or other device. Various general-purpose systems can also be used in conjunction with the teachings herein. The required structure for constructing such systems is apparent from the above description. Furthermore, this invention is not directed to any particular programming language. It should be understood that the contents of the invention described herein can be implemented using various programming languages, and the above description of specific languages ​​is for the purpose of disclosing the best mode of implementation of the invention.

[0073] Numerous specific details are set forth in the specification provided herein. However, it will be understood that embodiments of the invention may be practiced without these specific details. In some instances, well-known methods, structures, and techniques have not been shown in detail so as not to obscure the understanding of this specification.

[0074] Similarly, it should be understood that, in order to streamline this disclosure and aid in understanding one or more of the various inventive aspects, in the above description of exemplary embodiments of the invention, various features of the invention are sometimes grouped together in a single embodiment, figure, or description thereof. However, this method of disclosure should not be interpreted as reflecting an intention that the claimed invention requires more features than are expressly recited in each claim. Rather, as reflected in the claims, inventive aspects lie in fewer than all features of a single foregoing disclosed embodiment. Therefore, the claims following the detailed description are hereby expressly incorporated into that detailed description, wherein each claim itself is a separate embodiment of the invention.

[0075] Those skilled in the art will understand that modules in the device of the embodiments can be adaptively changed and placed in one or more devices different from that embodiment. Modules, units, or components in the embodiments can be combined into a single module, unit, or component, and further, they can be divided into multiple sub-modules, sub-units, or sub-components. Except where at least some of such features and / or processes or units are mutually exclusive, any combination can be used to combine all features disclosed in this specification (including the accompanying claims, abstract, and drawings) and all processes or units of any method or device so disclosed. Unless expressly stated otherwise, each feature disclosed in this specification (including the accompanying claims, abstract, and drawings) may be replaced by an alternative feature that serves the same, equivalent, or similar purpose.

[0076] Furthermore, those skilled in the art will understand that although some embodiments described herein include certain features but not others included in other embodiments, combinations of features from different embodiments are intended to be within the scope of the invention and form different embodiments. For example, in the claims, any of the claimed embodiments can be used in any combination.

[0077] The various component embodiments of the present invention can be implemented in hardware, or as software modules running on one or more processors, or a combination thereof. Those skilled in the art will understand that microprocessors or digital signal processors (DSPs) can be used in practice to implement some or all of the functions of some or all of the components in the narrowband IoT device battery availability prediction device according to embodiments of the present invention. The present invention can also be implemented as a device or apparatus program (e.g., a computer program and computer program product) for performing part or all of the methods described herein. Such programs implementing the present invention can be stored on a computer-readable medium or can be in the form of one or more signals. Such signals can be downloaded from an Internet website, provided on a carrier signal, or provided in any other form.

[0078] It should be noted that the above embodiments are illustrative of the invention and not restrictive, and that those skilled in the art can devise alternative embodiments without departing from the scope of the appended claims. In the claims, any reference signs placed between parentheses should not be construed as limiting the claims. The word "comprising" does not exclude the presence of elements or steps not listed in the claims. The word "a" or "an" preceding an element does not exclude the presence of a plurality of such elements. The invention can be implemented by means of hardware comprising several different elements and by means of a suitably programmed computer. In the unit claims enumerating several means, several of these means may be embodied by the same item of hardware. The use of the words first, second, and third, etc., does not indicate any order. These words can be interpreted as names.

Claims

1. A method for predicting battery availability in narrowband IoT devices, characterized in that, The methods include: Construct input feature data; the input feature data includes first input feature data and second input feature data; The first input feature data is constructed based on the acquired device operation data and battery basic data of the narrowband IoT device, and the second input feature data is calculated and generated from the battery basic data; the device operation data of the narrowband IoT device includes the following data: TAU cycle timer, power saving mode activation timer, extended idle state discontinuous reception cycle, paging time window, average daily data interaction count, average single data interaction amount, reference signal received power, signal-to-interference-to-noise ratio, and end-to-end latency; Based on the input feature data and the trained deep neural network, a corresponding output result is obtained. The deep neural network is trained offline using a TensorFlow model. A first specified function of the TensorFlow model is used to train the training input sample data, training labeled data, test input sample data, and test labeled data until a specified training epoch threshold is reached to complete the training of the deep neural network and determine its weights. A second specified function of the TensorFlow model is used to predict the input feature data to obtain the corresponding output result. The output result of the output layer is used to characterize whether the battery has a tendency to be damaged. The battery availability is determined based on the output results to notify whether the battery needs to be replaced.

2. The method according to claim 1, characterized in that, The deep neural network specifically includes an input layer, a hidden layer, and an output layer; the data dimension of the input layer is the feature dimension of the input feature data.

3. The method according to claim 2, characterized in that, The training process of the deep neural network is as follows: Construct input feature sample data and labeled data; the labeled data is set according to whether the narrowband IoT device has had its battery replaced. The input feature sample data and labeled data are divided into training input sample data, training labeled data, test input sample data and test labeled data according to a preset ratio; The Tensorflow model is used to train the training input sample data, training labeled data, test input sample data, and test labeled data using a first specified function until a specified training round threshold is reached, thereby completing the training of the deep neural network and determining the weight values ​​of the deep neural network.

4. The method according to claim 1, characterized in that, The step of determining battery availability based on the output result to notify whether to replace the battery further includes: Outputs indicating a tendency for battery damage are filtered out, and the corresponding narrowband IoT devices are identified to notify the user to replace the battery.

5. The method according to claim 4, characterized in that, The step of filtering out output results indicating batteries with a tendency to fail, and determining the corresponding narrowband IoT device for the output results to notify the user to replace the battery, further includes: A replacement work order is generated to notify the battery to be replaced.

6. A battery availability prediction device for narrowband Internet of Things (IoT) devices, characterized in that, The device includes: The construction module is suitable for constructing input feature data; the input feature data includes first input feature data and second input feature data. The first input feature data is constructed based on the acquired device operation data and battery basic data of the narrowband IoT device, and the second input feature data is calculated and generated from the battery basic data; the device operation data of the narrowband IoT device includes the following data: TAU cycle timer, power saving mode activation timer, extended idle state discontinuous reception cycle, paging time window, average daily data interaction count, average single data interaction amount, reference signal received power, signal-to-interference-to-noise ratio, and end-to-end latency; The prediction module is adapted to obtain a corresponding output result based on the input feature data and the trained deep neural network. The deep neural network is trained offline using a Tensorflow model. A first specified function of the Tensorflow model is used to train the training input sample data, training labeled data, test input sample data, and test labeled data until a specified training epoch threshold is reached to complete the training of the deep neural network and determine its weights. A second specified function of the Tensorflow model is used to predict the input feature data to obtain a corresponding output result. The output result indicates whether the battery has a tendency to be damaged. The determination module is adapted to determine the availability of the battery based on the output results, so as to notify whether the battery should be replaced.

7. An electronic device, comprising: The processor, memory, communication interface, and communication bus are provided, wherein the processor, memory, and communication interface communicate with each other via the communication bus. The memory is used to store at least one executable instruction that causes the processor to perform the operation corresponding to the narrowband IoT device battery availability prediction method as described in any one of claims 1-5.

8. A computer storage medium storing at least one executable instruction that causes a processor to perform an operation corresponding to the narrowband Internet of Things device battery availability prediction method as described in any one of claims 1-5.

Images