A panoramic visualization data monitoring and analysis system for voice lines based on big data

Abstract

This invention discloses a panoramic visualization data monitoring and analysis system for voice lines based on big data, belonging to the field of information monitoring technology. The invention includes a voice device positioning module, a voice device radiation area extraction module, a voice transmission interference analysis module, a voice line risk analysis module, a voice line visualization monitoring module, and a voice line management module. The system workflow sequentially passes through a data acquisition stage, a real-time processing stage, a risk warning stage, and a visualization update mechanism to achieve visualized monitoring and analysis of voice lines. This invention tracks and locates voice devices, combines this with risk analysis of the voice device's radiation area, and constructs a three-dimensional scene to facilitate visualized monitoring of the voice device's communication lines. Simultaneously, it identifies and locates interference sources, and marks interference sources and faulty devices in the three-dimensional scene, reducing processing costs and improving the efficiency and accuracy of real-time panoramic mapping.

Description

Technical Field

[0001] This invention belongs to the field of information monitoring technology, and in particular relates to a panoramic visualization data monitoring and analysis system for voice lines based on big data. Background Technology

[0002] Existing voice line monitoring technology faces the following core problems:

[0003] (1) Limitations of noise processing technology: Existing noise reduction algorithms can only achieve a noise reduction effect of about 20dB in strong noise environments (such as 60dB highway noise), and it is difficult to distinguish between human voices and environmental noise;

[0004] (2) Physical propagation limitations: 50dB human voice is significantly attenuated at 30 meters away, and the energy approaches zero at 100 meters away. Long-distance monitoring requires high-cost laser microphones (unit price exceeds 10,000 yuan).

[0005] (3) Lack of real-time panoramic mapping: There is no voice line visualization system that supports multi-dimensional (time / frequency / spatial domain) synchronous refresh rate ≥10Hz;

[0006] (4) Line management complexity: Outbound trunk lines have a call quality fluctuation rate of more than 20% during peak hours, and the interference rate increases by 35% when multiple lines are running concurrently. Summary of the Invention

[0007] The purpose of this invention is to provide a panoramic visualization data monitoring and analysis system for voice lines based on big data. It tracks and locates voice devices, performs risk analysis by combining the radiation area of ​​the voice devices, and constructs a three-dimensional scene to facilitate the visualization monitoring of the communication lines of voice devices. This solves the problems of physical propagation limitations, high processing costs, and lack of real-time panoramic mapping in existing voice line monitoring technologies.

[0008] To solve the above technical problems, the present invention is achieved through the following technical solutions:

[0009] This invention is a panoramic visualization data monitoring and analysis system for voice lines based on big data, including a voice device positioning module, a voice device radiation area extraction module, a voice transmission interference analysis module, a voice line risk analysis module, a voice line visualization monitoring module, and a voice line management module.

[0010] The voice device positioning module is used to deploy multimodal positioning beacons in the monitoring area, establish a signal strength fingerprint database, and output the three-dimensional coordinates of the voice device.

[0011] The voice device radiation area extraction module is used to construct a propagation loss model, generate a heat map coverage area, and dynamically delineate the effective communication boundary.

[0012] The voice transmission interference analysis module is used to monitor the signal-to-noise ratio in real time, build an interference source feature library, identify the type of interference, and locate the interference source.

[0013] The voice line risk analysis module is used to construct a risk assessment index system, which specifically includes link quality score, fault probability prediction and security threat detection.

[0014] The voice line visualization monitoring module is used to construct a three-dimensional scene of the voice equipment line and to perform AR annotation on faulty equipment in the three-dimensional scene;

[0015] The voice line management module is used for automatic routing and load balancing of voice lines.

[0016] As a preferred technical solution, the system includes the following steps:

[0017] Step S1: Model the scene area and deploy multimodal positioning beacons, and update the coordinates of different voice devices in the 3D scene in real time;

[0018] Step S2: Use a 5G spectrum analyzer to collect electromagnetic field strength data, construct a propagation loss model, and generate a heat map coverage area;

[0019] Step S3: Monitor the signal-to-noise ratio in real time to identify and locate the types of interference during transmission;

[0020] Step S4: Construct an evaluation index system to assess link quality, failure probability, and security threats;

[0021] Step S5: Construct a 3D scene of the site and perform AR fault annotation on the faulty equipment in the scene;

[0022] Step S6: Set up intelligent operation and maintenance strategies to automatically switch routes and balance load.

[0023] As a preferred technical solution, the multimodal positioning beacon is deployed in physical devices or technical nodes in the positioning environment. By transmitting multiple signals and cooperating with sensors in the environment, it constructs a positioning network covering complex indoor and outdoor scenarios. After receiving the signals, the terminal measures the distance or angle difference and calculates the position coordinates by combining the coordinates of multiple beacons.

[0024] When using Bluetooth RSSI ranging, the distance is estimated using the Received Signal Strength Indicator (RSSI) formula:

[0025] ;

[0026] In the formula, The distance between the terminal and the beacon. The reference signal strength is 1 meter from the beacon. Path loss index;

[0027] Given the coordinates of 3 beacons and distance measurement results Solve the system of equations:

[0028] ;

[0029] The terminal coordinates are solved using the least squares method. ;

[0030] Calculate the signal incident angle using the phase difference. ,distance Coordinates:

[0031] ;

[0032] The time difference of signal arrival at multiple base stations Establish the equation of the hyperbola:

[0033] ;

[0034] In the formula, To achieve the speed of light, at least four base stations are needed for three-dimensional positioning.

[0035] The process for establishing the signal strength fingerprint database is as follows:

[0036] Step S11: Model the site and deploy reference points; the deployment includes: regional network division, reference point coordinate labeling, and hotspot distribution mapping; regional network division is used to divide the positioning area into 1m*1m grid points; reference point coordinate labeling is used to record the coordinates of the center point of each grid, and hotspot distribution mapping is used to label the physical location of all APs;

[0037] Step S12: Mobile terminals supporting Wi-Fi scanning collect the MAC address and signal strength of the AP at the current coordinates; collect ≥30 sets of data for each reference point to eliminate instantaneous fluctuations, and collect data from multiple directions (0° / 90° / 180° / 270°) to improve robustness;

[0038] Step S13: Preprocess the data collected by the mobile terminal; preprocessing includes outlier removal, data smoothing, and feature vector construction;

[0039] Step S14: Construct a fingerprint database from the preprocessed data; the fingerprint database includes X coordinates, Y coordinates, MAC address of the AP, mean RSSI and standard deviation of RSSI for that point;

[0040] Step S15: Obtain the RSSI values ​​of each AP at the current location, and perform similarity calculation and location estimation.

[0041] As a preferred technical solution, in step S15, when obtaining the RSSI values ​​of each AP at the current location for fingerprint matching and positioning, the RSSI values ​​of each AP at the current location are collected for similarity calculation, and the specific formula is as follows: ;

[0042] Or weighted Euclidean distance: ;

[0043] In the formula, This is the total distance value, used to quantify the similarity between the real-time signal and the reference signal in the fingerprint database; For the fingerprint database stored as the first Signal strength reference value for each AP, For real-time measurement of the first Signal strength value of each AP, For the first The weighting coefficients of each AP;

[0044] Then, the average coordinates of the k reference points with the smallest distance are taken, and the calculation formula is as follows:

[0045] ;

[0046] In the formula, These represent the estimated x and y coordinates of the target point, respectively. The neighbor index is used to traverse the k nearest neighbors. They represent the first The actual values ​​of the x and y coordinates of each neighbor.

[0047] As a preferred technical solution, the specific steps for the voice device radiation area extraction module to construct the propagation loss model are as follows:

[0048] Step S21: Set the frequency (1000-2300MHz), Base Station Height (30-200m), height of the mobile platform (1-10m) and distance Basic parameters (1-20km);

[0049] Step S22: Determine the environment type and perform basic path calculation; the calculation formula is as follows:

[0050]

[0051] Step S23: Calculate correction terms based on different environment types ;

[0052] Small and medium-sized cities (model=1): ;

[0053] Low frequency in large cities (model=2): ;

[0054] High frequency in large cities (model=3): Different correction factors are set to adapt to various environments such as urban areas, suburbs, and rural areas, while also taking into account the influence of timing factors such as building density and antenna height.

[0055] Step S24: The calculation is encapsulated as a MATLAB function, with the path loss value directly output from the input parameters;

[0056] Step S25: Compare the actual measurement data with the model prediction results to improve the accuracy of parameter adjustment.

[0057] As a preferred technical solution, the workflow of the voice transmission interference analysis module in constructing the interference source feature library is as follows:

[0058] Step S31: Use an FPGA+RF chip architecture (such as Virtex-7 FPGA+AD9361 RF chip) to realize signal acquisition and processing, normalize and denoise the acquired interference signals, and use the PyWavelets library to implement signal decomposition.

[0059] Step S32: Calculate the wavelet coefficient energy at each scale using a multi-scale spatial energy distribution feature extraction method. In the formula, For the first Layer wavelet coefficients, and select wavelet basis functions for time-frequency analysis;

[0060] Step S33: Establish a six-dimensional feature space to store the interference pattern, using a hierarchical storage structure: interference type, wavelet scale, and feature vector;

[0061] Step S34: Select the most discriminative feature combination from the six-dimensional feature space, and use the recursive feature elimination method to reduce dimensionality and improve computational efficiency;

[0062] Step S35: Use the Grey Wolf Optimization Algorithm to optimize the penalty coefficient and kernel function parameters of the SVM, select the RBF kernel function to handle nonlinear classification problems, and train a multi-class SVM to distinguish five types of interference: co-channel, adjacent channel, out-of-band, intermodulation and blocking interference.

[0063] As a preferred technical solution, the specific process for establishing an evaluation index system in the voice line risk analysis module is as follows:

[0064] Step S41: Collect voice signal samples (16kHz sampling rate, 16-bit precision), simultaneously record network metrics (jitter ≤20ms, packet loss rate ≤1%), and set up the subjective evaluation module;

[0065] The scoring criteria used are as follows:

[0066] 5 points: Excellent voice quality; 4 points: Good call quality (VoIP industry standard); 3 points: Acceptable call quality; 2 points: Poor call quality; 1 point: Unable to communicate; Able to monitor voice transmission quality degradation in real time and identify quality problems such as encoding distortion and network jitter;

[0067] Step S42: Obtain time-domain features, frequency-domain features, and network features; time-domain features include short-time energy, zero-crossing rate, and harmonic-to-noise ratio; frequency-domain features are MFCC coefficients; network features include the number of SIP signaling anomalies and the RTP packet out-of-order rate.

[0068] Step S43: Train a fault probability prediction model based on the features;

[0069] Step S44: Use a sliding window to generate time series samples and define fault labels to validate the trained fault probability prediction model;

[0070] Step S45: Establish a dedicated rule base for voice and a real-time traffic mirroring generation and detection engine;

[0071] Step S46: Filter the quintuples to generate a behavioral analysis model;

[0072] During security testing, the Voice Line Security Threat Detection (IPS) system is an intelligent protection system designed for VoIP communication. Its core consists of three parts: a protocol analysis engine, behavioral baseline modeling, and a real-time response mechanism. The system uses Deep Packet Inspection (DPI) technology to analyze the SIP / RTP protocol, identifying 12 types of attack characteristics such as malformed INVITE packets and authentication flooding (detection rate ≥99.5%). Simultaneously, it uses an LSTM time-series model to establish behavioral baselines such as call frequency and packet length distribution, enabling the detection of minor anomalies caused by zero-day attacks (false positive rate <0.1%). The system deployment uses an Bypass mirror architecture, supporting 1Gbps line-rate processing. When a threat is detected, it triggers a three-level response: low-risk alarm (e.g., SSRC conflict), medium-risk rate limiting (e.g., >50 REGISTERs per second), and high-risk blocking (e.g., malicious payload injection), and automatically generates an ATT&CK matrix to visualize the attack path.

[0073] As a preferred technical solution, the voice line risk analysis module performs dynamic risk scoring based on the voice line risk index. Specifically, the final weight is determined by setting the maximum eigenvalue and consistency ratio. The index variance is recalculated hourly, and the objective weights are updated using the coefficient of variation method. The specific formula for fusing subjective and objective weights is as follows:

[0074] ;

[0075] In the formula, For the first The overall weight of each indicator The first step is to calculate the result using the layering analysis method. The subjective weight of each indicator, For the first The objective volatility quantification value of each indicator. Set the value to 0.6; generate risk indices from level 1 to 5 (corresponding to five-color warnings: red, orange, yellow, blue, and green), and support risk source analysis (visualization of the contribution of each indicator).

[0076] As a preferred technical solution, when the voice line visualization monitoring module performs AR fault annotation, it first establishes the basic AR environment by initializing the AR session through ARKit (iOS) or ARCore (Android), configuring camera permissions, and creating a 3D scene coordinate system. Next, it performs device identification and positioning by using computer vision algorithms to identify feature points of real-world devices, accurately aligning the 3D model with the physical device, and establishing a spatial mapping relationship. Then, it implements the annotation function by overlaying annotation elements such as text, icons, or outlines at the identification location using the drawing interface provided by the AR engine (e.g., AREngine.drawAnnotation()), supporting color grading and dynamic content updates. Finally, it constructs a data closed loop by associating and storing the annotation information with real-time IoT data (such as temperature and vibration values), enabling cloud synchronization and multi-user collaborative viewing, while also supporting historical record backtracking and maintenance plan push notifications.

[0077] As a preferred technical solution, when the voice line management module performs automatic routing switching, it centrally manages network devices through the OpenFlow protocol, detects link quality in real time, and dynamically calculates the optimal path based on the global topology view. When the performance of the primary path deteriorates, the controller issues flow table instructions within seconds to switch to the backup link, ensuring service continuity. When performing load balancing, it uses the Q-learning algorithm to optimize traffic allocation, constructs a Q-table to record the action value function under different link states, and iteratively updates the Q value through the Bellman equation.

[0078] Implement equipment lifecycle management covering the entire lifecycle from planning to scrapping: Collect equipment operation data through IoT sensors, train AI models by combining historical fault databases, identify hardware degradation trends in advance and trigger early warnings; automatically schedule inspection, firmware upgrade or spare parts replacement tasks based on equipment health status to reduce manual intervention; integrate CMMS system to realize digital management of equipment ledgers, maintenance records and decommissioning assessments, and optimize resource utilization.

[0079] The present invention has the following beneficial effects:

[0080] (1) This invention uses voice devices for tracking and positioning, combines the radiation area of ​​voice devices for risk analysis, and constructs a three-dimensional scene to facilitate the visual monitoring of the communication lines of voice devices; at the same time, it identifies and positions interference sources, and marks interference sources and faulty devices in the three-dimensional scene to reduce processing costs and improve the efficiency and accuracy of real-time panoramic mapping.

[0081] (2) The present invention achieves unified positioning by standardizing the fingerprint database through the RSSI differences of different terminals, supports dynamic updating of fingerprint data, and can cope with changes in scenarios such as new Wi-Fi hotspots and furniture layout changes. When new APs are added or the environment changes, local fingerprint data is updated through crowdsourcing, and Gaussian process regression (GPR) is used to predict the signal distribution of unsampled points, thereby achieving rapid positioning of different voice devices and improving positioning efficiency.

[0082] (3) This invention extracts radiation area information, collects electromagnetic field strength data, constructs a propagation loss model, accurately calculates signal strength attenuation at different locations, determines effective coverage boundaries, provides a theoretical basis for base station layout and network planning, and sets different correction factors to adapt to various environments such as cities, suburbs, and rural areas.

[0083] (4) This invention uses a three-dimensional scene to perform visual monitoring, creates topology maps, spectrum maps, and alarm dashboards, identifies and locates equipment, uses computer vision algorithms to identify feature points of real equipment, accurately aligns the 3D model with the physical equipment, establishes a spatial mapping relationship, realizes cloud synchronization and multi-person collaborative viewing, and supports historical record backtracking and maintenance plan push. The whole process requires the combination of 3D target detection technology to accurately locate the coordinates of faulty parts, and ensures the correct spatial hierarchy relationship between virtual annotations and real objects through depth testing.

[0084] (5) This invention centrally manages network devices through the OpenFlow protocol, detects link quality (packet loss rate, latency, etc.) in real time, and dynamically calculates the optimal path in combination with the global topology view. When the performance of the primary path deteriorates, the controller issues flow table instructions within seconds to switch to the backup link, ensuring service continuity.

[0085] Of course, any product implementing this invention does not necessarily need to achieve all of the advantages described above at the same time. Attached Figure Description

[0086] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the following briefly introduces the drawings required for describing the embodiments. 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 creative work.

[0087] Figure 1 This is a schematic diagram of the structure of a panoramic visualization data monitoring and analysis system for voice lines based on big data, according to the present invention. Detailed Implementation

[0088] The following will clearly and completely describe the technical solutions in the embodiments of the present invention in conjunction with the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making any creative efforts shall fall within the scope of protection of the present invention.

[0089] Furthermore, the technical features involved in the various embodiments of the present invention described below can be combined with each other as long as they do not conflict with each other.

[0090] In order to make the purpose, technical solutions and advantages of this application more clear, the following Figure 1 It should be understood that the specific embodiments described herein are only used to explain the present application and are not intended to limit the present application.

[0091] Please see Figure 1 As shown, the present invention is a panoramic visualization data monitoring and analysis system for voice lines based on big data, including a voice device positioning module, a voice device radiation area extraction module, a voice transmission interference analysis module, a voice line risk analysis module, a voice line visualization monitoring module, and a voice line management module.

[0092] The voice device positioning module is used to deploy multimodal positioning beacons in the monitored area, establish a signal strength fingerprint database, and output the three-dimensional coordinates of the voice device;

[0093] The voice device radiation area extraction module is used to construct a propagation loss model, generate a heat map coverage area, and dynamically delineate the effective communication boundary.

[0094] The voice transmission interference analysis module is used to monitor the signal-to-noise ratio in real time, build an interference source feature library, identify interference types, and locate interference sources.

[0095] The voice line risk analysis module is used to construct a risk assessment indicator system, which specifically includes link quality scoring, failure probability prediction, and security threat detection.

[0096] The voice line visualization monitoring module is used to construct a three-dimensional scene of the voice equipment line and to perform AR annotation on faulty equipment in the three-dimensional scene;

[0097] The voice line management module is used for automatic routing and load balancing of voice lines.

[0098] The system includes the following steps:

[0099] Step S1: Model the scene area and deploy multimodal positioning beacons, and update the coordinates of different voice devices in the 3D scene in real time;

[0100] Step S2: Use a 5G spectrum analyzer to collect electromagnetic field strength data, construct a propagation loss model, and generate a heat map coverage area;

[0101] Step S3: Monitor the signal-to-noise ratio in real time to identify and locate the types of interference during transmission;

[0102] Step S4: Construct an evaluation index system to assess link quality, failure probability, and security threats;

[0103] Step S5: Construct a 3D scene of the site and perform AR fault annotation on the faulty equipment in the scene;

[0104] Step S6: Set up intelligent operation and maintenance strategies to automatically switch routes and balance load.

[0105] Multimodal positioning beacons are physical devices or technical nodes deployed in the positioning environment. They transmit multiple signals and cooperate with sensors in the environment to build a positioning network covering complex indoor and outdoor scenarios. After receiving the signals, the terminal measures the distance or angle difference and calculates the position coordinates by combining the coordinates of multiple beacons.

[0106] When using Bluetooth RSSI ranging, the distance is estimated using the Received Signal Strength Indicator (RSSI) formula:

[0107] ;

[0108] In the formula, The distance between the terminal and the beacon. The reference signal strength is 1 meter from the beacon. Path loss index;

[0109] Given the coordinates of 3 beacons and distance measurement results Solve the system of equations:

[0110] ;

[0111] The terminal coordinates are solved using the least squares method. ;

[0112] Calculate the signal incident angle using the phase difference. ,distance Coordinates:

[0113] ;

[0114] The time difference of signal arrival at multiple base stations Establish the equation of the hyperbola:

[0115] ;

[0116] In the formula, To achieve the speed of light, at least four base stations are needed for three-dimensional positioning.

[0117] The process for establishing a signal strength fingerprint database is as follows:

[0118] Step S11: Model the site and deploy reference points; the deployment includes: regional network division, reference point coordinate labeling, and hotspot distribution mapping; regional network division is used to divide the positioning area into 1m*1m grid points; reference point coordinate labeling is used to record the coordinates of the center point of each grid, and hotspot distribution mapping is used to label the physical location of all APs;

[0119] Step S12: Mobile terminals supporting Wi-Fi scanning collect the MAC address and signal strength of the AP at the current coordinates; collect ≥30 sets of data for each reference point to eliminate instantaneous fluctuations, and collect data from multiple directions (0° / 90° / 180° / 270°) to improve robustness;

[0120] Step S13: Preprocess the data collected by the mobile terminal; preprocessing includes outlier removal, data smoothing, and feature vector construction;

[0121] Step S14: Construct a fingerprint database from the preprocessed data; the fingerprint database includes X coordinates, Y coordinates, MAC address of the AP, mean RSSI and standard deviation of RSSI for that point;

[0122] Step S15: Obtain the RSSI values ​​of each AP at the current location, and perform similarity calculation and location estimation.

[0123] In step S15, when obtaining the RSSI values ​​of each AP at the current location for fingerprint matching and localization, the RSSI values ​​of each AP at the current location are collected for similarity calculation. The specific formula is as follows: ;

[0124] Or weighted Euclidean distance: ;

[0125] In the formula, This is the total distance value, used to quantify the similarity between the real-time signal and the reference signal in the fingerprint database; For the fingerprint database stored as the first Signal strength reference value for each AP, For real-time measurement of the first Signal strength value of each AP, For the first The weighting coefficients of each AP;

[0126] Then, the average coordinates of the k reference points with the smallest distance are taken, and the calculation formula is as follows:

[0127] ;

[0128] In the formula, These represent the estimated x and y coordinates of the target point, respectively. The neighbor index is used to traverse the k nearest neighbors. They represent the first The actual values ​​of the x and y coordinates of each neighbor.

[0129] The specific steps for constructing a propagation loss model using the voice device radiation area extraction module are as follows:

[0130] Step S21: Set the frequency (1000-2300MHz), Base Station Height (30-200m), height of the mobile platform (1-10m) and distance Basic parameters (1-20km);

[0131] Step S22: Determine the environment type and perform basic path calculation; the calculation formula is as follows:

[0132]

[0133] Step S23: Calculate correction terms based on different environment types ;

[0134] Small and medium-sized cities (model=1): ;

[0135] Low frequency in large cities (model=2): ;

[0136] High frequency in large cities (model=3): Different correction factors are set to adapt to various environments such as urban areas, suburbs, and rural areas, while also taking into account the influence of timing factors such as building density and antenna height.

[0137] Step S24: The calculation is encapsulated as a MATLAB function, with the path loss value directly output from the input parameters;

[0138] Step S25: Compare the actual measurement data with the model prediction results to improve the accuracy of parameter adjustment.

[0139] The workflow for the voice transmission interference analysis module to build an interference source feature library is as follows:

[0140] Step S31: Use an FPGA+RF chip architecture (such as Virtex-7 FPGA+AD9361 RF chip) to realize signal acquisition and processing, normalize and denoise the acquired interference signals, and use the PyWavelets library to implement signal decomposition.

[0141] Step S32: Calculate the wavelet coefficient energy at each scale using a multi-scale spatial energy distribution feature extraction method. In the formula, For the first Layer wavelet coefficients, and select wavelet basis functions for time-frequency analysis;

[0142] Step S33: Establish a six-dimensional feature space to store the interference pattern, using a hierarchical storage structure: interference type, wavelet scale, and feature vector;

[0143] Step S34: Select the most discriminative feature combination from the six-dimensional feature space, and use the recursive feature elimination method to reduce dimensionality and improve computational efficiency;

[0144] Step S35: Use the Grey Wolf Optimization Algorithm to optimize the penalty coefficient and kernel function parameters of the SVM, select the RBF kernel function to handle nonlinear classification problems, and train a multi-class SVM to distinguish five types of interference: co-channel, adjacent channel, out-of-band, intermodulation and blocking interference.

[0145] The specific process for establishing an evaluation index system for the voice line risk analysis module is as follows:

[0146] Step S41: Collect voice signal samples (16kHz sampling rate, 16-bit precision), simultaneously record network metrics (jitter ≤20ms, packet loss rate ≤1%), and set up the subjective evaluation module;

[0147] The scoring criteria used are as follows:

[0148] 5 points: Excellent voice quality; 4 points: Good call quality (VoIP industry standard); 3 points: Acceptable call quality; 2 points: Poor call quality; 1 point: Unable to communicate; Able to monitor voice transmission quality degradation in real time and identify quality problems such as encoding distortion and network jitter;

[0149] Step S42: Obtain time-domain features, frequency-domain features, and network features; time-domain features include short-time energy, zero-crossing rate, and harmonic-to-noise ratio; frequency-domain features are MFCC coefficients; network features include the number of SIP signaling anomalies and the RTP packet out-of-order rate.

[0150] Step S43: Train a fault probability prediction model based on the features;

[0151] Step S44: Use a sliding window to generate time series samples and define fault labels to validate the trained fault probability prediction model;

[0152] Step S45: Establish a dedicated rule base for voice and a real-time traffic mirroring generation and detection engine;

[0153] Step S46: Filter the quintuples to generate a behavioral analysis model;

[0154] During security testing, the Voice Line Security Threat Detection (IPS) system is an intelligent protection system designed for VoIP communication. Its core consists of three parts: a protocol analysis engine, behavioral baseline modeling, and a real-time response mechanism. The system uses Deep Packet Inspection (DPI) technology to analyze the SIP / RTP protocol, identifying 12 types of attack characteristics such as malformed INVITE packets and authentication flooding (detection rate ≥99.5%). Simultaneously, it uses an LSTM time-series model to establish behavioral baselines such as call frequency and packet length distribution, enabling the detection of minor anomalies caused by zero-day attacks (false positive rate <0.1%). The system deployment uses an Bypass mirror architecture, supporting 1Gbps line-rate processing. When a threat is detected, it triggers a three-level response: low-risk alarm (e.g., SSRC conflict), medium-risk rate limiting (e.g., >50 REGISTERs per second), and high-risk blocking (e.g., malicious payload injection), and automatically generates an ATT&CK matrix to visualize the attack path.

[0155] The voice line risk analysis module dynamically scores risk based on the voice line risk index, specifically by setting the maximum eigenvalue and consistency ratio to determine the final weights. The indicator variance is recalculated hourly, and the objective weights are updated using the coefficient of variation method. The specific formula for fusing subjective and objective weights is as follows:

[0156] ;

[0157] In the formula, For the first The overall weight of each indicator The first step is to calculate the result using the layering analysis method. The subjective weight of each indicator, For the first The objective volatility quantification value of each indicator. Set the value to 0.6; generate risk indices from level 1 to 5 (corresponding to five-color warnings: red, orange, yellow, blue, and green), and support risk source analysis (visualization of the contribution of each indicator).

[0158] When the voice line visualization monitoring module performs AR fault annotation, it first establishes the basic AR environment by initializing the AR session through ARKit (iOS) or ARCore (Android), configuring camera permissions, and creating a 3D scene coordinate system. Next, it performs device identification and positioning, using computer vision algorithms to identify feature points of real-world devices, precisely aligning the 3D model with the physical device, and establishing a spatial mapping relationship. Then, it implements the annotation function, using the drawing interface provided by the AR engine (such as AREngine.drawAnnotation()) to overlay annotation elements such as text, icons, or outlines at the identified location, supporting color grading and dynamic content updates. Finally, it constructs a data loop, associating and storing the annotation information with real-time IoT data (such as temperature and vibration values), enabling cloud synchronization and multi-user collaborative viewing, while also supporting historical record backtracking and maintenance plan push notifications.

[0159] When the voice line management module performs automatic route switching, it centrally manages network devices through the OpenFlow protocol, detects link quality in real time, and dynamically calculates the optimal path based on the global topology view. When the performance of the primary path deteriorates, the controller issues flow table instructions within seconds to switch to the backup link to ensure service continuity. When performing load balancing, the Q-learning algorithm is used to optimize traffic allocation, and a Q-table is constructed to record the action value function under different link states. The Q value is iteratively updated through the Bellman equation.

[0160] Implement equipment lifecycle management covering the entire lifecycle from planning to scrapping: Collect equipment operation data through IoT sensors, train AI models by combining historical fault databases, identify hardware degradation trends in advance and trigger early warnings; automatically schedule inspection, firmware upgrade or spare parts replacement tasks based on equipment health status to reduce manual intervention; integrate CMMS system to realize digital management of equipment ledgers, maintenance records and decommissioning assessments, and optimize resource utilization.

[0161] It is worth noting that the various units included in the above system embodiments are only divided according to functional logic, but are not limited to the above division, as long as the corresponding functions can be achieved; in addition, the specific names of each functional unit are only for easy differentiation and are not used to limit the scope of protection of the present invention.

[0162] Furthermore, those skilled in the art will understand that all or part of the steps in the methods of the above embodiments can be implemented by a program instructing related hardware, and the corresponding program can be stored in a computer-readable storage medium.

[0163] The preferred embodiments of the present invention disclosed above are merely illustrative of the invention. These preferred embodiments do not exhaustively describe all details, nor do they limit the invention to the specific implementations described. Clearly, many modifications and variations can be made based on the content of this specification. This specification selects and specifically describes these embodiments to better explain the principles and practical applications of the invention, thereby enabling those skilled in the art to better understand and utilize the invention. The invention is limited only by the claims and their full scope and equivalents.

Claims

1. A panoramic visualization data monitoring and analysis system for voice lines based on big data, comprising a voice equipment positioning module, a voice equipment radiation area extraction module, a voice transmission interference analysis module, a voice line risk analysis module, a voice line visualization monitoring module, and a voice line management module, characterized in that: The voice device positioning module is used to deploy multimodal positioning beacons in the monitoring area, establish a signal strength fingerprint database, and output the three-dimensional coordinates of the voice device. The voice device radiation area extraction module is used to construct a propagation loss model, generate a heat map coverage area, and dynamically delineate the effective communication boundary. The voice transmission interference analysis module is used to monitor the signal-to-noise ratio in real time, build an interference source feature library, identify the type of interference, and locate the interference source. The voice line risk analysis module is used to construct a risk assessment index system, which specifically includes link quality score, fault probability prediction and security threat detection. The voice line visualization monitoring module is used to construct a three-dimensional scene of the voice equipment line and to perform AR annotation on faulty equipment in the three-dimensional scene; The voice line management module is used for automatic routing and load balancing of voice lines. The workflow for the voice transmission interference analysis module to build the interference source feature library is as follows: Step S31: Normalize and denoise the collected interference signal, and use the PyWavelets library to decompose the signal; Step S32: Using a multi-scale spatial energy distribution feature extraction method, calculate the energy of wavelet coefficients at each scale, and select wavelet basis functions for time-frequency analysis; Step S33: Establish a six-dimensional feature space to store the interference pattern; Step S34: Select the most discriminative feature combination from the six-dimensional feature space and reduce the dimensionality using the recursive feature elimination method; Step S35: Use the Grey Wolf Optimization Algorithm to optimize the penalty coefficient and kernel function parameters of the SVM, and train a multi-class SVM to distinguish five types of interference.

2. The panoramic visualization data monitoring and analysis system for voice lines based on big data as described in claim 1, characterized in that, The system performs the following steps: Step S1: Model the scene area and deploy multimodal positioning beacons, and update the coordinates of different voice devices in the 3D scene in real time; Step S2: Use a 5G spectrum analyzer to collect electromagnetic field strength data, construct a propagation loss model, and generate a heat map coverage area; Step S3: Monitor the signal-to-noise ratio in real time to identify and locate the types of interference during transmission; Step S4: Construct an evaluation index system to assess link quality, failure probability, and security threats; Step S5: Construct a 3D scene of the site and perform AR fault annotation on the faulty equipment in the scene; Step S6: Set up intelligent operation and maintenance strategies to automatically switch routes and balance load.

3. The panoramic visualization data monitoring and analysis system for voice lines based on big data as described in claim 1, characterized in that, The multimodal positioning beacons are deployed as physical devices or technical nodes in the positioning environment. By emitting multiple signals and cooperating with sensors in the environment, they construct a positioning network covering complex indoor and outdoor scenarios. After receiving the signals, the terminal measures the distance or angle difference and calculates the position coordinates by combining the coordinates of multiple beacons. The signal strength fingerprint database establishment process is as follows: Step S11: Model the site and deploy reference points; Step S12: Mobile terminals that support Wi-Fi scanning collect the MAC address and signal strength of the AP at the current coordinates; Step S13: Preprocess the data collected by the mobile terminal; Step S14: Construct a fingerprint database from the preprocessed data; Step S15: Obtain the RSSI values ​​of each AP at the current location, and perform similarity calculation and location estimation.

4. The panoramic visualization data monitoring and analysis system for voice lines based on big data as described in claim 3, characterized in that, In step S15, when obtaining the RSSI values ​​of each AP at the current location for fingerprint matching and positioning, the RSSI values ​​of each AP at the current location are collected for similarity calculation. The specific formula is as follows: ； Or weighted Euclidean distance: ; In the formula, This is the total distance value, used to quantify the similarity between the real-time signal and the reference signal in the fingerprint database; For the fingerprint database stored as the first Signal strength reference value for each AP, For real-time measurement of the first Signal strength value of each AP, For the first The weighting coefficients of each AP; Then, the average coordinates of the k reference points with the smallest distance are taken, and the calculation formula is as follows: ; Where, These represent the estimated x and y coordinates of the target point, respectively. The neighbor index is used to traverse the k nearest neighbors. They represent the first The actual values ​​of the x and y coordinates of each neighbor.

5. The panoramic visualization data monitoring and analysis system for voice lines based on big data as described in claim 1, characterized in that, The specific steps for constructing the propagation loss model using the voice device radiation area extraction module are as follows: Step S21: Set the frequency Base station height Mobile station height and distance Basic parameters; Step S22: Determine the environment type and perform basic path calculation; Step S23: Calculate the correction terms according to different environment types; Step S24: The calculation is encapsulated as a MATLAB function, with the path loss value directly output from the input parameters; Step S25: Compare the actual measurement data with the model prediction results.

6. The panoramic visualization data monitoring and analysis system for voice lines based on big data according to claim 1, characterized in that, The specific process for establishing the evaluation index system in the voice line risk analysis module is as follows: Step S41: Collect voice signal samples, record network indicators synchronously, and set up the subjective evaluation module; Step S42: Obtain time-domain features, frequency-domain features, and network features; Step S43: Train a fault probability prediction model based on the features; Step S44: Generate timing samples using a sliding window and define fault labels; Step S45: Establish a dedicated rule base for voice and a real-time traffic mirroring generation and detection engine; Step S46: Filter the quintuples to generate a behavior analysis model.

7. The panoramic visualization data monitoring and analysis system for voice lines based on big data as described in claim 1, characterized in that, The voice line risk analysis module dynamically scores the risk based on the voice line risk index, specifically by setting the maximum eigenvalue and consistency ratio to determine the final weight. The indicator variance is recalculated hourly, and the objective weights are updated using the coefficient of variation method. The specific formula for fusing subjective and objective weights is as follows: ； In the formula, For the first The combined weight of each indicator The first step is to calculate the result using the layering analysis method. The subjective weight of each indicator, For the first The objective volatility quantification value of each indicator. Take 0.

6.

8. The panoramic visualization data monitoring and analysis system for voice lines based on big data according to claim 1, characterized in that, When the voice line visualization monitoring module performs AR fault annotation, it first establishes the basic AR environment by initializing the AR session through ARKit or ARCore, configuring camera permissions, and creating a 3D scene coordinate system. Next, it performs device identification and positioning by using computer vision algorithms to identify feature points of real-world devices, precisely aligning the 3D model with the physical device, and establishing a spatial mapping relationship. Then, it implements the annotation function by overlaying annotation elements such as text, icons, or outlines at the identified location through the drawing interface provided by the AR engine, supporting color grading and dynamic content updates. Finally, it constructs a data loop by associating and storing the annotation information with real-time IoT data, enabling cloud synchronization and multi-user collaborative viewing, while also supporting historical record backtracking and maintenance plan push notifications.

9. A panoramic visualization data monitoring and analysis system for voice lines based on big data as described in claim 1, characterized in that, When the voice line management module performs automatic route switching, it centrally manages network devices through the OpenFlow protocol, detects link quality in real time, and dynamically calculates the optimal path in combination with the global topology view; when the performance of the primary path deteriorates, the controller issues flow table instructions in seconds to switch to the backup link. When performing load balancing, the Q-learning algorithm is used to optimize traffic allocation, a Q-table is constructed to record the action value function under different link states, and the Q value is updated iteratively through the Bellman equation.

Images