Abnormal radio signal monitoring method and system based on artificial intelligence and big data

Abstract

The invention relates to the technical field of signal detection, in particular to an abnormal radio signal monitoring method and system based on artificial intelligence and big data, and the method comprises the steps: obtaining radio signal data, collected by sensors and receivers in an Internet of Things environment, of each device in real time; the collected radio signal data is preprocessed; extracting abnormal signal features from the preprocessed radio signal data by using a feature engineering technology; training a convolutional neural network model based on the marked historical abnormal signal feature data; deploying the trained convolutional neural network model into a monitoring system, monitoring a radio signal of the Internet of Things equipment in real time, and automatically identifying an abnormal signal; and generating corresponding alarm information according to the severity of the abnormal signal, and sending an alarm notification through a corresponding channel. Therefore, the signal quality can be comprehensively judged according to the characteristics of various radio signals so as to monitor the fault condition, and the accuracy is high.

Description

Technical Field

[0001] The present invention relates to the field of signal detection technology, and in particular to an abnormal radio signal monitoring method and system based on artificial intelligence and big data. Background Art

[0002] Radio signal monitoring of IoT devices is one of the key technologies to ensure the normal operation, safety, reliability and efficient management of these devices. Through radio signal monitoring, important functions such as real-time monitoring of device status, fault detection and diagnosis, performance optimization and safety protection can be achieved.

[0003] Existing radio signal monitoring mostly uses a single signal to collect and monitor signals, and this method of using a single signal source to judge has low accuracy. Summary of the invention

[0004] The purpose of the present invention is to provide an abnormal radio signal monitoring method and system based on artificial intelligence and big data, aiming to use the characteristics of multiple radio signals to comprehensively judge the signal quality to monitor fault conditions, so that the accuracy is higher.

[0005] To achieve the above-mentioned purpose, in a first aspect, the present invention provides an abnormal radio signal monitoring method based on artificial intelligence and big data, comprising acquiring sensors and receivers distributed in an Internet of Things environment to collect radio signal data of each device in real time; Preprocessing the collected radio signal data; Extract abnormal signal features from preprocessed radio signal data using feature engineering techniques; The convolutional neural network model is trained based on the marked historical abnormal signal feature data; Deploy the trained convolutional neural network model into the monitoring system to monitor the radio signals of IoT devices in real time and automatically identify abnormal signals; According to the severity of the abnormal signal, the corresponding alarm information is generated and the alarm notification is sent through the corresponding channel.

[0006] Wherein, after generating corresponding alarm information according to the severity of the abnormal signal and sending the alarm notification through the corresponding channel, the method also includes: storing the monitored radio signal data and the abnormality detection result in a database.

[0007] After storing the monitored radio signal data and the abnormality detection results in a database, the method further includes displaying the abnormality detection results through a graphical interface.

[0008] The radio signal data includes signal strength, frequency, bandwidth and modulation mode.

[0009] The specific steps of obtaining the radio signal data of each device collected in real time by sensors and receivers distributed in the IoT environment include: Configuring acquisition parameters of sensors and receivers, including sampling frequency, sampling interval, and signal strength threshold; Configure the receiver's network parameters; Start the sensor and receiver to start collecting radio signal data in real time; Cache the collected radio signal data.

[0010] The specific steps of preprocessing the collected radio signal data include: Collect radio signal data from different sources; Use FIR filters to remove background noise and other interference from radio signal data; Correcting phase distortion in radio signal data due to the transmission path; The corrected radio signal data is normalized.

[0011] The specific steps of extracting abnormal signal features from the preprocessed radio signal data using feature engineering technology include: Calculate the mean, standard deviation and fluctuation range of signal strength in radio signal data; Calculate the mean, standard deviation, and drift rate of frequencies in radio signal data; Calculate the modulation error rate and phase error to obtain the abnormal signal characteristics.

[0012] The specific steps of training the convolutional neural network model based on the marked historical abnormal signal feature data include: Collect normal signals and abnormal signals in historical radio signal data to obtain a data set; Label each data sample in the dataset to identify whether it is an abnormal signal; Extract all abnormal signal features of samples in the data set and divide them into training set, validation set and test set; Train the convolutional neural network model based on the training set to obtain a recognition model; The recognition model is verified based on the validation set and tested using the test set.

[0013] In a second aspect, the present invention also provides an abnormal radio signal monitoring system based on artificial intelligence and big data, including a data acquisition module, a preprocessing module, a feature extraction module, a model generation module, a monitoring module and an alarm module: The data acquisition module is used to acquire the radio signal data of each device collected in real time by sensors and receivers distributed in the IoT environment; The preprocessing module is used to preprocess the collected radio signal data; The feature extraction module is used to extract abnormal signal features from the pre-processed radio signal data using feature engineering technology; The model generation module is used to train a convolutional neural network model based on the marked historical abnormal signal feature data; The monitoring module is used to deploy the trained convolutional neural network model into the monitoring system, monitor the radio signals of IoT devices in real time, and automatically identify abnormal signals; The alarm module is used to generate corresponding alarm information according to the severity of the abnormal signal and send the alarm notification through the corresponding channel.

[0014] The abnormal radio signal monitoring method and system based on artificial intelligence and big data of the present invention collects the radio signal data emitted by each device in real time through multiple sensors and receivers distributed in the Internet of Things environment. These sensors and receivers can cover a wide geographical area to ensure that the collected data is comprehensive and representative. The collected raw radio signal data often contains noise, missing values ​​or other non-ideal factors, so preprocessing is required. After data preprocessing, feature engineering technology is used to conduct in-depth analysis of the processed radio signal data to extract features that can effectively characterize abnormal signals. These features include changes in signal strength, spectral characteristics, time domain and frequency domain properties of the signal, etc. Through this process, the data dimension can be significantly reduced while retaining information that is critical to anomaly detection.

[0015] A convolutional neural network (CNN) model is trained using a dataset of labeled abnormal signal features. The trained CNN model will be deployed in an actual monitoring system. This system is able to continuously monitor the radio signals of IoT devices and automatically identify any abnormal signals using the trained model. This real-time monitoring capability is critical for responding quickly to potential security threats. When the monitoring system detects an abnormal signal, it automatically evaluates and generates corresponding alarm information based on the severity of the signal. These alarm messages can be sent to relevant personnel or systems through multiple channels, such as email, SMS, application push notifications, etc. The purpose of this is to ensure that relevant personnel can obtain information in a timely manner and take necessary measures to prevent or mitigate the impact of potential risks. BRIEF DESCRIPTION OF THE DRAWINGS

[0016] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0017] Figure 1 It is a flow chart of the abnormal radio signal monitoring method based on artificial intelligence and big data of the present invention.

[0018] Figure 2 It is a flow chart of the present invention for obtaining the radio signal data of each device collected in real time by sensors and receivers distributed in the Internet of Things environment.

[0019] Figure 3 It is a flow chart of preprocessing collected radio signal data according to the present invention.

[0020] Figure 4 It is a flow chart of the present invention for extracting abnormal signal features from preprocessed radio signal data using feature engineering technology.

[0021] Figure 5 It is a flow chart of the present invention for training a convolutional neural network model based on marked historical abnormal signal feature data.

[0022] Figure 6 It is a flow chart of the present invention for generating corresponding alarm information according to the severity of abnormal signals and sending alarm notifications through corresponding channels. DETAILED DESCRIPTION

[0023] Embodiments of the present invention are described in detail below, examples of which are shown in the accompanying drawings, wherein the same or similar reference numerals throughout represent the same or similar elements or elements having the same or similar functions. The embodiments described below with reference to the accompanying drawings are exemplary and are intended to be used to explain the present invention, and should not be construed as limiting the present invention.

[0024] First embodiment See also Figure 1 to Figure 6 The present invention provides an abnormal radio signal monitoring method based on artificial intelligence and big data, comprising: S101 obtains sensors and receivers distributed in the IoT environment to collect radio signal data of each device in real time; The radio signal data includes signal strength, frequency, bandwidth and modulation. In the IoT environment, sensors and receivers are usually deployed in a wide range of physical spaces to monitor radio signal activities in a specific area. These devices can capture various radio signals in the surrounding environment in real time, thereby providing raw materials for subsequent data analysis. The collected radio signal data is not limited to simple presence or absence, but also includes more detailed attributes such as signal strength (i.e., the strength of the signal), frequency (the wavelength or period of the signal), bandwidth (the spectrum range occupied by the signal) and modulation (how to encode information for transmission via radio waves).

[0025] The specific steps include: S201 configures acquisition parameters of the sensor and the receiver, wherein the acquisition parameters include sampling frequency, sampling interval, and signal strength threshold; In order to ensure that the collected data not only meets the needs of research or application, but also maintains high efficiency and accuracy, it is necessary to configure the working parameters of sensors and receivers in advance. This mainly includes the following aspects: The sampling frequency determines the number of times samples are extracted from the radio signal per second. A higher sampling frequency can obtain more refined signal features, but it also means a larger amount of data and higher processing requirements. The sampling interval corresponds to the sampling frequency and refers to the time difference between two consecutive samples. Properly setting the sampling interval helps to balance the relationship between data accuracy and resource consumption. The signal strength threshold sets a minimum signal strength standard, and only when the received signal exceeds this threshold will it be recorded. This can effectively filter out background noise and other non-relevant signals, and improve the relevance and availability of data.

[0026] S202 configures the network parameters of the receiver; In addition to the basic acquisition parameters, the network configuration of the receiver also needs to be adjusted according to the actual application scenario. This involves the selection of IP addresses, port numbers, communication protocols, etc., in order to ensure that the receiver can interact with other network components stably and efficiently, and smoothly upload the collected data to the central server or cloud platform for further processing.

[0027] S203 starts the sensor and the receiver to start collecting radio signal data in real time; After completing the above configuration, the next step is to activate the entire data collection system. At this point, all sensors and receivers will start working according to the preset parameters, continuously monitoring and recording the changes in radio signals within their coverage area. This process is automated and does not require human intervention, but it is necessary to regularly check the system's operating status to ensure that everything is operating normally.

[0028] S204 caches the collected radio signal data.

[0029] Since radio signal data is generated at a high speed, sending it directly to the final destination will result in data loss or delay. Therefore, it is very necessary to temporarily store this data in a cache before formal transmission. This not only reduces the network load, but also ensures the integrity and reliability of the data. Once conditions permit, the data in the cache can be uploaded in batches to achieve the best data management effect.

[0030] S102 pre-processes the collected radio signal data; The specific steps include: S301 collects radio signal data from different sources; Radio signal data needs to be collected from multiple different sensors and receivers. These data sources are located in different geographical locations and cover different environmental conditions, so the data they provide is diverse. It is crucial to ensure that all data sources can stably transmit data to the central processing unit in order to obtain a comprehensive and rich data set to lay the foundation for subsequent processing and analysis.

[0031] S302 uses an FIR filter to remove background noise and other interference from the radio signal data; In the collected radio signal data, some background noise and other forms of interference are inevitably mixed in. These interferences come from electromagnetic radiation from electronic devices, other wireless communication systems, or electromagnetic phenomena in nature. In order to improve the quality of the data, appropriate signal processing technology is needed to eliminate these interferences. In this process, FIR (Finite Impulse Response) filter is a commonly used tool. FIR filter can effectively remove unnecessary frequency components by designing specific frequency response characteristics while retaining useful signal information, thereby achieving the purpose of purifying data.

[0032] S303 corrects phase distortion in the radio signal data caused by the transmission path; During the transmission process, radio signals are affected by multipath effects, reflections, refractions and other factors, resulting in phase distortion of the signals received by the receiving end. This distortion affects the integrity of the signal, and in turn affects the decisions or analysis results based on these signals. Therefore, it is necessary to correct the signal through an algorithm to restore its original phase characteristics. Common methods include using channel estimation technology to compensate for phase offsets to ensure the authenticity and reliability of the signal.

[0033] S304 performs normalization processing on the corrected radio signal data.

[0034] After the previous steps, although most of the noise and interference have been removed and the phase distortion has been corrected, there are still amplitude differences between the data from different sources. In order to make these data comparable and analyzed on the same scale, they need to be normalized. Normalization refers to converting the data to a specific range (for example, between 0 and 1), which eliminates differences caused by different measurement units or device sensitivities, making the data more consistent and easier to analyze. In addition, normalization also helps to accelerate the training process of subsequent machine learning models and improve the predictive performance of the model.

[0035] By performing the above preprocessing steps on the collected radio signal data, the quality of the data can be significantly improved, providing a solid foundation for subsequent data analysis and application.

[0036] S103 uses feature engineering technology to extract abnormal signal features from the preprocessed radio signal data; The specific steps include: S401 calculates the mean, standard deviation and fluctuation range of the signal strength in the radio signal data; Calculating the average value of signal strength over a period of time can be used to indicate the general level of the signal. Abnormal signals can cause a sudden increase or decrease in signal strength, so the mean of abnormal signals will have a significant change compared to the mean under normal circumstances.

[0037] The standard deviation measures the dispersion of the signal strength, that is, the dispersion of the data points relative to the mean. The larger the standard deviation, the more unstable the signal strength changes. When the signal is disturbed or fails, its standard deviation tends to increase.

[0038] The fluctuation range refers to the difference between the maximum and minimum values ​​of the signal strength. Abnormal signals cause the fluctuation range of the signal strength to increase significantly, so this is also an important characteristic indicator.

[0039] S402 calculates the mean, standard deviation and drift rate of the frequency in the radio signal data; Similarly, the average value of the signal frequency over a period of time is calculated to reflect the basic level of the signal frequency. Abnormal signals can cause sudden changes in frequency, resulting in changes in the mean value.

[0040] The standard deviation is a measure of the stability of the signal frequency. If the signal is interfered with or if there is a problem with the equipment, the standard deviation of the frequency will increase.

[0041] The drift rate calculates the rate at which the signal frequency changes over time. Under normal circumstances, the signal frequency should be relatively stable, but if an abnormality occurs, the frequency will gradually deviate from its initial value, and the drift rate can quantify the trend of this change.

[0042] S403 calculates the modulation error rate and the phase error to obtain abnormal signal characteristics.

[0043] Modulation Error Rate is an important parameter to measure the quality of digital modulated signals. It reflects the deviation between the actual constellation points of the signal and its ideal position. High MER values ​​indicate poor signal quality, interference or equipment failure. In abnormal signal detection, MER can help identify signals with degraded modulation quality. Phase error refers to the deviation between the actual signal phase and the ideal phase. In digital communication systems, phase error will increase the bit error rate and affect the communication quality. By calculating the phase error, signals with unstable phase or phase offset can be found, which are usually signs of abnormal signals.

[0044] Through the above steps, multiple indicators reflecting abnormal signal characteristics can be extracted from the preprocessed radio signal data. These characteristics can not only be used to detect whether the signal is abnormal, but also to further analyze the cause of the abnormality, providing a basis for maintaining and optimizing the IoT system.

[0045] S104 trains a convolutional neural network model based on the marked historical abnormal signal feature data; The specific steps include: S501 collects normal signals and abnormal signals in historical radio signal data to obtain a data set; Collect a large amount of radio signal data from historical records. The data should cover a variety of scenarios and conditions, including signals under normal operating conditions and known abnormal signals. The data source can be sensors, receivers, or other monitoring devices. Perform preliminary cleaning on the collected data to remove invalid or erroneous data points to ensure the quality of the data set.

[0046] S502 labels each data sample in the data set to identify whether it is an abnormal signal; Define a labeling system using binary labels such as "normal" (0) and "abnormal" (1).

[0047] Organize a team of experts to manually label each sample in the dataset to ensure the accuracy and consistency of the labels. You can also use semi-supervised learning or active learning methods to generate labels by combining a small amount of high-quality manually labeled data with a large amount of unlabeled data.

[0048] Ensure that the number of normal signals and abnormal signals in the data set is relatively balanced to avoid the model being biased towards one category. If the data is unbalanced, oversampling (such as SMOTE) or undersampling techniques can be used to adjust the data distribution.

[0049] S503 extracts all abnormal signal features of samples in the data set and divides them into a training set, a validation set, and a test set; Using the feature engineering techniques mentioned above (such as S401, S402, and S403), extract features such as the mean and standard deviation of signal strength, fluctuation range, mean, standard deviation and drift rate of frequency, modulation error rate and phase error from each data sample. The dataset after feature extraction is randomly divided into training set, validation set and test set. The common ratio allocation is 70% for training, 15% for validation, and 15% for testing. Make sure that the ratio of normal signals to abnormal signals in each subset is roughly the same to maintain the representativeness of the data.

[0050] S504 trains a convolutional neural network model based on the training set to obtain a recognition model; Select a suitable convolutional neural network architecture, such as LeNet, VGG, ResNet, etc., and adjust it according to the characteristics of the dataset and task requirements. Perform preprocessing operations such as normalization and standardization on the data in the training set to ensure the consistency of the input data and the stability of the model. Use the training set data to train the convolutional neural network model. Set appropriate hyperparameters, such as learning rate, batch size, number of iterations, etc. Early stopping can be used to prevent overfitting.

[0051] Choose appropriate loss functions (such as cross entropy loss) and optimizers (such as Adam, SGD) to guide the model training process.

[0052] S505 verifies the recognition model based on the verification set and uses the test set for testing.

[0053] Use the validation set data to verify the trained model and evaluate the model's performance. Common evaluation indicators include accuracy, precision, recall, and F1 score. Adjust the model's structure or hyperparameters based on the results of the validation set to improve the model's generalization ability. Use independent test set data to test the final model to further verify the model's performance on unseen data. The results of the test set can be used as the final evaluation criteria for model performance.

[0054] Through the above steps, an effective convolutional neural network model can be trained based on the marked historical abnormal signal feature data to automatically detect abnormal signals in the IoT environment. This process not only improves the accuracy and efficiency of anomaly detection, but also provides strong support for subsequent fault diagnosis and system optimization.

[0055] S105 deploys the trained convolutional neural network model into the monitoring system to monitor the radio signals of IoT devices in real time and automatically identify abnormal signals; Export the trained CNN model from the training framework (such as TensorFlow, PyTorch, etc.) to a deployable format. Common export formats include ONNX, TensorFlow Lite, etc. Optimize the model for the deployment environment (such as edge devices, cloud servers, etc.). Optimization methods include pruning, quantization, model compression, etc., to reduce the model's computational workload and memory usage, and improve inference speed.

[0056] Choose the appropriate hardware platform based on actual needs. Edge devices (such as Raspberry Pi, NVIDIA Jetson, etc.) are suitable for low-power, real-time applications; cloud servers are suitable for large-scale data processing and high-performance computing. Install and configure the necessary software environment, including operating system, dependent libraries, inference engine, etc. Ensure that all components are compatible and can run stably.

[0057] Configure sensors and receivers to ensure that they can collect radio signal data from IoT devices in real time. The frequency and format of data collection should be consistent with the model input requirements. Preprocess the collected radio signal data, including normalization, feature extraction, etc., to make it meet the model input requirements. The preprocessing steps should be consistent with the training phase to ensure data consistency and model accuracy.

[0058] Load the exported CNN model in the deployment environment. Ensure that the model can be loaded quickly and is ready for inference. Input the preprocessed data into the model for real-time inference. The model will output the abnormal probability or classification result for each data sample. Based on the results output by the model, determine whether the signal is abnormal. Thresholds can be set to decide when to trigger an alarm or take other actions.

[0059] S106 generates corresponding alarm information according to the severity of the abnormal signal and sends an alarm notification through a corresponding channel.

[0060] The specific steps include: S601 defines a severity level according to the characteristics of the abnormal signal.

[0061] Determine the key indicators used to evaluate the severity of abnormal signals. Common evaluation indicators include: Abnormality duration: The length of time that the abnormal signal lasts. Long-term abnormality indicates a more serious fault.

[0062] Abnormal amplitude: the degree of deviation of characteristics such as signal strength, frequency, and phase. Larger deviations usually indicate more serious abnormalities.

[0063] Abnormal frequency: The number of times abnormal signals occur per unit time. Frequent abnormalities indicate potential problems in the system.

[0064] Impact scope: The number of devices or areas affected by the abnormal signal. A wider impact scope indicates a more serious fault.

[0065] Define severity levels: Based on the above evaluation indicators, the severity of abnormal signals is divided into different levels, for example: Low level (Level 1): Minor abnormality that does not affect normal system operation. Observation is recommended.

[0066] Level 2: Moderate anomaly, which has a certain impact on system performance. It is recommended to investigate as soon as possible.

[0067] Level 3: Severe anomalies that cause system failure or service interruption and require immediate action.

[0068] S602 designs different templates for alarms of different severity, including necessary information such as alarm level, anomaly type, occurrence time, location, recommended measures, etc.; Design standardized templates for alarms of different severity levels to ensure the integrity and consistency of information. The template should include the following necessary information: Clearly mark the severity level of the alarm (such as Level 1, Level 2, Level 3). Describe the type of abnormal signal, such as abnormal signal strength, frequency drift, phase distortion, etc.

[0069] Record the timestamp when the abnormal signal is first detected. Indicate the geographical location or device number where the abnormal signal occurs.

[0070] S603 selects an appropriate delivery channel according to the severity of the alarm and the preference of the recipient; Select the appropriate alert delivery channel based on the severity of the alert and the recipient's preferences. Common delivery channels include: Email: Available for low and medium severity alerts, providing detailed alert information and attachments.

[0071] SMS: Applicable to medium and high-level alarms, providing brief alarm information to ensure timely notification.

[0072] Push Notifications: Available for all alert levels, quickly alert the right people via mobile app or desktop notifications.

[0073] Telephone: For advanced alarms, ensuring that the responsible person can be contacted immediately in an emergency.

[0074] It also allows recipients to set the method and frequency of receiving alerts according to their needs and preferences, for example, they can choose to only receive phone notifications for high-level alerts.

[0075] S604: After detecting an abnormal signal and evaluating its severity, a corresponding sending mechanism is triggered.

[0076] When the monitoring system detects an abnormal signal and assesses its severity, it triggers the corresponding alarm sending mechanism according to pre-defined rules. The sending time and recipient of each alarm are recorded for subsequent tracking and auditing.

[0077] S107 stores the monitored radio signal data and abnormality detection results in a database; The monitored radio signal data and anomaly detection results are stored in a database to ensure data security and accessibility.

[0078] Design a reasonable database table structure, including but not limited to the following fields: Signal ID: A number that uniquely identifies each signal.

[0079] Signal characteristics: signal strength, frequency, phase and other characteristic values.

[0080] Acquisition time: The acquisition timestamp of the signal data.

[0081] Device ID: The device ID of the collected signal.

[0082] Location information: The geographical location where the signal is collected.

[0083] Abnormal flag: whether an abnormal signal is detected.

[0084] Exception type: The type of exception signal.

[0085] Abnormal Severity: The severity level of the abnormal signal.

[0086] Processing status: The processing status of the exception signal (such as unprocessed, processing, processed).

[0087] Data backup: Back up the database regularly to prevent data loss.

[0088] S108 displays the anomaly detection results through a graphical interface.

[0089] Develop a user-friendly graphical interface to display anomaly detection results and system status. Display real-time radio signal data and anomaly detection results, including dynamic charts of signal strength, frequency, phase and other characteristics. Provide historical data query function, users can view signal data and anomaly records in any time period.

[0090] Through the above steps, a complete abnormal signal detection and processing system can be established, which can not only identify and respond to abnormal signals in a timely manner, but also provide support for long-term management and optimization of the system through data storage and visualization. This system not only improves the efficiency and accuracy of abnormal detection, but also enhances the reliability and security of the system.

[0091] Second embodiment The present invention provides an abnormal radio signal monitoring system based on artificial intelligence and big data, comprising a data acquisition module, a preprocessing module, a feature extraction module, a model generation module, a monitoring module and an alarm module: the data acquisition module is used to obtain the radio signal data of each device collected in real time by sensors and receivers distributed in an Internet of Things environment; the preprocessing module is used to preprocess the collected radio signal data; the feature extraction module is used to extract abnormal signal features from the preprocessed radio signal data using feature engineering technology; the model generation module is used to train a convolutional neural network model based on marked historical abnormal signal feature data; the monitoring module is used to deploy the trained convolutional neural network model into a monitoring system, monitor the radio signals of Internet of Things devices in real time, and automatically identify abnormal signals; the alarm module is used to generate corresponding alarm information according to the severity of the abnormal signal, and send an alarm notification through a corresponding channel.

[0092] In this embodiment, the data acquisition module collects data from various sensors and receivers widely distributed in the IoT environment. These devices can capture the radio signals emitted by surrounding devices in real time and provide raw data for subsequent analysis. The preprocessing module cleans and formats the collected data to improve the data quality and the efficiency of subsequent processing. The preprocessing process includes steps such as removing interference signals and correcting frequency offsets. The feature extraction module uses feature engineering technology to mine representative features from the preprocessed radio signals. These features are crucial for distinguishing normal signals from abnormal signals and help improve the accuracy of anomaly detection. The model generation module performs model training based on known abnormal signal samples. Through a large amount of historical data learning, the model can gradually master the unique pattern of abnormal signals, thereby preparing for future detection tasks. Once the model training is completed, it will be integrated into the monitoring system to achieve continuous monitoring of the radio signals of IoT devices. This process is automated and can quickly identify any signals that deviate from the normal pattern and promptly detect potential security threats or technical failures. When the monitoring module detects an abnormal signal, the alarm module will trigger the corresponding level of alarm mechanism based on the severity evaluation results of the signal. This includes but is not limited to sending emails, SMS reminders or reporting directly to the designated security management system. This way, those involved can take immediate action to prevent the problem from getting worse.

[0093] In summary, this abnormal radio signal monitoring system based on artificial intelligence and big data not only improves the response speed to security threats in the IoT environment, but also enhances the stability and reliability of the overall network.

[0094] What is disclosed above is only a preferred embodiment of the present invention, and it certainly cannot be used to limit the scope of rights of the present invention. Ordinary technicians in this field can understand that all or part of the processes of the above embodiment and equivalent changes made according to the claims of the present invention still fall within the scope of the invention.

Claims

1. An abnormal radio signal monitoring method based on artificial intelligence and big data, characterized in that: include: Obtain sensors and receivers distributed in the IoT environment to collect radio signal data from each device in real time; Preprocessing the collected radio signal data; Extract abnormal signal features from preprocessed radio signal data using feature engineering techniques; The convolutional neural network model is trained based on the marked historical abnormal signal feature data; Deploy the trained convolutional neural network model into the monitoring system to monitor the radio signals of IoT devices in real time and automatically identify abnormal signals; According to the severity of the abnormal signal, the corresponding alarm information is generated and the alarm notification is sent through the corresponding channel.

2. The abnormal radio signal monitoring method based on artificial intelligence and big data according to claim 1, characterized in that: After generating corresponding alarm information according to the severity of the abnormal signal and sending the alarm notification through the corresponding channel, the method also includes: storing the monitored radio signal data and the abnormality detection result in a database.

3. The abnormal radio signal monitoring method based on artificial intelligence and big data as claimed in claim 2 is characterized in that: After storing the monitored radio signal data and the abnormality detection results in a database, the method further includes displaying the abnormality detection results through a graphical interface.

4. The abnormal radio signal monitoring method based on artificial intelligence and big data as claimed in claim 3 is characterized in that: The radio signal data includes signal strength, frequency, bandwidth and modulation mode.

5. The abnormal radio signal monitoring method based on artificial intelligence and big data according to claim 4, characterized in that: The specific steps of obtaining the sensors and receivers distributed in the IoT environment to collect the radio signal data of each device in real time include: Configuring acquisition parameters of sensors and receivers, including sampling frequency, sampling interval, and signal strength threshold; Configure the receiver's network parameters; Start the sensor and receiver to start collecting radio signal data in real time; Cache the collected radio signal data.

6. The abnormal radio signal monitoring method based on artificial intelligence and big data as claimed in claim 5, characterized in that: The specific steps of preprocessing the collected radio signal data include: Collect radio signal data from different sources; Use FIR filters to remove background noise and other interference from radio signal data; Correcting phase distortion in radio signal data due to the transmission path; The corrected radio signal data is normalized.

7. The abnormal radio signal monitoring method based on artificial intelligence and big data according to claim 6, characterized in that: The specific steps of extracting abnormal signal features from the pre-processed radio signal data using feature engineering technology include: Calculate the mean, standard deviation and fluctuation range of signal strength in radio signal data; Calculate the mean, standard deviation, and drift rate of frequencies in radio signal data; Calculate the modulation error rate and phase error to obtain the abnormal signal characteristics.

8. The abnormal radio signal monitoring method based on artificial intelligence and big data according to claim 7, characterized in that: The specific steps of training the convolutional neural network model based on the marked historical abnormal signal feature data include: Collect normal signals and abnormal signals in historical radio signal data to obtain a data set; Label each data sample in the dataset to identify whether it is an abnormal signal; Extract all abnormal signal features of samples in the data set and divide them into training set, validation set and test set; Train the convolutional neural network model based on the training set to obtain a recognition model; The recognition model is verified based on the validation set and tested using the test set.

9. An abnormal radio signal monitoring system based on artificial intelligence and big data, adopting the abnormal radio signal monitoring method based on artificial intelligence and big data as claimed in claim 1, characterized in that: Including data acquisition module, preprocessing module, feature extraction module, model generation module, monitoring module and alarm module: The data acquisition module is used to acquire the radio signal data of each device collected in real time by sensors and receivers distributed in the IoT environment; The preprocessing module is used to preprocess the collected radio signal data; The feature extraction module is used to extract abnormal signal features from the pre-processed radio signal data using feature engineering technology; The model generation module is used to train a convolutional neural network model based on the marked historical abnormal signal feature data; The monitoring module is used to deploy the trained convolutional neural network model into the monitoring system, monitor the radio signals of IoT devices in real time, and automatically identify abnormal signals; The alarm module is used to generate corresponding alarm information according to the severity of the abnormal signal and send the alarm notification through the corresponding channel.

Images