Remote monitoring method and system for broadcast television equipment based on hybrid wireless network

Abstract

The application relates to the technical field of broadcast television communication, and particularly discloses a broadcast television equipment remote monitoring method and system based on a hybrid wireless network, which comprises the following steps: synchronously collecting tower body structure safety parameters and equipment electrical parameters of a broadcast television transmitting tower through a sensor node; transmitting the collected various parameters by using a hybrid network transmission path; in the transmission process, dynamically selecting a transmission path based on network signal strength and parameter priority; performing time domain trend prediction and frequency domain anomaly detection on the transmitted various parameters; triggering different levels of alarms based on the detection results; and generating inspection or maintenance tasks for key parts of the broadcast television transmitting tower body and equipment according to the detection results. The application can realize high-precision monitoring, can early warn structure slowly changing hidden dangers, and can significantly improve operation and maintenance efficiency and safety.

Description

Technical Field

[0001] This invention relates to the field of broadcast television communication technology, and more specifically to a method and system for remote monitoring of broadcast television equipment based on a hybrid wireless network. Background Technology

[0002] Broadcast and television transmission towers, relay stations, and satellite receiving stations are the cornerstone of signal coverage and emergency communications. These towering structures (such as freestanding towers, monotube towers, and masts) are subject to long-term environmental erosion from wind, rain, temperature differences, icing, and strong ultraviolet radiation, making them prone to safety hazards such as tower tilting, foundation settlement, loose bolts, abnormal guy wire tension, structural deformation, and increased grounding resistance. Furthermore, their internal transmitters, power supply units, antennas, and other equipment also face risks of electrical performance degradation, overheating, and signal deterioration.

[0003] Traditional manual inspections suffer from blind spots, low efficiency, and delayed response, and cannot operate continuously in severe weather. Existing monitoring systems, such as tower safety monitoring systems, are often single-function, focusing only on structural tilt or monitoring only electrical parameters, lacking integrated analysis of multi-dimensional information about equipment, structure, and environment. Single communication networks have insufficient coverage in remote mountainous and high-altitude areas, and sensors are prone to failure at low temperatures, resulting in the inability to provide early warnings of catastrophic failures caused by gradual structural changes. Furthermore, existing systems generally lack energy self-sustaining designs for extreme low-temperature environments, as well as AI-based composite fault diagnosis and autonomous resource scheduling capabilities.

[0004] Therefore, how to achieve integrated, high-precision, and highly reliable remote monitoring of the structural health and operational status of broadcast television equipment in harsh environments such as high altitude, low temperature, and weak network conditions is a technical problem that urgently needs to be solved by those skilled in the art. Summary of the Invention

[0005] In view of the above problems, the present invention proposes a method and system for remote monitoring of broadcast television equipment based on a hybrid wireless network, so as to overcome the above problems or at least partially solve the above problems.

[0006] To achieve the above objectives, the present invention adopts the following technical solution:

[0007] In a first aspect, the present invention provides a method for remote monitoring of broadcast television equipment based on a hybrid wireless network, comprising: S1. Synchronously collect the structural safety parameters and electrical parameters of the broadcast television transmission tower through sensor nodes; S2. A hybrid network transmission path is used to transmit multiple collected parameters. During the transmission process, the transmission path is dynamically selected based on the network signal strength and parameter priority. S3. Perform time-domain trend prediction and frequency-domain anomaly detection on various transmission parameters, and trigger alarms of different levels based on the detection results; S4. Generate inspection or maintenance tasks for key parts of the broadcast and television transmission tower and its equipment based on the test results.

[0008] Furthermore, in S1, the safety parameters of the tower structure include one or more of the following: verticality, tilt, horizontal displacement, foundation settlement, deformation and torsion, guy wire tension, vibration, bolt loosening status, and grounding resistance.

[0009] Furthermore, the hybrid network transmission path in S2 includes GPRS network transmission path, 5G network transmission path, and BeiDou short message transmission path. The process of dynamically selecting the transmission path based on network signal strength and parameter priority includes: Calculate the signal strength and signal-to-noise ratio for each transmission path, perform normalization, and then calculate the overall signal quality index for each path. Q i :

[0010] in, Let be the normalized value of the signal strength of the i-th transmission path. This is the normalized value of the signal-to-noise ratio for the i-th transmission path; , Indicates weight; Based on the priority order of the parameters to be sent, the suitability score of each transmission path is calculated one by one. S i And select the appropriate transmission path for transmission according to priority order; wherein, the formula for calculating the suitability score is:

[0011] in, f ( P () represents a step function with respect to parameter priority P, where P is greater than or equal to a preset priority threshold. f ( P Take the maximum value. , Indicates weight; If the suitability scores of all transmission paths are higher than the threshold, then the suitability score is selected. S i The highest transmission path transmits the parameters currently to be sent; If the suitability scores of all transmission paths are below the threshold, and the current parameter to be sent has the highest priority, then the BeiDou short message transmission path will be forced to be used. If all three transmission paths are offline, the parameters to be sent will be buffered in priority order and retransmitted in priority order after the network is restored.

[0012] Furthermore, S3 includes: S31. Input the parameters with time-series characteristics one by one into the pre-trained LSTM model to obtain the time-domain prediction value of each parameter at future time. Based on the first derivative of the time-domain prediction value, determine whether the current parameter is in a stable state, a linearly deteriorating state, or an accelerated deteriorating state. S32. Perform fast Fourier transform on the current parameters and vibration parameters respectively to obtain the frequency domain spectrum. Use the pre-trained ConvLSTM model to extract the characteristics of the evolution of each frequency component over time to obtain the frequency domain prediction value. S33. When performing forward inference in the LSTM model and ConvLSTM model, perform T forward calculations on the same input parameter to obtain T sets of time-series prediction values ​​and T sets of frequency-domain prediction values. Calculate the standard deviation of each set as the time-domain prediction uncertainty and the frequency-domain prediction uncertainty. S34. Input the time-domain predicted value, frequency-domain predicted value, time-domain prediction uncertainty and frequency-domain prediction uncertainty into the Bayesian network to calculate the posterior conditional probability of the target composite fault. S35. For the multimodal parameters of the same equipment or the same structural part, calculate the residual between the predicted value and the measured value of each parameter, and use a weighted average method to merge the residuals of each parameter into a dimensionless comprehensive residual index as the fused residual; the fused residual characterizes the overall degree of anomaly of the current equipment or structural part. S36. Determine whether the posterior conditional probability exceeds the threshold value and whether the fusion residual exceeds the alarm threshold, and trigger the corresponding level of alarm based on the judgment result.

[0013] Furthermore, S34 includes: A directed acyclic graph is constructed based on prior knowledge of the physical mechanism of the fault; the fault nodes include single fault nodes and composite fault nodes defined by the combination of multiple single fault nodes; each evidence node represents the time domain prediction value or frequency domain prediction value of a parameter, and the directed edges point from the fault node to all evidence nodes that it can trigger. The current time-domain or frequency-domain predicted value is used as the current observation value of the corresponding evidence node, and the time-domain or frequency-domain prediction uncertainty is used as the standard deviation of the observation noise. The result is input into the Bayesian network, and the posterior conditional probability P of each composite fault node under the given current observation value is calculated using the connection tree algorithm or variational Bayesian inference algorithm.

[0014] Furthermore, in S36, the alarm threshold includes a static baseline threshold and a dynamic alarm threshold. The static baseline threshold is the standard deviation of the normal operating data of the equipment or structure under healthy conditions for a continuous period of time. The dynamic alarm threshold is obtained by automatically correcting the static baseline threshold using historical data and real-time meteorological information. The formula for calculating the dynamic alarm threshold is:

[0015] in, This indicates the corrected alarm threshold; Indicates the static baseline threshold; Indicates the weather correction factor. Indicates the historical trend correction factor; Weather correction factor The calculation formula is:

[0016] in, Indicates the meteorological impact coefficient. Indicates the current ambient temperature. Indicates the current relative humidity. Indicates standard ambient temperature. Indicates standard relative humidity; Historical trend correction factor The calculation formula is:

[0017] in, Indicates the influence coefficient of historical trends. This represents the average value of parameters monitored during the same historical period. This represents the standard deviation of parameters monitored during the same historical period. This represents the mean of historical health operation data corresponding to the static baseline threshold.

[0018] Furthermore, in S36, alarm levels are divided into Level 1 alarm, Level 2 alarm, and Level 3 alarm; When the fusion residual continuously exceeds the static baseline threshold and shows a slow deterioration trend, a level 3 alarm is triggered and a health assessment report is generated. When the fusion residual exceeds the dynamic alarm threshold and the posterior conditional probability of the composite fault is continuously greater than or equal to 0.5, a level 2 alarm is triggered, a pop-up warning is displayed, and a maintenance work order is generated. When the fusion residual exceeds twice the dynamic alarm threshold, or the posterior conditional probability of the composite fault is greater than or equal to 0.9, or the physical safety limit is exceeded, a level one alarm is triggered, alarm information is forcibly pushed to the terminal, and a drone is automatically dispatched to the site for verification.

[0019] Furthermore, S4 includes: Establish a digital twin model in the cloud that includes the three-dimensional geometric structure of the tower body, dynamic mapping of the health status of key equipment, an environmental field driven by historical and real-time meteorological data, as well as the dynamics and energy consumption model of the unmanned aerial vehicle; Randomly initialize the position, remaining power, health index of each device, wind speed vector, rainfall intensity and geographical obstacles of the unmanned aerial vehicle in the environmental field of the digital twin model to form a state space; take the movement of the unmanned aerial vehicle to adjacent waypoints as an action space; Adopt a DQN network, select the movement to adjacent waypoints in the action space with an ε-greedy strategy, calculate the immediate reward according to the reward function R after executing the action, and update the Q value; After iteratively training the DQN network for a preset number of generations, solidify and deploy the trained DQN network parameters; When an alarm is detected, first generate the path of the unmanned aerial vehicle in the digital twin model according to the real state of the current environmental field as a scheduling scheme, and issue the scheduling scheme to the unmanned aerial vehicle for execution.

[0020] In a second aspect, the present invention provides a remote monitoring system for radio and television equipment based on a hybrid wireless network, which is applicable to the above method, comprising: A local monitoring layer, which synchronously collects structural safety parameters of the radio and television transmission tower and electrical parameters of the equipment through sensor nodes; A remote transmission layer, which transmits the collected multiple types of parameters using a hybrid network transmission path, and dynamically selects the transmission path based on network signal strength and parameter priority during the transmission process; An intelligent diagnosis layer, which performs time-domain trend prediction and frequency-domain anomaly detection on the transmitted multiple types of parameters, and triggers alarms of different levels based on the detection results; A resource scheduling layer, which generates inspection or maintenance tasks for key parts of the radio and television transmission tower body and the equipment according to the detection results.

[0021] Further, the local monitoring layer further comprises: an energy management system, a solar panel, a supercapacitor and a self-heating circuit; The energy management system adopts a fuzzy control algorithm to control the PWM duty ratio of the solar panel and the state of the self-heating circuit; the fuzzy control algorithm is: Taking the ratio of the current output power of the solar panel to the instantaneous total power of the sensor node as the energy supply-demand ratio R, when R>0.8 and the SOC of the supercapacitor is ≥0.7, the PWM duty ratio of the solar panel is controlled to be 10%-30%; When 0.2<R<0.8 and 0.3≤SOC<0.7, the PWM duty ratio of the solar panel is controlled to be 30%-80%; When R≤0.2 and the SOC of the supercapacitor is <0.3, the PWM duty ratio of the solar panel is controlled to the maximum; When R≤0.2 and the SOC of the supercapacitor satisfies 0.3≤SOC<0.7, the PWM duty ratio of the solar panel is controlled to be 50%-80%; When the ambient temperature T > -20°C, the self-heating circuit is turned off; When the ambient temperature T≤-20°C, R>0.8 and SOC≥0.7, the continuous heating mode of the self-heating circuit is activated; When the ambient temperature T≤-20°C, 0.2<R≤0.8 and SOC<0.7, the pulse heating mode of the self-heating circuit is activated.

[0022] As can be seen from the above technical solutions, compared with the prior art, the present invention has the following beneficial effects: 1. The present invention adopts a 5G / GPRS / Beidou short message triple-mode hybrid network dynamic switching and network disconnection resume transmission mechanism based on signal strength and priority, which can automatically switch to Beidou short message or cache data when communication is interrupted in remote areas, so as to realize highly reliable communication and ensure the complete data transmission rate.

[0023] 2. The present invention adopts the multimodal fusion diagnosis technology of LSTM time domain trend prediction, ConvLSTM frequency domain anomaly detection and Bayesian composite fault diagnosis, combined with dynamic threshold adaptive correction, which can give early warning of potential structural gradual change hazards and improve the recognition accuracy of composite faults.

[0024] 3. The present invention realizes multi-objective optimization of energy consumption minimization and fault coverage maximization through UAV path planning based on deep reinforcement learning and digital twin, so as to realize intelligent autonomous operation and maintenance scheduling and significantly improve the inspection efficiency.

[0025] 4. The present invention ensures the continuous operation of equipment in a -40°C high-altitude environment and solves the problem of low-temperature power failure through the collaborative design of fuzzy logic intelligent energy management and self-heating circuit. Brief Description of the Drawings

[0026] In order to more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the following will briefly introduce the accompanying drawings required in the description of the embodiments or the prior art. Obviously, the accompanying drawings in the following description are only the embodiments of the present invention, and for those of ordinary skill in the art, other accompanying drawings can be obtained based on the provided accompanying drawings without any creative effort.

[0027] Figure 1 is a flowchart of the remote monitoring method for radio and television equipment based on a hybrid wireless network provided in the embodiment of the present invention; Figure 2 is a multi-level alarm flowchart provided in the embodiment of the present invention; Figure 3 is a schematic diagram of resource scheduling logic provided in the embodiment of the present invention; Figure 4 This is a schematic diagram illustrating the switching of a hybrid network transmission path provided in an embodiment of the present invention; Figure 5 This is a fuzzy control logic diagram of the energy management system provided in this embodiment of the invention; Figure 6 This is a framework diagram of a remote monitoring system for broadcast television equipment based on a hybrid wireless network, provided in an embodiment of the present invention. Detailed Implementation

[0028] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0029] like Figure 1 As shown, this embodiment of the invention discloses a method for remote monitoring of broadcast television equipment based on a hybrid wireless network, comprising: S1. Synchronously collect the structural safety parameters and electrical parameters of the broadcast television transmission tower through sensor nodes; S2. A hybrid network transmission path is used to transmit multiple collected parameters. During the transmission process, the transmission path is dynamically selected based on the network signal strength and parameter priority. S3. Perform time-domain trend prediction and frequency-domain anomaly detection on various transmission parameters, and trigger alarms of different levels based on the detection results; S4. Generate inspection or maintenance tasks for key parts of the broadcast and television transmission tower and its equipment based on the test results.

[0030] The specific implementation methods of the above steps will be further explained below.

[0031] S1. Local Monitoring: Composed of low-temperature resistant ZigBee sensor nodes based on the IEEE 802.15.4 protocol, deployed on key parts of the broadcast television transmission tower and its equipment, used to synchronously collect electrical parameters, environmental parameters, and tower structural safety parameters. Tower structural safety parameters include one or more of the following: verticality, tilt, horizontal displacement, foundation settlement, deformation and torsion, guy wire tension, vibration, bolt loosening status, and grounding resistance. For the acquisition of tower structural safety parameters, laser monitoring technology is used to measure horizontal displacement, foundation settlement, or deformation, with a monitoring accuracy ≤0.1mm.

[0032] ZigBee sensor nodes achieve low-power control based on an event-triggered wake-up mode. For example, they are activated when the ambient temperature changes by ≥5℃ or the device voltage fluctuates by ≥10%. During inactive periods, they enter deep sleep mode, and the MCU's default sleep current is ≤1μA. After waking up, they collect sensor data through the ADC and return to sleep mode within 10ms after sampling is completed.

[0033] In this step, the wireless sensor network is maintained through the ZigBee coordinator node. The specific maintenance process includes: network setup and address allocation, node joining and authentication (based on pre-set keys or MAC whitelists), topology monitoring and link repair (link quality (LQI) is detected through heartbeat packets, alternative routes are found through broadcast probe packets, and nodes are instructed to update their neighbor tables), and network congestion and collision management (the backoff window parameters are adjusted when the CSMA / CA collision rate exceeds the threshold).

[0034] After acquiring sensor data, the process also includes: normalizing the sensor data, transmitting the normalized data to the register via the MCU bus (AMBA AHB / APB), having the DMA controller move the data to the FIFO buffer, triggering the hardware encryption engine to execute the national cryptographic SM4 algorithm for encryption, and sending the encrypted data.

[0035] S2. Remote Transmission: A gateway supporting multiple communication standards is used to achieve encrypted transmission of the parameter data and automatic switching during network outages. The hybrid network transmission path includes GPRS, 5G, and BeiDou short message transmission paths, capable of dynamically selecting the transmission path based on signal strength and data priority, and possessing a self-healing mechanism for network outages, including automatic caching and backup channel switching. It adopts a communication standard based on the MQTT protocol, defining the topics and message formats for device reporting and server-side distribution.

[0036] Specifically, such as Figure 4 As shown, the process of dynamically selecting a transmission path based on network signal strength and parameter priority includes: 1) Calculate the signal strength and signal-to-noise ratio for each transmission path, perform normalization, and then calculate the overall signal quality index for each path. Q i :

[0037] in, Let be the normalized value of the signal strength of the i-th transmission path. This is the normalized value of the signal-to-noise ratio for the i-th transmission path; , Indicates weight; 2) Calculate the suitability score for each transmission path according to the priority order of the parameters to be sent. Si And select the appropriate transmission path for transmission according to priority order; wherein, the formula for calculating the suitability score is:

[0038] in, f ( P () represents a step function with respect to parameter priority P, where P is greater than or equal to a preset priority threshold. f ( P Take the maximum value. , Indicates weight; 3) If the suitability scores of all transmission paths are higher than the threshold, then select the suitability score. S i The highest transmission path transmits the parameters currently to be sent; If the suitability scores of all transmission paths are below the threshold, and the current parameter to be sent has the highest priority, then the BeiDou short message transmission path will be forced to be used. If all three transmission paths are offline, the parameters to be sent are cached according to priority, and retransmitted according to priority after the network is restored. Specifically, during offline resume transmission, a 32MB NOR Flash cache is built into the ZigBee node gateway, storing data in a FIFO queue (supporting up to 72 hours of data), using the BeiDou short message transmission protocol, with data packets of 196 bytes / frame, appended with BCH(31,21) error correction codes, and sent to the designated receiving ID via a BeiDou civilian card. Failed retransmissions employ a binary exponential backoff algorithm (initial backoff window 16ms, maximum retries 3 times).

[0039] In addition, during data transmission, the CSMA / CA mechanism of the IEEE 802.15.4 MAC layer is used to prevent data loss. Specifically, the data is fragmented into 128-byte frames, and a CRC-16 checksum is added to the data frame. Before transmission, two idle channel assessments (CCA) are performed. If the channel is busy, random backoff is performed, and the backoff exponent BE is incremented and CCA is repeated. After transmission, the system waits for an ACK confirmation frame. If no ACK is received within the timeout period, the system retryes (up to 3 times). If the limit is exceeded, the data is discarded and an error is reported.

[0040] S3. Intelligent Diagnosis: A fault prediction platform based on LabVIEW, integrating dynamic threshold adjustment and 3D geographic visualization functions, used for multimodal data fusion diagnosis of equipment electrical parameters and tower structural safety parameters; adopting a multimodal data fusion diagnosis platform, which is based on LabVIEW and integrates a Long Short-Term Memory (LSTM) network, used for time-domain trend prediction and frequency-domain anomaly detection of equipment and structural parameters, supports composite fault diagnosis, and can automatically correct alarm thresholds by combining historical data and real-time meteorological information.

[0041] Specifically, such as Figure 2 As shown, S3 includes: S31. Input the parameters with time-series characteristics one by one into the pre-trained LSTM model to obtain the time-domain prediction value of each parameter at future time. Based on the first derivative of the time-domain prediction value, determine whether the current parameter is in a stable state, a linearly deteriorating state, or an accelerated deteriorating state.

[0042] S32. Perform fast Fourier transform on the current parameters and vibration parameters respectively to obtain the frequency domain spectrum. Use the pre-trained ConvLSTM model to extract the characteristics of the evolution of each frequency component over time to obtain the frequency domain prediction value.

[0043] S33. When performing forward inference in the LSTM model and ConvLSTM model, perform T forward calculations on the same input parameters to obtain T sets of time-series prediction values ​​and T sets of frequency-domain prediction values. Calculate the standard deviation of each value as the time-domain prediction uncertainty and the frequency-domain prediction uncertainty.

[0044] S34. Input the time-domain predicted value, frequency-domain predicted value, time-domain prediction uncertainty, and frequency-domain prediction uncertainty into a Bayesian network to calculate the posterior conditional probability of the target composite fault; specifically including: A directed acyclic graph is constructed based on prior knowledge of the physical mechanism of the fault; the fault nodes include single fault nodes and composite fault nodes defined by the combination of multiple single fault nodes; each evidence node represents the time domain prediction value or frequency domain prediction value of a parameter, and the directed edges point from the fault node to all evidence nodes that it can trigger. The current time-domain or frequency-domain predicted value is used as the current observation value of the corresponding evidence node, and the time-domain or frequency-domain prediction uncertainty is used as the standard deviation of the observation noise. The result is input into the Bayesian network, and the posterior conditional probability P of each composite fault node under the given current observation value is calculated using the connection tree algorithm or variational Bayesian inference algorithm.

[0045] S35. For the multimodal parameters of the same equipment or the same structural part, calculate the residual between the predicted value and the measured value of each parameter, and use a weighted average method to merge the residuals of each parameter into a dimensionless comprehensive residual index as the fused residual; the fused residual characterizes the overall degree of anomaly of the current equipment or structural part.

[0046] S36. Determine whether the posterior conditional probability exceeds the threshold value and whether the fusion residual exceeds the alarm threshold, and trigger the corresponding level of alarm based on the judgment result.

[0047] The alarm threshold includes a static baseline threshold and a dynamic alarm threshold. The static baseline threshold is the standard deviation of the equipment's or structure's normal operating data over a continuous period of time in a healthy state. The dynamic alarm threshold is obtained by automatically correcting the static baseline threshold using historical data and real-time meteorological information. The formula for calculating the dynamic alarm threshold is:

[0048] in, This indicates the corrected alarm threshold; Indicates the static baseline threshold; Indicates the weather correction factor. Indicates the historical trend correction factor; Weather correction factor The calculation formula is:

[0049] in, Indicates the meteorological impact coefficient. Indicates the current ambient temperature. Indicates the current relative humidity. Indicates standard ambient temperature. Indicates standard relative humidity; Historical trend correction factor The calculation formula is:

[0050] in, Indicates the influence coefficient of historical trends. This represents the average value of parameters monitored during the same historical period. This represents the standard deviation of parameters monitored during the same historical period. This represents the mean of historical health operation data corresponding to the static baseline threshold.

[0051] Alarm levels are divided into three levels: Level 1, Level 2, and Level 3; Level 1 is the level of concern, Level 2 is the level of warning, and Level 3 is the level of emergency. The specific criteria for judgment are as follows: 1) When the fusion residual exceeds the static baseline threshold for 1 consecutive hour and shows a slow deterioration trend, a level 3 alarm is triggered, and a health assessment report is generated in the background. 2) When the fusion residual exceeds the dynamic alarm threshold and lasts for 5 minutes, and the posterior conditional probability of the composite fault is greater than or equal to 0.5, a level 2 alarm is triggered, a pop-up warning is displayed, and a maintenance work order is generated. 3) When the fusion residual exceeds twice the dynamic alarm threshold, or the posterior conditional probability of the composite fault is greater than or equal to 0.9, or the physical safety limit is exceeded, a level one alarm is triggered, and alarm information is forcibly pushed to the terminal via SMS or APP, and a drone is automatically dispatched to the site for verification.

[0052] The specific alarm triggering conditions are shown in Table 1.

[0053] Table 1. Specific triggering conditions and actions for Level 3 alarms

[0054] S4. Based on the diagnostic results, generate and execute inspection or maintenance tasks. By integrating the DQN network and digital twin model, and combining real-time meteorological data, equipment health index, and geographic information, dynamically generate the optimal inspection path for the UAV, achieving multi-objective optimization of minimizing energy consumption and maximizing fault coverage. Figure 3 As shown, it specifically includes: 1) Establish a digital twin model in the cloud that includes the three-dimensional geometric structure of the tower, dynamic mapping of the health status of key equipment, environmental field driven by historical and real-time meteorological data, and dynamics and energy consumption model of the drone. 2) Randomly initialize the drone's position, remaining battery power, health index of each device, wind speed vector, rainfall intensity, and geographical obstacles in the environment field of the digital twin model to form the state space; move the drone to a nearby waypoint as the action space; 3) A DQN network is used for offline training in the digital twin model. An ε-greedy policy is used to select nearby waypoints in the action space for movement. After the action is executed, the immediate reward is calculated according to the reward function R, and the Q value is updated. 4) After training the DQN network iteratively a preset number of times, solidify and deploy the trained DQN network parameters; 5) When an alarm is detected, a drone path is generated in the digital twin model based on the actual state of the current environment using the DQN network. This path serves as a scheduling scheme, which is then sent to the drone for execution. Specifically, multiple scheduling schemes (such as dispatching different drones, different paths, and different actions) can be quickly simulated in the digital twin environment. The scheme with the highest reward value is then selected and sent to the drone for execution, achieving a closed loop of "perception-twin-simulation-decision-execution".

[0055] Subsequently, the drone was equipped with a visible light camera, an infrared thermal imager, and a robotic arm to perform image recognition inspections and contact maintenance.

[0056] In one embodiment, such as Figure 6 As shown, the present invention also provides a remote monitoring system for broadcast television equipment based on a hybrid wireless network, which employs the above-described method, including: The local monitoring layer synchronously collects the tower structure safety parameters and equipment electrical parameters of the radio and television transmission tower through sensor nodes; The remote transmission layer uses a hybrid network transmission path to transmit the collected multiple parameters, and dynamically selects the transmission path based on network signal strength and parameter priority during the transmission process; The intelligent diagnosis layer performs time-domain trend prediction and frequency-domain anomaly detection on the transmitted multiple parameters, and triggers alarms of different levels based on the detection results; The resource scheduling layer generates inspection or maintenance tasks for key parts of the radio and television transmission tower body and equipment according to the detection results.

[0057] Wherein, the local monitoring layer further includes: an energy management system, a solar panel, a supercapacitor and a self-heating circuit; the energy management system adopts a fuzzy control algorithm to control the PWM duty ratio of the solar panel and the state of the self-heating circuit; as Figure 5 shown, the fuzzy control algorithm is: The ratio of the current output power of the solar panel to the instantaneous total power of the sensor node is taken as the energy supply-demand ratio R. When R>0.8 and the SOC of the supercapacitor is ≥0.7, it is regarded that the energy supply-demand ratio is sufficient and the SOC is high. At this time, the PWM duty ratio of the solar panel is controlled to be 10%-30%; When 0.2<R<0.8 and 0.3≤SOC<0.7, it is regarded that the energy supply-demand ratio is balanced and the SOC is medium. At this time, the PWM duty ratio of the solar panel is controlled to be 30%-80%; When R≤0.2 and the SOC of the supercapacitor is <0.3, it is regarded that the energy supply-demand ratio is severely insufficient and the SOC is low. At this time, the PWM duty ratio of the solar panel is controlled to be the maximum; When R≤0.2 and 0.3≤SOC<0.7 of the supercapacitor, the PWM duty ratio of the solar panel is controlled to be 50%-80%; When the ambient temperature T > -20°C, the self-heating circuit is turned off; When the ambient temperature T ≤ -20°C, R>0.8 and SOC≥0.7, the continuous heating mode of the self-heating circuit is started; When the ambient temperature T ≤ -20°C, 0.2<R≤0.8 and SOC<0.7, the pulse heating mode of the self-heating circuit is started.

[0058] The embodiments in this specification are described in a progressive manner, each embodiment focuses on the differences from other embodiments, and the same and similar parts between the embodiments can be referred to each other.

[0059] The above description of the disclosed embodiments enables those skilled in the art to make or use the invention. Various modifications to these embodiments will be readily apparent to those skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of the invention. Therefore, the invention is not to be limited to the embodiments shown herein, but is to be accorded the widest scope consistent with the principles and novel features disclosed herein.

Claims

1. A method for remote monitoring of broadcast television equipment based on a hybrid wireless network, characterized in that, include: S1. Synchronously collect the structural safety parameters and electrical parameters of the broadcast television transmission tower through sensor nodes; S2. A hybrid network transmission path is used to transmit multiple collected parameters. During the transmission process, the transmission path is dynamically selected based on the network signal strength and parameter priority. S3. Perform time-domain trend prediction and frequency-domain anomaly detection on various transmission parameters, and trigger alarms of different levels based on the detection results; S4. Generate inspection or maintenance tasks for key parts of the broadcast and television transmission tower and its equipment based on the test results.

2. The remote monitoring method for broadcast television equipment based on a hybrid wireless network as described in claim 1, characterized in that, In S1, the safety parameters of the tower structure include one or more of the following: verticality, tilt, horizontal displacement, foundation settlement, deformation and torsion, guy wire tension, vibration, bolt loosening status, and grounding resistance.

3. The remote monitoring method for broadcast television equipment based on a hybrid wireless network as described in claim 1, characterized in that, The hybrid network transmission path in S2 includes GPRS network transmission path, 5G network transmission path, and BeiDou short message transmission path. The process of dynamically selecting the transmission path based on network signal strength and parameter priority includes: Calculate the signal strength and signal-to-noise ratio for each transmission path, perform normalization, and then calculate the overall signal quality index for each path. Q i : in, Let be the normalized value of the signal strength of the i-th transmission path. This is the normalized value of the signal-to-noise ratio for the i-th transmission path; , Indicates weight; Based on the priority order of the parameters to be sent, the suitability score of each transmission path is calculated one by one. S i And select the appropriate transmission path for transmission according to priority order; wherein, the formula for calculating the suitability score is: in, f ( P () represents a step function with respect to parameter priority P, where P is greater than or equal to a preset priority threshold. f ( P Take the maximum value. , Indicates weight; If the suitability scores of all transmission paths are higher than the threshold, then the suitability score is selected. S i The highest transmission path transmits the parameters currently to be sent; If the suitability scores of all transmission paths are below the threshold, and the current parameter to be sent has the highest priority, then the BeiDou short message transmission path will be forced to be used. If all three transmission paths are offline, the parameters to be sent will be buffered in priority order and retransmitted in priority order after the network is restored.

4. The remote monitoring method for broadcast television equipment based on a hybrid wireless network as described in claim 1, characterized in that, S3 include: S31. Input the parameters with time-series characteristics one by one into the pre-trained LSTM model to obtain the time-domain prediction value of each parameter at future time. Based on the first derivative of the time-domain prediction value, determine whether the current parameter is in a stable state, a linearly deteriorating state, or an accelerated deteriorating state. S32. Perform fast Fourier transform on the current parameters and vibration parameters respectively to obtain the frequency domain spectrum. Use the pre-trained ConvLSTM model to extract the characteristics of the evolution of each frequency component over time to obtain the frequency domain prediction value. S33. When performing forward inference in the LSTM model and ConvLSTM model, perform T forward calculations on the same input parameter to obtain T sets of time-series prediction values ​​and T sets of frequency-domain prediction values. Calculate the standard deviation of each set as the time-domain prediction uncertainty and the frequency-domain prediction uncertainty. S34. Input the time-domain predicted value, frequency-domain predicted value, time-domain prediction uncertainty and frequency-domain prediction uncertainty into the Bayesian network to calculate the posterior conditional probability of the target composite fault. S35. For the multimodal parameters of the same equipment or the same structural part, calculate the residual between the predicted value and the measured value of each parameter, and use a weighted average method to merge the residuals of each parameter into a dimensionless comprehensive residual index as the fused residual; the fused residual characterizes the overall degree of anomaly of the current equipment or structural part. S36. Determine whether the posterior conditional probability exceeds the threshold value and whether the fusion residual exceeds the alarm threshold, and trigger the corresponding level of alarm based on the judgment result.

5. The remote monitoring method for broadcast television equipment based on a hybrid wireless network as described in claim 4, characterized in that, S34 includes: A directed acyclic graph is constructed based on prior knowledge of the physical mechanism of the fault; the fault nodes include single fault nodes and composite fault nodes defined by the combination of multiple single fault nodes; each evidence node represents the time domain prediction value or frequency domain prediction value of a parameter, and the directed edges point from the fault node to all evidence nodes that it can trigger. The current time-domain or frequency-domain predicted value is used as the current observation value of the corresponding evidence node, and the time-domain or frequency-domain prediction uncertainty is used as the standard deviation of the observation noise. The result is input into the Bayesian network, and the posterior conditional probability P of each composite fault node under the given current observation value is calculated using the connection tree algorithm or variational Bayesian inference algorithm.

6. The remote monitoring method for broadcast television equipment based on a hybrid wireless network as described in claim 4, characterized in that, In S36, the alarm threshold includes a static baseline threshold and a dynamic alarm threshold. The static baseline threshold is the standard deviation of the normal operating data of the equipment or structure under healthy conditions for a continuous period of time. The dynamic alarm threshold is obtained by automatically correcting the static baseline threshold using historical data and real-time meteorological information. The formula for calculating the dynamic alarm threshold is: in, This indicates the corrected alarm threshold; Indicates the static baseline threshold; Indicates the weather correction factor. Indicates the historical trend correction factor; Weather correction factor The calculation formula is: in, Indicates the meteorological impact coefficient. Indicates the current ambient temperature. Indicates the current relative humidity. Indicates standard ambient temperature. Indicates standard relative humidity; Historical trend correction factor The calculation formula is: in, Indicates the influence coefficient of historical trends. This represents the average value of parameters monitored during the same historical period. This represents the standard deviation of parameters monitored during the same historical period. This represents the mean of historical health operation data corresponding to the static baseline threshold.

7. The remote monitoring method for broadcast television equipment based on a hybrid wireless network as described in claim 6, characterized in that, In S36, alarm levels are divided into Level 1 alarm, Level 2 alarm and Level 3 alarm; When the fusion residual continuously exceeds the static baseline threshold and shows a slow deterioration trend, a level 3 alarm is triggered and a health assessment report is generated. When the fusion residual exceeds the dynamic alarm threshold and the posterior conditional probability of the composite fault is continuously greater than or equal to 0.5, a level 2 alarm is triggered, a pop-up warning is displayed, and a maintenance work order is generated. When the fusion residual exceeds twice the dynamic alarm threshold, or the posterior conditional probability of the composite fault is greater than or equal to 0.9, or the physical safety limit is exceeded, a level one alarm is triggered, alarm information is forcibly pushed to the terminal, and a drone is automatically dispatched to the site for verification.

8. The remote monitoring method for broadcast television equipment based on a hybrid wireless network as described in claim 1, characterized in that, S4 include: A digital twin model is created in the cloud, including the three-dimensional geometric structure of the tower, the dynamic mapping of the health status of key equipment, the environmental field driven by historical and real-time meteorological data, and the dynamics and energy consumption model of the drone. In the environment of the digital twin model, the drone's position, remaining battery power, health index of each device, wind speed vector, rainfall intensity, and geographical obstacles are randomly initialized to form the state space; the movement of the drone to a nearby waypoint is used as the action space. A DQN network is used to select nearby waypoints in the action space with an ε-greedy strategy. After the action is executed, the immediate reward is calculated according to the reward function R, and the Q value is updated. After training the DQN network iteratively a preset number of times, the trained DQN network parameters are solidified and deployed. When an alarm is detected, a drone path is first generated in the digital twin model based on the actual state of the current environment, which serves as a scheduling scheme. The scheduling scheme is then sent to the drone for execution.

9. A remote monitoring system for broadcast television equipment based on a hybrid wireless network, characterized in that, It is applicable to the method as described in any one of claims 1-8, comprising: The local monitoring layer synchronously collects the structural safety parameters and electrical parameters of the broadcast television transmission tower through sensor nodes; The remote transmission layer uses a hybrid network transmission path to transmit various collected parameters. During the transmission process, the transmission path is dynamically selected based on network signal strength and parameter priority. The intelligent diagnosis layer performs time-domain trend prediction and frequency-domain anomaly detection on multiple transmitted parameters, and triggers alarms of different levels based on the detection results; The resource scheduling layer generates inspection or maintenance tasks for key parts of the broadcast television transmitting tower body and equipment according to the detection results.

10. The remote monitoring system for broadcast television equipment based on a hybrid wireless network as described in claim 9, characterized in that, The local monitoring layer further includes: an energy management system, a solar panel, a supercapacitor, and a self-heating circuit; The energy management system adopts a fuzzy control algorithm to control the PWM duty ratio of the solar panel and the state of the self-heating circuit; the fuzzy control algorithm is as follows: Taking the ratio of the current output power of the solar panel to the instantaneous total power of the sensor node as the energy supply-demand ratio R, when R>0.8 and the SOC of the supercapacitor is ≥0.7, the PWM duty ratio of the solar panel is controlled to be 10%-30%; When 0.2<R<0.8 and 0.3≤SOC<0.7, the PWM duty ratio of the solar panel is controlled to be 30%-80%; When R≤0.2 and the SOC of the supercapacitor is <0.3, the PWM duty ratio of the solar panel is controlled to the maximum value; When R≤0.2 and 0.3≤SOC<0.7 for the supercapacitor, the PWM duty ratio of the solar panel is controlled to be 50%-80%; When the ambient temperature T>-20°C, the self-heating circuit is turned off; When the ambient temperature T≤-20°C, R>0.8 and SOC≥0.7, the continuous heating mode of the self-heating circuit is activated; When the ambient temperature T≤-20°C, 0.2<R≤0.8 and SOC<0.7, the pulse heating mode of the self-heating circuit is activated.

Images