Internet of Things system and device based on fifth generation mobile communication technology, and medium

Abstract

The invention provides an Internet of Things system, device and medium based on the fifth generation mobile communication technology, and relates to the technical field of Internet of Things, the system comprises a 5G connection module and an anomaly detection module; the 5G connection module is connected with terminal equipment and acquires equipment data of the terminal equipment; the anomaly detection module is used for generating predicted reporting time data of the next moment according to historical reporting time data and the environment data, and generating a data timeout anomaly judgment result according to the predicted reporting time data and the actual reporting time data; the analysis module is used for analyzing a statistical feature value of each window of the time sequence data and a weight corresponding to the statistical feature value, determining an extreme value anomaly probability according to the time sequence data and the statistical feature value, and generating an extreme value anomaly judgment result according to the statistical feature value, the weight and the extreme value anomaly probability; the method is also used for generating an extreme value proportion abnormity judgment result. The abnormal connection condition can be analyzed.

Description

Technical Field

[0001] The present invention relates to the technical field of the Internet of Things, and in particular, to an Internet of Things system, device, and medium based on the fifth-generation mobile communication technology. Background Art

[0002] An Internet of Things system refers to a series of software solutions designed to achieve the interconnection and interoperability between objects and the connection between objects and the Internet. The Internet of Things system allows devices, sensors, electrical appliances, etc. to exchange data through a network and can perform specific tasks or provide services based on the collected data.

[0003] In the related art, most Internet of Things systems can only view the normal connection status of devices when connecting devices and cannot analyze the abnormal connection conditions of devices. Summary of the Invention

[0004] The problem solved by the present invention is how to analyze the abnormal connection conditions of Internet of Things devices.

[0005] To solve the above problems, the present invention provides an Internet of Things system, device, and medium based on the fifth-generation mobile communication technology.

[0006] In a first aspect, the present invention provides an Internet of Things system based on the fifth-generation mobile communication technology, including a 5G connection module and an anomaly detection module; The 5G connection module is based on the fifth-generation mobile communication technology, connects to a terminal device, and obtains device data of the terminal device, where the device data includes historical reporting time data, actual reporting time data, environmental data, and time series data; The anomaly detection module includes a data timeout unit and an extreme value unit. The data timeout unit is used to generate predicted reporting time data for the next moment according to the historical reporting time data and the environmental data by using a bidirectional LSTM model, and generate a data timeout anomaly judgment result according to the predicted reporting time data and the actual reporting time data; The extreme value unit is used to analyze the statistical feature values of each window of the time series data and the weights corresponding to the statistical feature values by using a multi-scale sliding window, and determine the extreme value anomaly probability by using a hidden Markov model according to the time series data and the statistical feature values. According to the statistical feature values, the weights, and the extreme value anomaly probability, generate an extreme value comprehensive anomaly score, and generate an extreme value anomaly judgment result according to the extreme value comprehensive anomaly score, where the statistical feature values include the mean and variance; The extreme value unit is further configured to generate a fitted mean and a fitted variance according to the time series data by using the maximum likelihood estimation method, and determine an extreme value theory proportion by using a cumulative distribution function based on the fitted mean and the fitted variance, determine an actual extreme value proportion according to the mean and the variance, and generate an extreme value proportion anomaly judgment result according to the extreme value theory proportion and the actual extreme value proportion.

[0007] Optionally, the step of generating the predicted reporting time data at the next moment according to the historical reporting time data and the environmental data by using the bidirectional LSTM model includes: Using the bidirectional LSTM model, generate preliminary prediction data according to the historical reporting time data and a preliminary prediction formula, where the preliminary prediction formula includes: ; ; where h t is the output of the hidden layer of the bidirectional LSTM model, is the historical reporting time data, n is n moments, t is t moments, BiLSTM is the calculation rule of the hidden layer of the bidirectional LSTM model, is the trainable parameter of the bidirectional LSTM model, including a weight matrix and a bias term, is the preliminary prediction data, is the weight matrix that maps the output of the hidden layer of the bidirectional LSTM model to the prediction time, is the bias term of the prediction time; Using an attention mechanism, weight and sum the environmental data according to an importance weight formula to generate environmental impact degree data, where the importance weight formula includes: ; ; where, is the importance weight of the environmental data, softmax is the attention calculation rule, is the environmental data, is the weight matrix that maps the environmental data to the attention weight space, is the bias term of the attention mechanism, is the environmental impact degree data, is the importance weight of the i-th environmental data, is the i-th environmental data; Using a fusion formula to fuse the preliminary prediction data and the environmental impact degree data to generate the predicted reporting time data, where the fusion formula includes: T = f ( T t+1 , ct ; θ f ) Among them, T is the predicted reporting time data, and f is the fusion rule, θ f are the trainable parameters for fusion, including the weight matrix and the bias term.

[0008] Optionally, generating a data timeout exception judgment result according to the predicted reporting time data and the actual reporting time data includes: Determining a dynamic timeout threshold according to the predicted reporting time data and the environmental data by using a dynamic timeout threshold formula, where the dynamic timeout threshold formula includes: ; Among them, is the dynamic timeout threshold, is the predicted reporting time data, is the safety buffer coefficient, is the standard deviation of the historical reporting time data, is the weight coefficient, is the degree of change of the environmental data; Generating the data timeout exception judgment result according to the actual reporting time data and the dynamic timeout threshold.

[0009] Optionally, analyzing the statistical eigenvalue of each window of the time series data and the weight corresponding to the statistical eigenvalue by using a multi-scale sliding window includes: Analyzing the statistical eigenvalue of each window of the time series data and the uncertainty entropy value of the statistical eigenvalue of each window by using the multi-scale sliding window; Determining the weight according to the uncertainty entropy value, where the uncertainty entropy value is inversely proportional to the corresponding weight.

[0010] Optionally, generating an extreme value comprehensive anomaly score according to the statistical eigenvalue, the weight, and the extreme value anomaly probability includes: Generating the extreme value comprehensive anomaly score according to the statistical eigenvalue, the weight, and the extreme value anomaly probability by using a comprehensive formula, where the comprehensive formula includes: ; Among them, is the extreme value comprehensive anomaly score, is the weight, is the time series data, is the mean value, the variance, is the adjustment parameter, is the extreme value anomaly probability.

[0011] Optionally, the terminal device includes a camera, the device data includes historical video data and current video data corresponding to the camera, and the anomaly detection module further includes an occlusion detection unit; The occlusion detection unit is used to extract the unoccluded background features of the historical video data by using a deep learning model, and based on the unoccluded background features, construct an unoccluded virtual model by using an adversarial network, and input the current video data into the trained unoccluded virtual model to generate an occlusion detection result.

[0012] Optionally, the anomaly detection module further includes a color screen detection unit; The color screen detection unit is used to convert the current video data into the Lab color space, and analyze the video data in the Lab color space by using wavelet transform to generate a color screen detection result.

[0013] Optionally, the anomaly detection module further includes a blur detection unit; The blur detection unit is used to perform frequency domain analysis on the current video data by using fast Fourier transform to generate a blur detection result.

[0014] In a second aspect, the present invention provides an electronic device, including a memory and a processor; The memory is used to store a computer program; The processor is used to, when executing the computer program, implement the following steps: Based on the fifth-generation mobile communication technology, connect to the terminal device and obtain the device data of the terminal device, where the device data includes historical reporting time data, actual reporting time data, environmental data, and time series data; Adopt a bidirectional LSTM model to generate predicted reporting time data for the next moment according to the historical reporting time data and the environmental data, and generate a data timeout anomaly judgment result according to the predicted reporting time data and the actual reporting time data; Adopt a multi-scale sliding window to analyze the statistical feature values of each window of the time series data and the weights corresponding to the statistical feature values, and according to the time series data and the statistical feature values, use a hidden Markov model to determine the extreme value anomaly probability, and generate an extreme value comprehensive anomaly score according to the statistical feature values, the weights, and the extreme value anomaly probability, and generate an extreme value anomaly judgment result according to the extreme value comprehensive anomaly score, where the statistical feature values include the mean value and the variance; According to the time series data, the maximum likelihood estimation method is used to generate the fitted mean and the fitted variance. Based on the fitted mean and the fitted variance, the cumulative distribution function is used to determine the extreme value theory proportion. According to the mean and the variance, the actual extreme value proportion is determined. According to the extreme value theory proportion and the actual extreme value proportion, an extreme value proportion anomaly judgment result is generated.

[0015] In a third aspect, the present invention provides a computer-readable storage medium, on which a computer program is stored. When the computer program is executed by a processor, the following steps are implemented: Based on the fifth-generation mobile communication technology, connect to the terminal device and obtain the device data of the terminal device, where the device data includes historical reporting time data, actual reporting time data, environmental data, and time series data; Adopt a bidirectional LSTM model to generate predicted reporting time data for the next moment according to the historical reporting time data and the environmental data, and generate a data timeout anomaly judgment result according to the predicted reporting time data and the actual reporting time data; Adopt a multi-scale sliding window to analyze the statistical feature values of each window of the time series data and the weights corresponding to the statistical feature values. According to the time series data and the statistical feature values, use the hidden Markov model to determine the extreme value anomaly probability. According to the statistical feature values, the weights, and the extreme value anomaly probability, generate an extreme value comprehensive anomaly score. According to the extreme value comprehensive anomaly score, generate an extreme value anomaly judgment result, where the statistical feature values include the mean and the variance; According to the time series data, the maximum likelihood estimation method is used to generate the fitted mean and the fitted variance. Based on the fitted mean and the fitted variance, the cumulative distribution function is used to determine the extreme value theory proportion. According to the mean and the variance, the actual extreme value proportion is determined. According to the extreme value theory proportion and the actual extreme value proportion, an extreme value proportion anomaly judgment result is generated.

[0016] The beneficial effects of the Internet of Things system, device, and medium based on the fifth-generation mobile communication technology of the present invention are: The present invention uses a 5G connection module based on the fifth-generation mobile communication technology to connect to terminal devices, providing a stable connection channel and a fast data transmission channel, thereby stably and quickly obtaining the device data of the terminal devices. Through the data timeout unit of the anomaly detection module, a bidirectional LSTM model is used for forward and backward propagation to capture the long-term dependence relationship of the time series of historical reporting time data, and considering the influence of environmental data, accurate predicted reporting time data can be generated, that is, the normal reporting time of the terminal devices. Then, the actual reported time data is compared with the predicted reporting time data to accurately determine whether the data upload of the terminal devices connected to the Internet of Things is normal. Further, through the extreme value unit of the anomaly detection module, a multi-scale sliding window is used to analyze the statistical eigenvalue of each window of the time series data and the corresponding weight of the statistical eigenvalue. Different statistical eigenvalues at different time scales can be captured through multiple window sizes, that is, the mean and variance, so as to comprehensively describe the distribution characteristics of the data. And through the corresponding weight, the model pays more attention to the windows with high certainty. Then, according to the time series data and the statistical eigenvalues, a hidden Markov model is used to determine the extreme value anomaly probability of the extreme values in the time series data, so as to facilitate subsequent judgment on whether the extreme values in the current data belong to the abnormal state. Finally, considering the statistical eigenvalues, weights, and extreme value anomaly probability comprehensively, it is reasonably judged whether the extreme values of the terminal devices connected to the Internet of Things are abnormal. The extreme value unit also uses the maximum likelihood estimation method and the cumulative distribution function to determine the proportion of extreme values in the theoretical case, that is, the theoretical proportion of extreme values. Then, the theoretical proportion of extreme values is compared with the actual proportion of extreme values to determine whether the proportion of extreme values of the terminal devices connected to the Internet of Things is abnormal. The present invention analyzes and judges the data upload time, extreme values, and the proportion of extreme values of the terminal devices connected to the Internet of Things, and can further analyze whether the connection of the terminal devices is abnormal. Description of the Drawings

[0017] Figure 1 FIG. is a schematic structural diagram of an Internet of Things system based on the fifth-generation mobile communication technology provided by an embodiment of the present invention; Figure 2 FIG. is a schematic structural diagram of an electronic device provided by an embodiment of the present invention. Detailed Embodiments

[0018] To make the above objects, features, and advantages of the present invention more obvious and understandable, the following detailed description of the specific embodiments of the present invention is provided in conjunction with the accompanying drawings. Although some embodiments of the present invention are shown in the drawings, it should be understood that the present invention can be implemented in various forms and should not be construed as limited to the embodiments described herein. On the contrary, these embodiments are provided to more thoroughly and completely understand the present invention. It should be understood that the drawings and embodiments of the present invention are only for exemplary purposes and are not used to limit the protection scope of the present invention.

[0019] It should be understood that the various steps recited in the method embodiments of the present invention can be executed in different orders and / or executed in parallel. In addition, the method embodiments may include additional steps and / or omit the steps shown. The scope of the present invention is not limited in this regard.

[0020] As used herein, the term "comprising" and its variations are open-ended, that is, "including but not limited to"; the term "based on" is "at least partially based on"; the term "one embodiment" means "at least one embodiment"; the term "another embodiment" means "at least one additional embodiment"; the term "some embodiments" means "at least some embodiments"; the term "optionally" means "optional embodiment". The relevant definitions of other terms will be given in the following description. It should be noted that the concepts such as "first" and "second" mentioned in the present invention are only used to distinguish different devices, modules or units, and are not used to limit the order or interdependence of the functions performed by these devices, modules or units.

[0021] It should be noted that the modifications of "one" and "plural" mentioned in the present invention are illustrative rather than restrictive. Those skilled in the art should understand that unless otherwise clearly stated in the context, it should be understood as "one or more".

[0022] The names of the messages or information exchanged between multiple devices in the embodiments of the present invention are only for illustrative purposes and are not used to limit the scope of these messages or information.

[0023] In view of the problems existing in the above related technologies, this embodiment provides an Internet of Things system, device and medium based on the fifth-generation mobile communication technology.

[0024] As Figure 1 shown, an Internet of Things system based on the fifth-generation mobile communication technology provided by an embodiment of the present invention includes a 5G connection module and an anomaly detection module; The 5G connection module is based on the fifth-generation mobile communication technology, connects to the terminal device, and obtains the device data of the terminal device, where the device data includes historical reporting time data, actual reporting time data, environmental data, and time series data.

[0025] Specifically, the 5G connection module is based on the fifth-generation mobile communication technology to connect to terminal devices, providing a stable connection channel and a fast data transmission channel, so as to stably and quickly obtain the device data of the terminal devices. The device data includes historical reporting time data, actual reporting time data, environmental data, and time series data. The terminal devices may include mobile phones, computers, sensors, monitors, cameras, etc. The environmental data refers to the data of the environment where the terminal device is located. For example, network latency, packet loss rate, and device power. The time series data refers to the data related to the time dimension. For example, sensor data, device performance data, behavior data, and timestamp data.

[0026] The anomaly detection module includes a data timeout unit and an extreme value unit. The data timeout unit is used to adopt a bidirectional LSTM model to generate predicted reporting time data for the next moment according to the historical reporting time data and the environmental data, and generate a data timeout anomaly judgment result according to the predicted reporting time data and the actual reporting time data.

[0027] Specifically, the bidirectional LSTM model, that is, the BiLSTM model, combines the forward LSTM and the reverse LSTM, can capture the context information of the past and future in the time series data at the same time, and introduces the influence of environmental variables (such as network latency, packet loss rate, etc.) on the prediction result, improving the adaptability of the model to complex scenarios, obtaining the predicted reporting time data for the next moment of the historical reporting time data, and comparing the predicted reporting time data with the actual reporting time data. If the predicted reporting time data is different from the actual reporting time data, the data timeout anomaly judgment result is that the data reporting is timed out and the reporting time is abnormal. If the predicted reporting time data is the same as the actual reporting time data, the data timeout anomaly judgment result is that the data reporting is not timed out and the reporting time is normal. Among them, the minimum unit of the predicted reporting time data and the actual reporting time data is seconds, and the difference between the predicted reporting time data and the actual reporting time data is within 1 second. By default, the predicted reporting time data and the actual reporting time data are the same.

[0028] The extreme value unit is used to adopt a multi-scale sliding window to analyze the statistical eigenvalue of each window of the time series data and the weight corresponding to the statistical eigenvalue, and determine the extreme value anomaly probability according to the time series data and the statistical eigenvalue by using a hidden Markov model. Generate an extreme value comprehensive anomaly score according to the statistical eigenvalue, the weight, and the extreme value anomaly probability, and generate an extreme value anomaly judgment result according to the extreme value comprehensive anomaly score. Among them, the statistical eigenvalue includes the mean value and the variance.

[0029] Specifically, the extreme value unit adopts a multi-scale sliding window, captures statistical features such as mean and variance at different time scales through multiple window sizes to comprehensively describe the distribution characteristics of the data. The multi-scale sliding window can also determine the weights corresponding to the statistical feature values to ensure the reasonable utilization of multi-scale information. Then, a Hidden Markov Model (HMM) is used to identify potential abnormal patterns, so as to facilitate subsequent judgment on whether the extreme values in the current data belong to the abnormal state. The HMM can be trained to learn the transition probability matrix and observation probability matrix in the normal state, and the Bayesian formula is used to calculate the probability that the extreme value belongs to the abnormal state to obtain the extreme value abnormal probability. Among them, the Hidden Markov Model (HMM) is a statistical model, and its use for identifying potential abnormal patterns is prior art and will not be elaborated here. Then, by synthesizing the statistical feature values, weights, and extreme value abnormal probability, an extreme value comprehensive abnormal score is generated. If the extreme value comprehensive abnormal score is greater than the preset extreme value threshold, the extreme value abnormal judgment result is that the extreme value is abnormal. If the extreme value comprehensive abnormal score is less than or equal to the preset extreme value threshold, the extreme value abnormal judgment result is that the extreme value is normal. Among them, the preset extreme value threshold can be obtained according to actual tests.

[0030] The extreme value unit is further configured to generate a fitted mean and a fitted variance according to the time series data by using the maximum likelihood estimation method, and determine the theoretical proportion of extreme values based on the fitted mean and the fitted variance by using the cumulative distribution function, determine the actual proportion of extreme values according to the mean and the variance, and generate an extreme value proportion abnormal judgment result according to the theoretical proportion of extreme values and the actual proportion of extreme values.

[0031] Specifically, the extreme value unit is further configured to assume that the time series data follows a normal distribution, fit the distribution parameters and solve them by using the maximum likelihood estimation method to obtain the fitted mean and the fitted variance, and thus determine the theoretical proportion of extreme values by using the cumulative distribution function (CDF) according to the fitted mean and the fitted variance. Among them, calculating the proportion of extreme values by using the cumulative distribution function (CDF) is an existing and widely used technology and will not be elaborated here. Then, the theoretical proportion of extreme values and the actual proportion of extreme values are compared. If the actual proportion of extreme values is within the preset difference range of the theoretical proportion of extreme values, the extreme value proportion abnormal judgment result is that the extreme value proportion is normal. If the actual proportion of extreme values is not within the preset difference range of the theoretical proportion of extreme values, the extreme value proportion abnormal judgment result is that the extreme value proportion is abnormal. The preset difference range can be obtained according to actual tests.

[0032] In this embodiment, the present invention connects to a terminal device through a 5G connection module based on the fifth-generation mobile communication technology, which can provide a stable connection channel and a fast data transmission channel, so as to stably and quickly obtain the device data of the terminal device. Through the data timeout unit of the anomaly detection module, a bidirectional LSTM model is used to capture the long-term dependence relationship of the time series of historical reporting time data through forward and backward propagation, and the influence of environmental data is considered, and accurate predicted reporting time data can be generated, that is, the normal reporting time of the terminal device. Then, the actual actually reported time data is compared with the predicted reporting time data to accurately judge whether the data upload of the terminal device connected to the Internet of Things is normal. Then, through the extreme value unit of the anomaly detection module, a multi-scale sliding window is used to analyze the statistical eigenvalue of each window of the time series data and the weight corresponding to the statistical eigenvalue. Different statistical eigenvalues at different time scales can be captured through multiple window sizes, that is, the mean and variance, so as to comprehensively describe the distribution characteristics of the data. And through the corresponding weights, the model pays more attention to the windows with high certainty. Then, according to the time series data and the statistical eigenvalues, a hidden Markov model is used to determine the extreme value anomaly probability of the extreme values in the time series data, so as to facilitate subsequent judgment of whether the extreme values in the current data belong to the abnormal state. Finally, considering the statistical eigenvalues, weights, and extreme value anomaly probability comprehensively, it is reasonably judged whether the extreme values of the terminal device connected to the Internet of Things are abnormal. The extreme value unit also uses the maximum likelihood estimation method and the cumulative distribution function to determine the proportion of extreme values in the theoretical situation, that is, the theoretical proportion of extreme values, and then compares the theoretical proportion of extreme values with the actual proportion of extreme values to judge whether the proportion of extreme values of the terminal device connected to the Internet of Things is abnormal. The present invention analyzes and judges the data upload time, extreme values, and the proportion of extreme values of the terminal device connected to the Internet of Things, and can further analyze whether the connection of the terminal device is abnormal.

[0033] Optionally, the step of using the bidirectional LSTM model to generate the predicted reporting time data for the next moment according to the historical reporting time data and the environmental data includes: Using the bidirectional LSTM model, according to the historical reporting time data and the preliminary prediction formula, preliminary prediction data is generated. The preliminary prediction formula includes: ; ; where h t is the output of the hidden layer of the bidirectional LSTM model, is the historical reporting time data, n is n moments, t is t moments, BiLSTM is the calculation rule of the hidden layer of the bidirectional LSTM model, are the trainable parameters of the bidirectional LSTM model, including the weight matrix and the bias term, is the preliminary prediction data, is the weight matrix that maps the output of the hidden layer of the bidirectional LSTM model to the prediction time, is the bias term for the prediction time; An attention mechanism is adopted. According to the importance weight formula, the environmental data is weighted and summed to generate environmental impact degree data. The importance weight formula includes: ; ; where, is the importance weight of the environmental data, softmax is the attention calculation rule, is the environmental data, is the weight matrix that maps the environmental data to the attention weight space, is the bias term of the attention mechanism, is the environmental impact degree data, is the importance weight of the i-th environmental data, is the i-th environmental data; A fusion formula is used to fuse the preliminary prediction data and the environmental impact degree data to generate the predicted reporting time data. The fusion formula includes: T = f ( T t+1 , ct ; θ f ) where, T is the predicted reporting time data, f is the fusion rule, θ f are the trainable parameters of the fusion, including the weight matrix and the bias term.

[0034] Specifically, a bidirectional LSTM model is adopted. According to the historical reporting time data and the preliminary prediction formula, preliminary prediction is carried out to generate preliminary prediction data. Then, considering the influence of environmental variables (such as network delay, packet loss rate, etc.) on the prediction result, the adaptability of the model to complex scenarios is improved. The attention mechanism is used to calculate the importance weight of each environmental data, and all environmental data are weighted and summed to obtain the environmental impact degree data. Finally, a fusion formula is used to fuse the preliminary prediction data and the environmental impact degree data to generate more accurate predicted reporting time data.

[0035] Optionally, generating a data timeout exception judgment result according to the predicted reporting time data and the actual reporting time data includes: Based on the predicted reporting time data and the environmental data, a dynamic timeout threshold is determined using a dynamic timeout threshold formula, and the dynamic timeout threshold formula includes: ; wherein, is the dynamic timeout threshold, is the predicted reporting time data, is the safety buffer coefficient, is the standard deviation of the historical reporting time data, is the weight coefficient, is the degree of change of the environmental data; Based on the actual reporting time data and the dynamic timeout threshold, the data timeout exception judgment result is generated.

[0036] Specifically, based on the predicted reporting time data, historical reporting time data, and environmental data, a dynamic timeout threshold formula is used to dynamically adjust the timeout threshold to consider the impacts brought by various situations, and then accurately judge whether it times out. If the actual reporting time data does not exceed the dynamic timeout threshold, the exception judgment result is that the data reporting does not time out and the reporting time is normal. If the actual reporting time data exceeds the dynamic timeout threshold, the exception judgment result is that the data reporting times out and the reporting time is abnormal. The safety buffer coefficient refers to the safety buffer set according to the volatility of historical data, and the degree of change of the environmental data can be obtained based on actual tests.

[0037] Optionally, the use of a multi-scale sliding window to analyze the statistical feature values of each window of the time series data and the weights corresponding to the statistical feature values includes: Using the multi-scale sliding window to analyze the statistical feature values of each window of the time series data and the uncertainty entropy value of the statistical feature value of each window; Based on the uncertainty entropy value, the weight is determined, where the uncertainty entropy value is inversely proportional to the corresponding weight.

[0038] Specifically, a multi-scale sliding window is used to analyze the statistical feature values of each window of the time series data and the uncertainty entropy value of the statistical feature value of each window; based on the uncertainty entropy value, the weight is determined, where the uncertainty entropy value is inversely proportional to the corresponding weight, that is, the lower the uncertainty entropy value, the more stable the data and the higher the weight, and the higher the uncertainty entropy value, the more unstable the data and the lower the weight.

[0039] Optionally, the generation of the extreme value comprehensive anomaly score based on the statistical feature value, the weight, and the extreme value anomaly probability includes: Based on the statistical eigenvalue, the weight, and the extreme value anomaly probability, use a comprehensive formula to generate the extreme value comprehensive anomaly score, where the comprehensive formula includes: ; where, is the extreme value comprehensive anomaly score, is the weight, is the time series data, is the mean value, is the variance, is a regulation parameter, is the extreme value anomaly probability.

[0040] Specifically, the first term in the comprehensive formula can measure the deviation degree between the extreme value and the window mean value, and the second term introduces the extreme value anomaly probability for anomaly recognition, increasing the attention to the abnormal state, so as to obtain a more comprehensive extreme value comprehensive anomaly score for final judgment.

[0041] Optionally, the terminal device includes a camera, the device data includes historical video data and current video data corresponding to the camera, and the anomaly detection module further includes an occlusion detection unit; The occlusion detection unit is used to extract the unoccluded background features of the historical video data by using a deep learning model, and based on the unoccluded background features, construct an unoccluded virtual model by using an adversarial network, and input the current video data into the trained unoccluded virtual model to generate an occlusion detection result.

[0042] Specifically, the occlusion detection unit uses a deep learning model, such as a convolutional neural network, to extract the unoccluded background features of the historical video data. The unoccluded background features include background and foreground features, and based on the unoccluded background features, uses an adversarial network to construct a high-resolution unoccluded virtual model for the normal unoccluded scene, input the current video data into the trained unoccluded virtual model, and compare the image in the unoccluded virtual model with the video frame in the current video data to determine whether occlusion occurs and generate an occlusion detection result.

[0043] Optionally, the anomaly detection module further includes a screen freeze detection unit; The screen freeze detection unit is used to convert the current video data into the Lab color space and use wavelet transform to analyze the video data in the Lab color space to generate a screen freeze detection result.

[0044] Specifically, the screen freeze detection unit converts the current video data into the Lab color space to reduce the influence of light, and uses wavelet transform to analyze the video data in the Lab color space to identify noise or distortion regions, so as to perform screen freeze detection and generate a screen freeze detection result.

[0045] Optionally, the anomaly detection module further includes a blur detection unit; The blur detection unit is configured to perform frequency domain analysis on the current video data by using fast Fourier transform to generate a blur detection result.

[0046] Specifically, the blur detection unit performs frequency domain analysis on the image by using fast Fourier transform (FFT) to identify the blur caused by excessive low-frequency components and generates a blur detection result.

[0047] Such as Figure 2 As shown, an electronic device 200 provided by an embodiment of the present invention includes a memory 210 and a processor 220; the memory 210 is used to store a computer program; the processor 220 is configured to, when executing the computer program, implement the following steps: Based on the fifth-generation mobile communication technology, connect to the terminal device and obtain the device data of the terminal device, where the device data includes historical reporting time data, actual reporting time data, environmental data, and time series data; Adopt a bidirectional LSTM model to generate predicted reporting time data for the next moment according to the historical reporting time data and the environmental data, and generate a data timeout anomaly judgment result according to the predicted reporting time data and the actual reporting time data; Adopt a multi-scale sliding window to analyze the statistical feature values of each window of the time series data and the weights corresponding to the statistical feature values, and use a hidden Markov model according to the time series data and the statistical feature values to determine the extreme value anomaly probability. Generate an extreme value comprehensive anomaly score according to the statistical feature values, the weights, and the extreme value anomaly probability, and generate an extreme value anomaly judgment result according to the extreme value comprehensive anomaly score, where the statistical feature values include the mean and variance; According to the time series data, adopt the maximum likelihood estimation method to generate a fitted mean and a fitted variance, and based on the fitted mean and the fitted variance, use the cumulative distribution function to determine the theoretical proportion of extreme values, and determine the actual proportion of extreme values according to the mean and the variance. Generate an extreme value proportion anomaly judgment result according to the theoretical proportion of extreme values and the actual proportion of extreme values.

[0048] A computer-readable storage medium provided by an embodiment of the present invention has a computer program stored thereon. When the computer program is executed by a processor, the following steps are implemented: Based on the fifth-generation mobile communication technology, connect to the terminal device and obtain the device data of the terminal device, where the device data includes historical reporting time data, actual reporting time data, environmental data, and time series data; Adopt a bidirectional LSTM model to generate predicted reporting time data for the next moment based on the historical reported time data and the environmental data, and generate a data timeout exception judgment result based on the predicted reporting time data and the actual reported time data; Adopt a multi-scale sliding window to analyze the statistical eigenvalue of each window of the time series data and the weight corresponding to the statistical eigenvalue, and adopt a hidden Markov model based on the time series data and the statistical eigenvalue to determine the extreme value anomaly probability. Generate an extreme value comprehensive anomaly score according to the statistical eigenvalue, the weight, and the extreme value anomaly probability, and generate an extreme value anomaly judgment result according to the extreme value comprehensive anomaly score, where the statistical eigenvalue includes the mean value and the variance; Adopt the maximum likelihood estimation method based on the time series data to generate a fitted mean value and a fitted variance, and based on the fitted mean value and the fitted variance, adopt the cumulative distribution function to determine the extreme value theoretical proportion. Determine the extreme value actual proportion according to the mean value and the variance, and generate an extreme value proportion anomaly judgment result according to the extreme value theoretical proportion and the extreme value actual proportion.

[0049] Now, an electronic device 200 that can be a server or a client of the present invention will be described. It is an example of a hardware device that can be applied to various aspects of the present invention. The electronic device 200 is intended to represent various forms of digital electronic computer devices, such as, laptop computers, desktop computers, workstations, personal digital assistants, servers, blade servers, mainframe computers, and other suitable computers. The electronic device 200 can also represent various forms of mobile devices, such as, personal digital processors, cellular phones, smart phones, wearable devices, and other similar computing devices. The components shown herein, their connections and relationships, and their functions are merely examples and are not intended to limit the implementation of the present invention described and / or claimed herein.

[0050] The electronic device 200 includes a computing unit that can perform various appropriate actions and processes according to a computer program stored in a read-only memory (ROM) or a computer program loaded from a storage unit into a random access memory (RAM). In the RAM, various programs and data required for device operation can also be stored. The computing unit, the ROM, and the RAM are connected to each other through a bus. An input / output (I / O) interface is also connected to the bus.

[0051] Those of ordinary skill in the art can understand that all or part of the processes in the methods of the above embodiments can be completed by instructing relevant hardware through a computer program. The program can be stored in a computer-readable storage medium. When the program is executed, it can include the processes of the embodiments of the above methods. Among them, the storage medium can be a magnetic disk, an optical disk, a read-only memory (ROM), or a random access memory (RAM), etc. In this application, the units described as separate components may or may not be physically separated. The components shown as units may or may not be physical units, that is, they may be located in one place or distributed to multiple network units. Some or all of the units can be selected according to actual needs to achieve the purpose of the solution of the embodiments of the present invention. In addition, the functional units in the various embodiments of the present invention can be integrated in a processing unit, or each unit can exist physically alone, or two or more units can be integrated in one unit. The above integrated units can be implemented in the form of hardware or in the form of software functional units.

[0052] Although the present invention is disclosed as above, the scope of protection of the present invention is not limited thereto. Those skilled in the art can make various changes and modifications without departing from the spirit and scope of the present invention, and these changes and modifications will all fall within the scope of protection of the present invention.

Claims

1. An Internet of Things system based on the fifth-generation mobile communication technology, characterized in that It includes a 5G connection module and an anomaly detection module; The 5G connection module is based on the fifth-generation mobile communication technology, connects to the terminal device, and obtains the device data of the terminal device, where the device data includes historical reporting time data, actual reporting time data, environmental data, and time series data; The anomaly detection module includes a data timeout unit and an extreme value unit. The data timeout unit is used to generate predicted reporting time data for the next moment according to the historical reporting time data and the environmental data by using a bidirectional LSTM model, and generate a data timeout anomaly judgment result according to the predicted reporting time data and the actual reporting time data; The extreme value unit is used to analyze the statistical eigenvalue of each window of the time series data and the weight corresponding to the statistical eigenvalue by using a multi-scale sliding window, and determine the extreme value anomaly probability by using a hidden Markov model according to the time series data and the statistical eigenvalue. According to the statistical eigenvalue, the weight, and the extreme value anomaly probability, generate an extreme value comprehensive anomaly score, and generate an extreme value anomaly judgment result according to the extreme value comprehensive anomaly score, where the statistical eigenvalue includes the mean value and the variance; The extreme value unit is also used to generate a fitted mean value and a fitted variance by using the maximum likelihood estimation method according to the time series data, and determine the theoretical proportion of extreme values by using the cumulative distribution function based on the fitted mean value and the fitted variance. Determine the actual proportion of extreme values according to the mean value and the variance, and generate an extreme value proportion anomaly judgment result according to the theoretical proportion of extreme values and the actual proportion of extreme values.

2. The Internet of Things system based on the fifth-generation mobile communication technology according to claim 1, characterized in that The method of generating predicted reporting time data for the next moment by using the bidirectional LSTM model according to the historical reporting time data and the environmental data includes: Using the bidirectional LSTM model to generate preliminary prediction data according to the historical reporting time data and a preliminary prediction formula, and the preliminary prediction formula includes: ； ； Among them, h t is the output of the hidden layer of the bidirectional LSTM model, is the historical reporting time data, n represents n moments, t represents t moments, and BiLSTM is the calculation rule of the hidden layer of the bidirectional LSTM model, are the trainable parameters of the bidirectional LSTM model, including the weight matrix and the bias term, is the preliminary prediction data, is the weight matrix that maps the output of the hidden layer of the bidirectional LSTM model to the prediction time, is the bias term of the prediction time; Using an attention mechanism to weight and sum the environmental data according to an importance weight formula to generate environmental impact degree data, and the importance weight formula includes: ； ； Among them, is the importance weight of the environmental data, softmax is the attention calculation rule, is the environmental data, is the weight matrix for mapping the environmental data to the attention weight space, is the bias term of the attention mechanism, is the environmental impact degree data, is the importance weight of the i-th environmental data, is the i-th environmental data; Using a fusion formula to fuse the preliminary prediction data and the environmental impact degree data to generate the predicted reporting time data, and the fusion formula includes: T = f ( T t+1 , ct ; θ f ) Among them, T is the predicted reporting time data, and f is the fusion rule. θ f are the trainable parameters for fusion, including the weight matrix and the bias term.

3. The Internet of Things system based on the fifth-generation mobile communication technology according to claim 1, characterized in that The method of generating a data timeout anomaly judgment result according to the predicted reporting time data and the actual reporting time data includes: Determining a dynamic timeout threshold according to the predicted reporting time data and the environmental data by using a dynamic timeout threshold formula, and the dynamic timeout threshold formula includes: ； Among them, is the dynamic timeout threshold, is the predicted reporting time data, is the safety buffer coefficient, is the standard deviation of the historical reporting time data, is the weight coefficient, is the degree of change of the environmental data; Generating the data timeout anomaly judgment result according to the actual reporting time data and the dynamic timeout threshold.

4. The Internet of Things system based on the fifth-generation mobile communication technology according to claim 1, characterized in that, The method of analyzing the statistical eigenvalue of each window of the time series data and the weight corresponding to the statistical eigenvalue by using a multi-scale sliding window includes: Using the multi-scale sliding window to analyze the statistical eigenvalue of each window of the time series data and the uncertainty entropy value of the statistical eigenvalue of each window; Determine the weight according to the uncertainty entropy value, where the uncertainty entropy value is inversely proportional to the corresponding weight.

5. The Internet of Things system based on the fifth-generation mobile communication technology according to claim 1, wherein The generating of the extreme value comprehensive anomaly score according to the statistical eigenvalue, the weight, and the extreme value anomaly probability includes: Generating the extreme value comprehensive anomaly score according to the statistical eigenvalue, the weight, and the extreme value anomaly probability by using a comprehensive formula, where the comprehensive formula includes: ； Among them, is the comprehensive extreme anomaly score, is the weight, is the time series data, is the mean value, is the variance, is the adjustment parameter, is the extreme anomaly probability.

6. The Internet of Things system based on the fifth-generation mobile communication technology according to claim 1, characterized in that The terminal device includes a camera, the device data includes historical video data and current video data corresponding to the camera, and the anomaly detection module further includes an occlusion detection unit; The occlusion detection unit is configured to extract the unoccluded background features of the historical video data by using a deep learning model, and based on the unoccluded background features, construct an unoccluded virtual model by using an adversarial network, and input the current video data into the trained unoccluded virtual model to generate an occlusion detection result.

7. The Internet of Things system based on the fifth-generation mobile communication technology according to claim 6, wherein The anomaly detection module further includes a screen freeze detection unit; The screen freeze detection unit is configured to convert the current video data into the Lab color space, and analyze the video data in the Lab color space by using wavelet transform to generate a screen freeze detection result.

8. The Internet of Things system based on the fifth-generation mobile communication technology according to claim 6, characterized in that, The anomaly detection module further includes a blur detection unit; The blur detection unit is configured to perform frequency domain analysis on the current video data by using fast Fourier transform to generate a blur detection result.

9. An electronic device, characterized in that, Includes a memory and a processor; The memory is used to store a computer program; The processor is configured to, when executing the computer program, implement the following steps: Based on the fifth-generation mobile communication technology, connect to the terminal device and obtain the device data of the terminal device, where the device data includes historical reporting time data, actual reporting time data, environmental data, and time series data; Adopt a bidirectional LSTM model to generate predicted reporting time data for the next moment according to the historical reporting time data and the environmental data, and generate a data timeout anomaly judgment result according to the predicted reporting time data and the actual reporting time data; Adopt a multi-scale sliding window to analyze the statistical eigenvalue of each window of the time series data and the weight corresponding to the statistical eigenvalue, and determine the extreme value anomaly probability according to the time series data and the statistical eigenvalue by using a hidden Markov model, generate an extreme value comprehensive anomaly score according to the statistical eigenvalue, the weight, and the extreme value anomaly probability, and generate an extreme value anomaly judgment result according to the extreme value comprehensive anomaly score, where the statistical eigenvalue includes the mean and the variance; Generate a fitted mean and a fitted variance according to the time series data by using the maximum likelihood estimation method, and determine the extreme value theoretical proportion by using the cumulative distribution function based on the fitted mean and the fitted variance, determine the extreme value actual proportion according to the mean and the variance, and generate an extreme value proportion anomaly judgment result according to the extreme value theoretical proportion and the extreme value actual proportion.

10. A computer-readable storage medium, characterized in that, A computer program is stored on the storage medium, and when the computer program is executed by the processor, the following steps are implemented: Based on the fifth-generation mobile communication technology, connect to the terminal device and obtain the device data of the terminal device, where the device data includes historical reporting time data, actual reporting time data, environmental data, and time series data; Adopt a bidirectional LSTM model to generate predicted reporting time data for the next moment according to the historical reporting time data and the environmental data, and generate a data timeout exception judgment result according to the predicted reporting time data and the actual reporting time data; Adopt a multi-scale sliding window to analyze the statistical eigenvalue of each window of the time series data and the weight corresponding to the statistical eigenvalue, and according to the time series data and the statistical eigenvalue, use the hidden Markov model to determine the extreme value anomaly probability. According to the statistical eigenvalue, the weight, and the extreme value anomaly probability, generate an extreme value comprehensive anomaly score, and generate an extreme value anomaly judgment result according to the extreme value comprehensive anomaly score, where the statistical eigenvalue includes the mean and variance; According to the time series data, use the maximum likelihood estimation method to generate a fitted mean and a fitted variance, and based on the fitted mean and the fitted variance, use the cumulative distribution function to determine the extreme value theoretical proportion. According to the mean and the variance, determine the extreme value actual proportion. According to the extreme value theoretical proportion and the extreme value actual proportion, generate an extreme value proportion anomaly judgment result.

Images