Communication iron tower data processing method and system based on edge calculation

Abstract

The invention discloses a communication iron tower data processing method and system based on edge computing, and relates to the field of edge computing, and the method comprises the steps: responding to the received communication data for any edge computing node, determining the current state of a corresponding communication iron tower based on the communication data, the current state comprises a wireless signal state, a node flow state and an iron tower operation state; the current state is sent to a central computing node through a regional coordination gateway, and the central computing node receives the current state of each communication iron tower and determines a communication structure of the communication iron tower group based on the current states of all the communication iron towers; and the central computing node determines a communication optimization strategy of the communication iron tower group based on the communication structure of the communication iron tower group, and issues the communication optimization strategy to the corresponding edge computing node through the regional coordination gateway so as to optimize communication of the communication iron tower group. According to the invention, data transmission delay can be effectively reduced.

Description

Technical Field

[0001] The embodiments of the present application relate to the field of edge computing, and in particular, to a communication tower data processing method and system based on edge computing. Background Art

[0002] With the rapid development of 5G technology, communication towers, as the core infrastructure of communications, have a data processing capability that directly affects the performance and user experience of the entire communication network.

[0003] At present, the data processing of communication towers mainly relies on centralized cloud platforms. Under this architecture, all data collected from communication towers will be transmitted to remote cloud computing centers for processing and analysis. However, with the explosive growth of IoT devices and the high real-time requirements of 5G networks, all data is transmitted to the cloud through the wide area network, which will not only cause serious congestion of network bandwidth, but also increase the delay of data transmission. Especially in application scenarios such as autonomous driving and telemedicine, millisecond-level delays may have serious consequences. Therefore, as the amount of data processing increases, the data transmission delay of traditional communication tower data processing methods is high, which is difficult to meet the real-time response requirements of massive IoT devices in the 5G era. Summary of the invention

[0004] The embodiments of the present application provide a communication tower data processing method and system based on edge computing, which are used to effectively reduce data transmission delays.

[0005] To achieve the above objectives, the embodiments of the present application adopt the following technical solutions: In a first aspect, a communication tower data processing method based on edge computing is provided, which is applied to an edge computing system, wherein the edge computing system includes n edge computing nodes, a regional coordination gateway, and a central computing node, each of which is deployed inside a communication tower, and the number of the communication towers is n, where n is an integer greater than 2, and the method includes: For any one of the edge computing nodes, in response to receiving the communication data, determining the current state of the corresponding communication tower based on the communication data, wherein the current state includes the wireless signal state, the node traffic state and the tower operation state; The current state is sent to the central computing node through the regional coordination gateway, and the central computing node receives the current state of each of the communication towers and determines the communication structure of the communication tower group based on the current state of all the communication towers; The central computing node determines the communication optimization strategy of the communication tower group based on the communication structure of the communication tower group, and sends the communication optimization strategy to the corresponding edge computing node through the regional coordination gateway to optimize the communication of the communication tower group.

[0006] In a possible implementation manner of the first aspect, the communication data includes base station wireless signal data, node traffic data, and communication tower sensor data, each of the edge computing nodes is deployed with an LSTM lightweight model, and determining the current state of the corresponding communication tower based on the communication data includes: In the case where the communication data is the base station wireless signal data, a historical base station load data set and a real-time user distribution heat map are obtained, and a time series analysis is performed on the historical base station load data set to extract periodic load characteristics; Marking an area in the real-time user distribution heat map where the user density exceeds a preset threshold as a high-load area; Based on the periodic load characteristics and the high load area, generating a base station signal coverage strength matrix; Determining a wireless signal state according to a difference between the signal coverage strength matrix and a preset signal quality standard; In the case where the communication data is the node traffic data, the number of data packets in a unit time is counted, and if the number of data packets exceeds a preset traffic threshold, the node traffic state is determined to be a high load state, and if the number of data packets does not exceed the preset traffic threshold, the node traffic state is determined to be a low load state; In the case where the communication data is the communication tower sensor data, the LSTM lightweight model is used to process the communication tower sensor data to obtain the tower operation status.

[0007] In another possible implementation manner of the first aspect, if the node traffic state of any one of the communication towers is in a high load state, after determining the node traffic state, the method further includes: Using a preset NFV instance to offload local traffic of an edge computing node corresponding to the communication tower; Acquire the network bandwidth of the edge computing node, and perform dynamic bit rate conversion on the video streaming media of the edge computing node based on the network bandwidth; Priority marking is performed on the communication tower sensor data received by the edge computing node.

[0008] In another possible implementation manner of the first aspect, the dynamically converting the video streaming media of the edge computing node based on the network bandwidth includes: If the network bandwidth drops to a preset bandwidth threshold, a preset bit rate calculation formula is used to calculate a reduction value of the video streaming media according to the network bandwidth, and the encoding bit rate of the video streaming media is reduced based on the reduction value.

[0009] In another possible implementation manner of the first aspect, the step of using the LSTM lightweight model to process the communication tower sensor data to obtain the tower operation status includes: When the node traffic state of any one of the communication towers is in a high load state, the priority-marked communication tower sensor data is input into the LSTM lightweight model to obtain the tower operation state of the communication tower; Among them, the LSTM lightweight model includes an input layer, a first hidden layer, a second hidden layer and an output layer. The input layer is used to receive the communication tower sensor data. The first hidden layer is a sparsely connected structure and includes 32 LSTM units. The second hidden layer is connected to the first hidden layer through a gated jump connection and includes 16 LSTM units. The output layer is a fully connected layer for outputting the operating status of the tower.

[0010] In another possible implementation manner of the first aspect, determining the communication structure of the communication tower group based on the current states of all the communication towers includes: According to the wireless signal status and node traffic status of each communication tower, a network topology diagram is constructed with signal coverage strength as weight and traffic load as node attribute, wherein the communication tower corresponding to the traffic load greater than the preset load threshold is regarded as a high-load node, and the communication tower corresponding to the traffic load not greater than the preset load threshold is regarded as a low-load node; According to the network topology diagram, a communication structure of a communication tower group is determined by adopting a communication path planning strategy; Wherein, the communication path planning strategy includes: Using a preset minimum spanning tree algorithm to generate a minimum spanning tree of the network topology graph, and marking the path connecting the high-load node in the minimum spanning tree as the optimal communication path; According to the optimal communication path, a communication structure of the communication tower group is generated, wherein the communication structure includes a trunk link and a redundant link, wherein the trunk link is used for high priority data transmission, and the redundant link is used for load balancing.

[0011] In another possible implementation manner of the first aspect, the central computing node determines the communication optimization strategy of the communication tower group based on the communication structure of the communication tower group, including: For the communication tower on the backbone link, a transmission power adjustment formula is used to adjust the wireless signal transmission power of the communication tower; determining a high load area and a low load area according to the communication structure; For any communication tower located in the high-load area, excess traffic exceeding the local processing capacity is distributed to adjacent communication towers located in the low-load area according to the idle resource ratio of adjacent nodes; Determining whether there is a faulty communication tower in the communication structure according to the operating status of the tower; If there is a faulty communication tower, the faulty communication tower is removed from the communication structure, and the communication path planning strategy is used to replan the optimal communication path to bypass the faulty communication tower.

[0012] In another possible implementation manner of the first aspect, a preset fitness function is used to determine the fitness of each of the edge computing nodes, and the edge computing node with the highest fitness is used as the optimal node, and all edge computing nodes except the optimal node are used as backup nodes; Adopting a first processing strategy to allocate a first preset value of system bandwidth and a second preset value of computing resources to the optimal node; Adopting the second processing strategy, setting the resource occupancy upper limit of each backup node to a third preset value; The first processing strategy and the second processing strategy are encapsulated through the regional coordination gateway to obtain a JSON instruction, and the JSON instruction is sent to the corresponding edge computing node.

[0013] In a second aspect, the present application provides a machine-readable storage medium having instructions stored thereon, the instructions being used to enable a machine to execute the above-mentioned communication tower data processing method based on edge computing.

[0014] In a third aspect, the present application provides an edge computing system, including: n edge computing nodes, wherein each of the edge computing nodes is deployed inside a communication tower, the number of the communication towers is n, and n is an integer greater than 2; A regional coordination gateway connected to each of the edge computing nodes; The central computing node is connected to the regional coordination gateway.

[0015] Through the above technical solution, by deploying edge computing nodes inside communication towers, localized data processing is realized, which significantly reduces data transmission delay and improves real-time response capability. The localized processing capability of edge computing nodes effectively reduces data transmission delay. At the same time, the global optimization capability of central computing nodes ensures the efficient operation of communication tower groups. Edge computing nodes can process communication data in real time, quickly determine the current status of communication towers, and transmit status data to central computing nodes through regional coordination gateways. Central computing nodes build communication structures of communication tower groups based on status data of all communication towers and formulate corresponding optimization strategies. The optimization strategies are sent to edge computing nodes through regional coordination gateways to ensure communication efficiency and stability of communication tower groups. In this way, not only data transmission delay is significantly reduced and real-time response capability is improved, but also the efficient operation of communication tower groups is ensured through global optimization strategies. In addition, the localized processing capability of edge computing nodes reduces dependence on centralized cloud platforms and reduces pressure on network bandwidth. In summary, data transmission delay is significantly reduced, real-time response capability is improved, efficient operation of communication tower groups is ensured, and the requirements of massive IoT devices for real-time and efficiency in the 5G era are met.

[0016] Other features and advantages of the embodiments of the present application will be described in detail in the subsequent specific implementation section. BRIEF DESCRIPTION OF THE DRAWINGS

[0017] Figure 1 A flow chart of a communication tower data processing method based on edge computing provided in an embodiment of the present application; Figure 2 A schematic diagram of the architecture of an LSTM lightweight model provided in an embodiment of the present application; Figure 3 A schematic diagram of the structure of an edge system provided in an embodiment of the present application. DETAILED DESCRIPTION

[0018] In order to make the purpose, technical scheme and advantages of the embodiments of the present application clearer, the technical scheme in the embodiments of the present application will be clearly and completely described below in conjunction with the drawings in the embodiments of the present application. It should be understood that the specific implementation methods described herein are only used to illustrate and explain the embodiments of the present application, and are not used to limit the embodiments of the present application. Based on the embodiments in the present application, all other embodiments obtained by ordinary technicians in this field without making creative work are within the scope of protection of this application.

[0019] It should be noted that if the embodiments of the present application involve directional indications (such as up, down, left, right, front, back...), such directional indications are only used to explain the relative position relationship, movement status, etc. between the components under a certain specific posture (as shown in the accompanying drawings). If the specific posture changes, the directional indication will also change accordingly.

[0020] In addition, if there are descriptions involving "first", "second", etc. in the embodiments of the present application, the descriptions of "first", "second", etc. are only used for descriptive purposes and cannot be understood as indicating or suggesting their relative importance or implicitly indicating the number of technical features indicated. Therefore, the features defined as "first" and "second" may explicitly or implicitly include at least one of the features. In addition, the technical solutions between the various embodiments can be combined with each other, but they must be based on the ability of ordinary technicians in the field to implement them. When the combination of technical solutions is contradictory or cannot be implemented, it should be deemed that such combination of technical solutions does not exist and is not within the scope of protection required by this application.

[0021] Figure 1 The flowchart of a communication tower data processing method based on edge computing according to an embodiment of the present application is schematically shown. Figure 1 As shown, an embodiment of the present application provides a communication tower data processing method based on edge computing, which is applied to an edge computing system. The edge computing system includes n edge computing nodes, a regional coordination gateway and a central computing node. Each edge computing node is deployed inside a communication tower. The number of communication towers is n, where n is an integer greater than 2. The method may include the following steps.

[0022] S110, for any edge computing node, in response to receiving communication data, determining the current state of the corresponding communication tower based on the communication data, the current state including the wireless signal state, the node traffic state and the tower operation state; S120, sending the current state to the central computing node through the regional coordination gateway, the central computing node receives the current state of each communication tower, and determines the communication structure of the communication tower group based on the current state of all communication towers; S130. The central computing node determines the communication optimization strategy of the communication tower group based on the communication structure of the communication tower group, and sends the communication optimization strategy to the corresponding edge computing node through the regional coordination gateway to optimize the communication of the communication tower group.

[0023] In this embodiment, the edge computing node and the central computing node can be a tablet computer, a desktop, a laptop, a handheld computer, a wearable device, a notebook computer, an ultra-mobile personal computer (UMPC), a netbook, or other devices with a processor. Of course, the edge computing node and the central computing node can also be a server. The embodiment of the present application does not impose any special restrictions on the specific forms of the edge computing node and the central computing node.

[0024] In the edge computing system, each edge computing node is deployed inside the communication tower and is responsible for receiving and processing communication data from the communication tower in real time. Communication data includes base station wireless signal data, node traffic data, and communication tower sensor data. When the edge computing node receives the communication data, it first needs to classify and process the data to determine the current status of the communication tower. The determination of the wireless signal status depends on the base station wireless signal data. By analyzing the historical base station load data set and the real-time user distribution heat map, the periodic load characteristics are extracted, and the areas where the user density exceeds the preset threshold are marked as high-load areas, thereby generating a base station signal coverage strength matrix. According to the difference between the signal coverage strength matrix and the preset signal quality standard, it can be accurately judged whether the wireless signal status meets the requirements.

[0025] The node traffic state is determined by counting the number of data packets per unit time. If the number of data packets exceeds the preset traffic threshold, the node traffic state is determined to be a high-load state, otherwise it is a low-load state. The tower operation state is determined by using the LSTM lightweight model to process the communication tower sensor data. The LSTM model includes an input layer, a first hidden layer, a second hidden layer, and an output layer. The input layer receives sensor data. The first hidden layer is a sparsely connected structure containing 32 LSTM units. The second hidden layer is connected to the first hidden layer through a gated jump connection and contains 16 LSTM units. The output layer is a fully connected layer for outputting the tower operation state.

[0026] After determining the current status of the communication tower, the edge computing node sends these status data to the central computing node through the regional coordination gateway. The regional coordination gateway is responsible for transferring and coordinating data between the edge computing node and the central computing node to ensure the stability and efficiency of data transmission. After the central computing node receives the current status of all communication towers, it needs to determine the communication structure of the communication tower group based on these data. First, according to the wireless signal status and node traffic status of each communication tower, a network topology diagram with signal coverage strength as weight and traffic load as node attribute is constructed. In the network topology diagram, the communication tower corresponding to the traffic load greater than the preset load threshold is regarded as a high-load node, and the communication tower corresponding to the traffic load not greater than the preset load threshold is regarded as a low-load node. Then, the communication path planning strategy is used to determine the communication structure of the communication tower group. The communication path planning strategy includes using a preset minimum spanning tree algorithm to generate a minimum spanning tree of the network topology diagram, and marking the path connecting the high-load nodes in the minimum spanning tree as the optimal communication path. According to the optimal communication path, the communication structure of the communication tower group is generated. The communication structure includes a backbone link and a redundant link. The backbone link is used for high-priority data transmission, and the redundant link is used for load balancing. In summary, the central computing node can build an efficient and stable communication structure, providing a structural basis for subsequent communication optimization.

[0027] After determining the communication structure of the communication tower group, the central computing node needs to formulate a communication optimization strategy based on the structure and send the optimization strategy to the corresponding edge computing node through the regional coordination gateway. First, for the communication towers on the backbone link, the transmission power adjustment formula is used to adjust the wireless signal transmission power of the communication tower to ensure the balance between signal coverage strength and communication quality. Secondly, according to the communication structure, the high-load area and the low-load area are determined. For the communication towers located in the high-load area, the excess traffic exceeding the local processing capacity is allocated to the adjacent communication towers located in the low-load area according to the idle resource ratio of the neighboring nodes to achieve load balancing. In addition, according to the tower operation status, it is determined whether there is a faulty communication tower in the communication structure. If there is a faulty communication tower, the faulty communication tower is removed from the communication structure, and the communication path planning strategy is used to re-plan the optimal communication path to bypass the faulty communication tower. Finally, the preset fitness function is used to determine the fitness of each edge computing node, and the edge computing node with the highest fitness is used as the optimal node, and all edge computing nodes except the optimal node are used as backup nodes. Allocate the system bandwidth of the first preset value and the computing resources of the second preset value to the optimal node, and set the resource occupancy limit of each backup node to the third preset value. Encapsulate the above optimization strategy through the regional coordination gateway, obtain the JSON instruction, and send the JSON instruction to the corresponding edge computing node. In summary, the communication efficiency and stability of the communication tower group can be significantly improved.

[0028] In this embodiment, a corresponding processing flow may also be generated based on the current state of each communication tower, as follows: For the wireless signal status, the signal optimization process is performed in the following order: (1) The antenna direction angle is adjusted through the beamforming algorithm so that the main lobe is aimed at the user-dense area; (2) The QPSK / 16QAM modulation mode is dynamically switched according to the channel quality indicator; (3) The transmission power is calibrated based on the preset power adjustment formula; For the node traffic status, the traffic priority processing flow is executed in the following order: (1) identifying urgent communication packets in the traffic and allocating dedicated transmission channels to them; (2) enabling the cache queue mechanism for delay-tolerant data.

[0029] For the tower operation status, the fault alarm processing process is executed in the following order: (1) cut off the power supply to the faulty equipment and activate the backup power supply; (2) send an alarm message containing the fault code and location information to the maintenance terminal.

[0030] This embodiment realizes localized data processing by deploying edge computing nodes inside communication towers, significantly reduces data transmission delays, and improves real-time response capabilities. The localized processing capabilities of edge computing nodes effectively reduce data transmission delays. At the same time, the global optimization capabilities of central computing nodes ensure the efficient operation of communication tower groups. Edge computing nodes can process communication data in real time, quickly determine the current status of communication towers, and transmit status data to central computing nodes through regional coordination gateways. Central computing nodes build communication structures of communication tower groups based on status data of all communication towers, and formulate corresponding optimization strategies. The optimization strategies are sent to edge computing nodes through regional coordination gateways to ensure communication efficiency and stability of communication tower groups. In this way, not only data transmission delays are significantly reduced and real-time response capabilities are improved, but also efficient operation of communication tower groups is ensured through global optimization strategies. In addition, the localized processing capabilities of edge computing nodes reduce dependence on centralized cloud platforms and reduce pressure on network bandwidth. In summary, data transmission delays are significantly reduced, real-time response capabilities are improved, and efficient operation of communication tower groups is ensured, meeting the real-time and efficiency requirements of massive IoT devices in the 5G era.

[0031] In one implementation of this embodiment, the communication data includes base station wireless signal data, node traffic data, and communication tower sensor data. Each edge computing node is deployed with an LSTM lightweight model. The current state of the corresponding communication tower is determined based on the communication data, including the following steps: S210, when the communication data is base station wireless signal data, obtaining a historical base station load data set and a real-time user distribution heat map, and performing a time series analysis on the historical base station load data set to extract periodic load characteristics; S220, marking an area in the real-time user distribution heat map where the user density exceeds a preset threshold as a high-load area; S230, generating a base station signal coverage strength matrix based on periodic load characteristics and high load areas; S240, determining a wireless signal state according to a difference between the signal coverage strength matrix and a preset signal quality standard; S250, when the communication data is node flow data, counting the number of data packets in a unit time, and determining that the node flow state is a high load state if the number of data packets exceeds a preset flow threshold, and determining that the node flow state is a low load state if the number of data packets does not exceed the preset flow threshold; S260: When the communication data is communication tower sensor data, use an LSTM lightweight model to process the communication tower sensor data to obtain the tower operation status.

[0032] When the communication data is base station wireless signal data, it is necessary to first obtain the historical base station load data set and the real-time user distribution heat map. The historical base station load data set contains the load conditions of the base station in the past period of time. These data are usually stored in the form of time series, recording the base station load values ​​at different time points. In order to extract periodic load characteristics, time series analysis methods such as Fourier transform or autoregressive integrated moving average model (ARIMA) can be used. Fourier transform can convert time series data into frequency domain data, thereby identifying the periodic components in the data. For example, by Fourier transforming the historical base station load data set, the peak load period and the valley load period of each day can be found. These periodic characteristics are of great significance for predicting the future base station load conditions. The autoregressive integrated moving average model can predict future load trends by fitting historical data. Through time series analysis, the periodic characteristics of base station load can be extracted, such as the peak load period of each day and the load change law of each week. These periodic characteristics provide basic data support for the subsequent generation of the base station signal coverage strength matrix, which helps to more accurately judge the wireless signal status.

[0033] The real-time user distribution heat map reflects the distribution of users in the current area, usually displayed in the form of a heat map. The darker the color, the higher the user density. In order to identify high-load areas, a preset user density threshold needs to be set. For example, assuming the preset threshold is 1,000 users per square kilometer, then in the real-time user distribution heat map, areas with a user density of more than 1,000 will be marked as high-load areas. These high-load areas usually correspond to crowded areas such as commercial centers and transportation hubs. By marking the high-load areas, you can more intuitively understand the current network load. For example, in a city's real-time user distribution heat map, the user density in commercial areas and transportation hubs may exceed the preset threshold, and these areas will be marked as high-load areas.

[0034] After extracting the periodic load characteristics and marking the high-load areas, it is necessary to generate the base station signal coverage strength matrix. The base station signal coverage strength matrix is ​​a two-dimensional matrix, and each element in the matrix represents the signal coverage strength of a certain area at a certain time point. The process of generating the base station signal coverage strength matrix includes the following steps: First, based on the periodic load characteristics, predict the base station load at various time points in the future period. For example, if historical data shows that 10 am to 12 pm every day is the peak load period, then the load situation in the same period in the next few days can be predicted. Secondly, adjust the base station signal coverage strength according to the high-load area in the real-time user distribution heat map. For example, in the high-load area, the base station's transmission power can be increased to improve the signal coverage strength; in the low-load area, the transmission power can be appropriately reduced to save energy. Finally, the predicted load situation and the adjusted signal coverage strength are combined to generate the base station signal coverage strength matrix. For example, assuming that the predicted load of a certain area at 10 am is high load, and the area is marked as a high-load area, then in the base station signal coverage strength matrix, the signal coverage strength of the area at 10 am will be set to a higher value. In this way, the base station signal coverage strength matrix can accurately reflect the signal coverage conditions of various areas in the future and provide a basis for network optimization.

[0035] After generating the base station signal coverage strength matrix, it is necessary to determine the wireless signal status based on the difference between the matrix and the preset signal quality standard. The preset signal quality standard usually includes indicators such as signal strength, signal-to-noise ratio, and bit error rate. For example, assuming that the preset signal quality standard is that the signal strength is not less than -85dBm, the signal-to-noise ratio is not less than 20dB, and the bit error rate is not higher than 0.1%. Then, in the base station signal coverage strength matrix, the signal coverage strength of each area will be compared with these standards. If the signal coverage strength of an area is lower than the preset standard, the wireless signal status of the area will be judged as poor; if the signal coverage strength reaches or exceeds the preset standard, the wireless signal status will be judged as good. For example, assuming that the signal coverage strength of a certain area at 10 am is -90dBm, which is lower than the preset -85dBm standard, the wireless signal status of the area at 10 am will be judged as poor.

[0036] When the communication data is node traffic data, it is necessary to count the number of packets per unit time to determine the node traffic status. The unit time is usually set to 1 second or 1 minute, depending on the scale and traffic conditions of the network. For example, assuming that the preset traffic threshold is 1000 packets per second, then within the unit time, if the number of packets of a node exceeds 1000, the traffic state of the node will be judged as a high load state; if the number of packets does not exceed 1000, the traffic state will be judged as a low load state. For example, assuming that a node receives 1200 packets in 1 second, which exceeds the preset threshold of 1000 packets, the traffic state of the node will be judged as a high load state.

[0037] When the communication data is communication tower sensor data, the LSTM lightweight model needs to be used to process this data to obtain the tower operation status. Figure 2 A schematic diagram of the architecture of an LSTM lightweight model provided in an embodiment of the present application is shown, Figure 2As shown in the figure, the LSTM (Long Short-Term Memory) model is a special recursive neural network that can process time series data and capture long-term dependencies in the data. The LSTM lightweight model includes an input layer, a first hidden layer, a second hidden layer, and an output layer. The input layer receives the sensor data of the communication tower. The first hidden layer is a sparsely connected structure containing 32 LSTM units. The second hidden layer is connected to the first hidden layer through a gated jump connection and contains 16 LSTM units. The output layer is a fully connected layer for outputting the tower operation status. For example, assume that the sensor data of the communication tower includes environmental parameters such as temperature, humidity, and wind speed, as well as structural parameters such as vibration and tilt of the tower. These data are input into the LSTM model in the form of time series, and after being processed by the first hidden layer and the second hidden layer, the tower operation status, such as normal, warning, fault, etc., is finally obtained in the output layer. In this way, the operation status of the communication tower can be monitored in real time, potential problems can be discovered in time, and the stable operation of the communication network can be ensured.

[0038] This implementation method analyzes historical base station load data sets and real-time user distribution heat maps, extracts periodic load characteristics, and generates a base station signal coverage strength matrix to accurately determine the status of wireless signals. Secondly, by counting the number of data packets per unit time, high-load nodes can be quickly identified to provide a basis for traffic scheduling and optimization. Finally, the LSTM lightweight model is used to process the communication tower sensor data, monitor the tower operation status in real time, and detect potential problems in a timely manner. In summary, the data processing capability and network performance of the communication tower are significantly improved, ensuring the stability and efficiency of the communication network. Through the application of edge computing technology, data transmission delays are reduced, real-time response capabilities are improved, and the needs of massive IoT devices in the 5G era are met.

[0039] In one implementation of this embodiment, if the node flow state of any communication tower is in a high load state, after determining the node flow state, the following steps are also included: S310, using a preset NFV instance to offload local traffic of an edge computing node corresponding to a communication tower; S320, obtaining the network bandwidth of the edge computing node, and performing dynamic bit rate conversion on the video streaming media of the edge computing node based on the network bandwidth; S330. Priority marking is performed on the communication tower sensor data received by the edge computing node.

[0040] When the node traffic state of the communication tower is in a high-load state, a preset NFV (network function virtualization) instance needs to be used to unload the local traffic of the edge computing node corresponding to the communication tower. NFV is a technology that virtualizes the functions of traditional network devices and implements network functions such as firewalls and load balancing through software. In edge computing nodes, NFV instances can be dynamically created and destroyed to cope with different network load conditions. Under high load conditions, the local traffic of edge computing nodes may exceed their processing capacity, resulting in increased network delay and packet loss rate. In order to alleviate this situation, NFV instances can be used to unload local traffic. The specific implementation method includes: first, creating multiple NFV instances, each of which is responsible for processing a part of the local traffic. For example, assuming that the local traffic of the edge computing node is 5,000 packets per second, 5 NFV instances can be created, each of which processes 1,000 packets. Secondly, the local traffic is distributed to these NFV instances, and the load balancing algorithm is used to ensure the load balancing of each instance. For example, a polling algorithm can be used to distribute data packets to each NFV instance in turn. Finally, the load of each NFV instance is monitored, and the number of instances and traffic distribution strategy are dynamically adjusted. For example, if the load of a certain NFV instance is too high, the number of instances can be increased and the traffic can be redistributed. In this way, the local traffic of the edge computing node can be effectively unloaded, the node load can be reduced, and the network performance can be improved.

[0041] Under high load conditions, the network bandwidth of the edge computing node may be limited, affecting the transmission quality of the video streaming. In order to cope with this situation, it is necessary to obtain the network bandwidth of the edge computing node and perform dynamic bit rate conversion on the video streaming based on the network bandwidth. Dynamic bit rate conversion is a technology that adjusts the bit rate of video streaming in real time according to the network bandwidth to ensure smooth video playback. The specific implementation method includes: first, real-time monitoring of the network bandwidth of the edge computing node. For example, assuming that the current network bandwidth is 10Mbps, the maximum bit rate of the video streaming can be calculated. Secondly, the reduction value of the video streaming is calculated according to the network bandwidth. For example, assuming that the preset bit rate calculation formula is: reduction value = current bit rate - (current bit rate × (1-network bandwidth / maximum bandwidth)), where the maximum bandwidth is 20Mbps and the current bit rate is 5Mbps, then the reduction value = 5-(5×(1-10 / 20)) = 2.5Mbps. Then, the encoding bit rate of the video streaming is reduced based on the reduction value. For example, the bit rate of the video streaming is reduced from 5Mbps to 2.5Mbps. Finally, the adjusted video streaming is transmitted to the user. In this way, the smooth playback of video streaming under different network bandwidths can be ensured, and the user experience can be improved.

[0042] Under high load conditions, the communication tower sensor data received by the edge computing node may face processing delays. In order to ensure the timely processing of key data, the communication tower sensor data needs to be prioritized. Priority marking can ensure that high-priority data can be processed first. A specific implementation of an embodiment includes: first, determining the priority standard of the communication tower sensor data. For example, the sensor data can be divided into high priority, medium priority and low priority. High priority data includes tower structure safety data (such as vibration, tilt, etc.), medium priority data includes environmental parameters (such as temperature, humidity, etc.), and low priority data includes historical record data. Secondly, the sensor data is marked according to the priority standard. For example, the tower structure safety data is marked as high priority, the environmental parameters are marked as medium priority, and the historical record data is marked as low priority. Finally, the sensor data is processed according to the priority marking. For example, high priority data is processed first to ensure real-time monitoring and processing of tower structure safety data. In this way, the timely processing of key data can be ensured, and the safety and stability of the communication tower can be improved.

[0043] In another embodiment, priority marking of communication tower sensor data received by an edge computing node includes the following steps: 1. Increasing the collection frequency of sensor data under high load conditions to twice the original frequency; 2. Marking device temperature and voltage parameters as high priority data, and marking the remaining parameters as low priority data.

[0044] This implementation method uses NFV instances to offload local traffic from edge computing nodes, reduce node load, and improve network performance. Secondly, dynamic bit rate conversion is performed on video streaming media based on network bandwidth to ensure smooth video playback and improve user experience. Finally, priority marking is performed on communication tower sensor data to ensure timely processing of key data and improve the security and stability of communication towers. In summary, the data processing capability and network performance of communication towers are significantly improved, the stability and efficiency of communication networks are ensured, data transmission delays are reduced, real-time response capabilities are improved, and the needs of massive IoT devices in the 5G era are met.

[0045] In one implementation of this embodiment, dynamically converting the video streaming media of the edge computing node based on the network bandwidth includes the following steps: S410: If the network bandwidth drops to a preset bandwidth threshold, a preset bit rate calculation formula is used to calculate a reduction value of the video streaming media according to the network bandwidth, and the encoding bit rate of the video streaming media is reduced based on the reduction value.

[0046] Under high load conditions, the network bandwidth of the edge computing node may be limited, affecting the transmission quality of the video streaming. In order to cope with this situation, it is necessary to dynamically convert the video streaming bit rate according to the dynamic changes in the network bandwidth. Dynamic bit rate conversion is a technology that adjusts the bit rate of video streaming in real time according to the network bandwidth to ensure smooth video playback. The specific implementation method includes: first, real-time monitoring of the network bandwidth of the edge computing node. For example, assuming that the current network bandwidth is 10Mbps and the preset bandwidth threshold is 8Mbps, then when the network bandwidth drops to 8Mbps, the dynamic bit rate conversion is triggered. Secondly, the preset bit rate calculation formula is used to calculate the reduction value of the video streaming according to the network bandwidth. The preset bit rate calculation formula can be designed according to actual needs. For example, assuming that the preset bit rate calculation formula is: reduction value = current bit rate - (current bit rate × (1-network bandwidth / maximum bandwidth)), where the maximum bandwidth is 20Mbps and the current bit rate is 5Mbps, then the reduction value = 5-(5×(1-8 / 20)) = 3Mbps. Then, the encoding bit rate of the video streaming is reduced based on the reduction value. For example, the bit rate of the video streaming is reduced from 5Mbps to 2Mbps. Finally, the adjusted video streaming media is transmitted to the user. In this way, the smooth playback of video streaming media under different network bandwidths can be ensured, thereby improving the user experience.

[0047] In specific implementation, the process of dynamic bit rate conversion can be divided into the following steps: First, monitor the changes in network bandwidth in real time. For example, assuming that the network bandwidth drops from 10Mbps to 8Mbps, dynamic bit rate conversion needs to be triggered. Secondly, according to the preset bit rate calculation formula, calculate the reduction value of the video streaming media. For example, assuming that the preset bit rate calculation formula is: reduction value = current bit rate - (current bit rate × (1-network bandwidth / maximum bandwidth)), where the maximum bandwidth is 20Mbps and the current bit rate is 5Mbps, then the reduction value = 5-(5×(1-8 / 20)) = 3Mbps. Then, reduce the encoding bit rate of the video streaming media based on the reduction value. For example, reduce the bit rate of the video streaming media from 5Mbps to 2Mbps. Finally, transmit the adjusted video streaming media to the user. In this way, the smooth playback of video streaming media under different network bandwidths can be ensured, thereby improving the user experience.

[0048] The effects of dynamic bitrate conversion are mainly reflected in the following aspects: First, by adjusting the bitrate of video streaming in real time, it can ensure smooth video playback under different network bandwidths and avoid video freezes and buffering. For example, when the network bandwidth drops from 10Mbps to 8Mbps, reducing the bitrate of video streaming from 5Mbps to 2Mbps can ensure smooth video playback and avoid freezes. Second, dynamic bitrate conversion can improve the utilization of network bandwidth and avoid bandwidth waste. For example, when the network bandwidth is sufficient, a higher bitrate can be used to provide high-quality video; when the network bandwidth is insufficient, a lower bitrate can be used to ensure smooth video playback. Finally, dynamic bitrate conversion can improve user experience and ensure that users can get a good video viewing experience in different network environments. For example, when users watch videos in an environment with low network bandwidth, dynamic bitrate conversion can ensure smooth video playback and avoid users having a bad experience due to video freezes.

[0049] In this embodiment, when the network bandwidth drops to a preset bandwidth threshold, dynamic bit rate conversion is triggered, and a preset bit rate calculation formula is used to calculate the reduction value of the video streaming media according to the network bandwidth, and the encoding bit rate of the video streaming media is reduced based on the reduction value. Finally, the adjusted video streaming media is transmitted to the user to ensure smooth video playback. The transmission quality and user experience of video streaming media are significantly improved, the utilization rate of network bandwidth is improved, and bandwidth waste is avoided. Through the application of dynamic bit rate conversion technology, the dynamic changes of network bandwidth can be effectively responded to, ensuring the smooth playback of video streaming media in different network environments.

[0050] In one implementation of this embodiment, the LSTM lightweight model is used to process the communication tower sensor data to obtain the tower operation status, including the following steps: S510, when the node flow state of any communication tower is in a high load state, the priority-marked communication tower sensor data is input into the LSTM lightweight model to obtain the tower operation state of the communication tower; Among them, the LSTM lightweight model includes an input layer, a first hidden layer, a second hidden layer and an output layer. The input layer is used to receive communication tower sensor data. The first hidden layer is a sparsely connected structure and includes 32 LSTM units. The second hidden layer is connected to the first hidden layer through a gated jump connection and includes 16 LSTM units. The output layer is a fully connected layer for outputting the tower operation status.

[0051] Specifically, the input layer is used to receive a sensor data sequence with a time series length of 10 and an input dimension of 6, corresponding to temperature, humidity, voltage, vibration, wind speed, and tilt angle parameters; the first hidden layer contains 32 LSTM units and adopts a sparse connection structure, in which each unit is only connected to the same-position unit at the next moment and the three adjacent units; the second hidden layer contains 16 LSTM units, which are connected to the first hidden layer through gated jump connections; the output layer is a fully connected layer, which is used to output the operating status of the tower.

[0052] When the node traffic state of the communication tower is in a high-load state, in order to ensure the timely processing of key data, the priority-marked communication tower sensor data needs to be input into the LSTM (Long Short-Term Memory) lightweight model to obtain the tower's operating status. LSTM is a special recursive neural network that can process time series data and capture long-term dependencies in the data. The LSTM lightweight model is designed to reduce the occupation of computing resources while maintaining high prediction accuracy. The structure of the model includes an input layer, a first hidden layer, a second hidden layer, and an output layer. The input layer is used to receive communication tower sensor data, which usually includes environmental parameters such as temperature, humidity, wind speed, and structural parameters such as vibration and tilt of the tower. The first hidden layer is a sparsely connected structure containing 32 LSTM units. The sparsely connected structure can reduce the amount of model calculation and improve processing efficiency. The second hidden layer is connected to the first hidden layer through a gated jump connection and contains 16 LSTM units. The gated jump connection can enhance the model's expression ability and capture more complex features. The output layer is a fully connected layer, which is used to output the tower's operating status, such as normal, warning, and fault.

[0053] In the specific implementation, the processing process of the LSTM lightweight model can be divided into the following steps: First, the priority-marked communication tower sensor data is input into the input layer of the model. For example, assuming that the sensor data includes parameters such as temperature, humidity, wind speed, vibration, and tilt, these data are input into the model in the form of time series. Secondly, the data is processed by the first hidden layer, and the sparse connection structure of the first hidden layer can reduce the amount of calculation and improve the processing efficiency. For example, assuming that the input data is sensor data with a time step length of 100, the 32 LSTM units in the first hidden layer will process these data and extract the features in the time series. Then, the data is processed by the second hidden layer, and the second hidden layer and the first hidden layer are connected through gated jumps to capture more complex features. For example, the 16 LSTM units in the second hidden layer will further process the features output by the first hidden layer to extract higher-level features. Finally, the data is processed by the output layer, which is a fully connected layer for outputting the tower operation status. For example, the output layer can output a probability distribution indicating the probability that the tower operation status is normal, warning, or faulty. In this way, the operating status of communication towers can be monitored in real time, potential problems can be discovered in time, and the stable operation of the communication network can be ensured.

[0054] The effects of the LSTM lightweight model are mainly reflected in the following aspects: First, through the sparse connection structure and gated skip connection, the model can maintain high prediction accuracy while reducing the computing resource usage. For example, the sparse connection structure can reduce the model's computational workload and improve processing efficiency, while the gated skip connection can enhance the model's expressiveness and capture more complex features. Second, the model can process time series data and capture long-term dependencies in the data, thereby improving the accuracy of predictions. For example, the model can predict the future tower operation status based on historical sensor data and detect potential problems in a timely manner. Finally, the model can monitor the operation status of communication towers in real time to ensure the stable operation of the communication network. For example, the model can output the tower operation status in real time, such as normal, warning, or failure, to help operation and maintenance personnel detect and handle problems in a timely manner.

[0055] The LSTM lightweight model in this implementation reduces computing resource usage and improves processing efficiency through sparse connection structure and gated jump connection, while maintaining high prediction accuracy. Secondly, the model can process time series data, capture long-term dependencies in the data, and improve the accuracy of predictions. Finally, the model can output the tower operation status in real time, such as normal, warning or fault, to help operation and maintenance personnel to promptly discover and handle problems, significantly improving the data processing capabilities and network performance of communication towers, and ensuring the stability and efficiency of communication networks. Through the application of the LSTM lightweight model, data transmission delays are reduced, real-time response capabilities are improved, and the needs of massive IoT devices in the 5G era are met.

[0056] In one implementation of this embodiment, determining the communication structure of a communication tower group based on the current status of all communication towers includes the following steps: S610, according to the wireless signal status and node traffic status of each communication tower, construct a network topology diagram with signal coverage strength as weight and traffic load as node attribute, wherein the communication tower corresponding to the traffic load greater than the preset load threshold is regarded as a high-load node, and the communication tower corresponding to the traffic load not greater than the preset load threshold is regarded as a low-load node; S620, according to the network topology diagram, using a communication path planning strategy to determine the communication structure of the communication tower group; Among them, the communication path planning strategy includes: S1. Generate a minimum spanning tree of the network topology graph using a preset minimum spanning tree algorithm, and mark the path connecting the high-load nodes in the minimum spanning tree as the optimal communication path; S2. Generate a communication structure of the communication tower group according to the optimal communication path, wherein the communication structure includes a trunk link and a redundant link, wherein the trunk link is used for high-priority data transmission and the redundant link is used for load balancing.

[0057] Before determining the communication structure of the communication tower group, it is necessary to construct a network topology diagram based on the wireless signal status and node traffic status of each communication tower. The network topology diagram is a graphical tool for representing nodes and connection relationships in the network, where nodes represent communication towers and connection relationships represent communication links. In this network topology diagram, signal coverage strength is used as weight and traffic load is used as node attribute. Signal coverage strength reflects the quality of the wireless signal of the communication tower, usually expressed in dBm, and the higher the value, the stronger the signal. Traffic load reflects the data processing pressure of the communication tower, usually expressed in the number of packets per second. In order to distinguish between high-load nodes and low-load nodes, a preset load threshold needs to be set. For example, assuming that the preset load threshold is 1000 packets per second, the edge computing node with a traffic load greater than 1000 will be marked as a high-load node, and the edge computing node with a traffic load not greater than 1000 will be marked as a low-load node. In this way, the high-load area and the low-load area in the network can be clearly identified, providing a basis for subsequent communication path planning.

[0058] In the specific implementation, the process of constructing the network topology map can be divided into the following steps: First, collect the wireless signal status and node traffic status data of each communication tower. For example, assuming there are 10 communication towers, the wireless signal status and node traffic status data of each tower are as follows: the signal coverage strength of tower A is -80dBm, and the traffic load is 1200; the signal coverage strength of tower B is -85dBm, and the traffic load is 800; the signal coverage strength of tower C is -90dBm, and the traffic load is 1500; .... Secondly, construct a network topology map based on these data. For example, tower A, tower B, tower C, etc. are taken as nodes, the signal coverage strength is taken as the connection weight, and the traffic load is taken as the node attribute. Then, according to the preset load threshold, the nodes are divided into high-load nodes and low-load nodes. For example, the traffic load of tower A and tower C is greater than 1000, and they are marked as high-load nodes; the traffic load of tower B is not greater than 1000, and it is marked as a low-load node. Finally, a network topology map reflecting the actual status of the communication tower group can be constructed.

[0059] After constructing the network topology, a communication path planning strategy is needed to determine the communication structure of the communication tower group. The communication path planning strategy includes two main steps: first, a preset minimum spanning tree algorithm is used to generate a minimum spanning tree of the network topology, and the path connecting the high-load nodes in the minimum spanning tree is marked as the optimal communication path; second, based on the optimal communication path, the communication structure of the communication tower group is generated. The communication structure includes a trunk link and a redundant link. The trunk link is used for high-priority data transmission, and the redundant link is used for load balancing.

[0060] In the specific implementation, the execution process of the communication path planning strategy can be divided into the following steps: First, the minimum spanning tree algorithm is used to generate the minimum spanning tree of the network topology graph. The minimum spanning tree is a tree structure that contains all nodes and has the smallest sum of connection weights. For example, assuming that there are 10 nodes in the network topology graph, the minimum spanning tree algorithm will find the minimum weight path connecting these 10 nodes. Then, the path connecting the high-load nodes in the minimum spanning tree is marked as the optimal communication path. For example, assuming that the path weight connecting tower A and tower C in the minimum spanning tree is the smallest, then this path will be marked as the optimal communication path. Next, based on the optimal communication path, the communication structure of the communication tower group is generated. The communication structure includes a trunk link and a redundant link. The trunk link is used for high-priority data transmission, and the redundant link is used for load balancing. For example, assuming that the optimal communication path includes tower A, tower B and tower C, then the trunk link will connect these three towers for transmitting high-priority data; the redundant link will connect other towers for load balancing. In this way, an efficient and stable communication structure can be constructed to ensure the efficient operation of the communication tower group.

[0061] The minimum spanning tree algorithm is an algorithm for finding a tree structure that connects all nodes and has the smallest total weight in a weighted graph. Common minimum spanning tree algorithms include Kruskal's algorithm and Prim's algorithm. In communication path planning, the minimum spanning tree algorithm is used to generate a minimum spanning tree of a network topology graph and mark the path connecting high-load nodes as the optimal communication path. The specific implementation method includes: first, initialize an empty minimum spanning tree and a set containing all nodes. For example, assuming that there are 10 nodes in the network topology graph, initialize an empty minimum spanning tree and a set containing these 10 nodes. Secondly, select the edge with the smallest weight and add it to the minimum spanning tree. For example, assuming that the edge connecting tower A and tower B has the smallest weight, then add this edge to the minimum spanning tree. Then, repeat the above steps until the minimum spanning tree contains all nodes. For example, assuming that the minimum spanning tree already contains tower A, tower B and tower C, then continue to select the edge with the smallest weight until all nodes are included in the minimum spanning tree. Finally, mark the path connecting high-load nodes in the minimum spanning tree as the optimal communication path. For example, if the path connecting tower A and tower C in the minimum spanning tree has the smallest weight, then this path will be marked as the optimal communication path. In this way, the optimal communication path connecting high-load nodes can be found, ensuring the efficient transmission of high-priority data.

[0062] After determining the optimal communication path, it is necessary to generate the communication structure of the communication tower group according to the optimal communication path. The communication structure includes a trunk link and a redundant link. The trunk link is used for high-priority data transmission, and the redundant link is used for load balancing. The specific implementation method includes: first, according to the optimal communication path, determine the connection relationship of the trunk link. For example, assuming that the optimal communication path includes tower A, tower B and tower C, then the trunk link will connect these three towers for transmitting high-priority data. Secondly, according to the network topology diagram, determine the connection relationship of the redundant link. The redundant link is used to connect other towers to achieve load balancing. For example, assuming that there are towers D, tower E and tower F in the network topology diagram, then the redundant link will connect these towers for load sharing. Then, configure the transmission parameters of the trunk link and the redundant link. For example, the trunk link can use higher bandwidth and lower latency to ensure the rapid transmission of high-priority data; the redundant link can use lower bandwidth and higher latency to share the load. Finally, monitor the operating status of the trunk link and the redundant link, and dynamically adjust the transmission parameters. For example, if the load on the backbone link is too high, the bandwidth of the redundant link can be increased to share part of the load. In this way, an efficient and stable communication structure can be built to ensure the efficient operation of the communication tower group.

[0063] This implementation method constructs a network topology diagram with signal coverage strength as weight and traffic load as node attribute according to the wireless signal status and node traffic status of each communication tower, and clearly identifies high-load nodes and low-load nodes. Secondly, the minimum spanning tree algorithm is used to generate the minimum spanning tree of the network topology diagram, and the path connecting the high-load nodes is marked as the optimal communication path to ensure the efficient transmission of high-priority data. Finally, the communication structure of the communication tower group is generated according to the optimal communication path, including a trunk link and a redundant link. The trunk link is used for high-priority data transmission, and the redundant link is used for load balancing. The implementation of these steps significantly improves the data processing capability and network performance of the communication tower group, and ensures the stability and efficiency of the communication network. The communication structure of the communication tower group is determined based on the current status of all communication towers, which reduces data transmission delays and can effectively improve the stability and efficiency of the communication network.

[0064] In one implementation of this embodiment, the central computing node determines the communication optimization strategy of the communication tower group based on the communication structure of the communication tower group, including the following steps: S710, for the communication tower on the backbone link, adjusting the wireless signal transmission power of the communication tower using a transmission power adjustment formula; S720, determining a high-load area and a low-load area according to the communication structure; S730: for any communication tower located in a high-load area, allocating excess traffic exceeding the local processing capacity to adjacent communication towers located in a low-load area according to the idle resource ratio of adjacent nodes; S740, determining whether there is a faulty communication tower in the communication structure according to the tower operation status; S750: If there is a faulty communication tower, remove the faulty communication tower from the communication structure, and use a communication path planning strategy to replan an optimal communication path to bypass the faulty communication tower.

[0065] In the communication structure of a communication tower group, the backbone link is used to transmit high-priority data, so it is necessary to ensure that the communication towers on the backbone link have the best wireless signal coverage. To achieve this goal, the transmission power adjustment formula can be used to dynamically adjust the wireless signal transmission power of the communication tower. The transmission power adjustment formula is usually designed based on multiple factors such as signal coverage strength, network load, and interference level.

[0066] The transmission power adjustment formula of an embodiment is: adjusted transmission power = current transmission power × (target signal coverage strength / actual signal coverage strength) × (1-network load / maximum load) × (1-interference level / maximum interference level), where the target signal coverage strength is -80dBm, the actual signal coverage strength is -85dBm, the network load is 70%, the maximum load is 100%, the interference level is 30%, and the maximum interference level is 100%. Then the adjusted transmission power = current transmission power × (-80 / -85) × (1-0.7) × (1-0.3) = current transmission power × 0.94 × 0.3 × 0.7 = current transmission power × 0.1974. In this way, the transmission power of the communication tower can be dynamically adjusted to ensure that the signal coverage strength reaches the target value, while avoiding energy waste and increased interference caused by excessive transmission power.

[0067] In the specific implementation, the process of transmission power adjustment can be divided into the following steps: First, monitor the signal coverage strength, network load and interference level of the communication tower on the backbone link in real time. For example, assuming that the current signal coverage strength is -85dBm, the network load is 70%, and the interference level is 30%. Secondly, according to the transmission power adjustment formula, calculate the adjusted transmission power. For example, assuming that the current transmission power is 20W, then the adjusted transmission power = 20 × 0.1974 = 3.948W. Then, apply the adjusted transmission power to the communication tower. For example, adjust the transmission power of the communication tower from 20W to 3.948W. Finally, monitor the adjusted signal coverage strength, network load and interference level to ensure that the expected effect is achieved. For example, if the adjusted signal coverage strength is still lower than the target value, the transmission power can be further adjusted. In this way, it can be ensured that the communication tower on the backbone link has the best wireless signal coverage and improve the stability and efficiency of data transmission.

[0068] In the communication structure of a communication tower group, the determination of high-load areas and low-load areas is crucial for load balancing and resource optimization. High-load areas generally refer to areas where the traffic load exceeds a preset threshold, and low-load areas refer to areas where the traffic load is lower than the preset threshold. In order to determine high-load areas and low-load areas, it is necessary to analyze the traffic load data of each communication tower in the communication structure. For example, assuming that the preset traffic load threshold is 1000 packets per second, the area where the communication towers with a traffic load exceeding 1000 are located is marked as a high-load area, and the area where the communication towers with a traffic load below 1000 are located is marked as a low-load area. In this way, the high-load areas and low-load areas in the network can be clearly identified, providing a basis for subsequent load balancing and resource optimization.

[0069] In the specific implementation, first, the traffic load data of each communication tower in the communication structure is collected. For example, assuming that there are 10 communication towers, the traffic load data of each tower is as follows: Tower A is 1200, Tower B is 800, Tower C is 1500, and so on. Secondly, according to the preset traffic load threshold, the communication towers are divided into high-load nodes and low-load nodes. For example, the traffic load of Tower A and Tower C exceeds 1000 and is marked as a high-load node; the traffic load of Tower B is less than 1000 and is marked as a low-load node. Then, according to the distribution of high-load nodes and low-load nodes, high-load areas and low-load areas are determined. For example, assuming that Tower A and Tower C are located in the same area, then the area is marked as a high-load area; Tower B is located in another area, then the area is marked as a low-load area. In this way, the high-load area and low-load area in the network can be clearly identified, providing a basis for subsequent load balancing and resource optimization.

[0070] In high-load areas, communication towers may face excess traffic that exceeds local processing capabilities, resulting in increased network delays and packet loss rates. In order to alleviate this situation, it is necessary to allocate excess traffic to neighboring communication towers in low-load areas according to the proportion of idle resources of neighboring nodes. The specific implementation method includes: first, determining the excess traffic of communication towers in high-load areas. For example, assuming that the traffic load of communication tower A in the high-load area is 1200 and the local processing capacity is 1000, then the excess traffic is 200. Secondly, determine the idle resources of communication towers in neighboring low-load areas. For example, assuming that the idle resources of communication tower B in the neighboring low-load area are 300, and the idle resources of communication tower C are 200. Then, allocate excess traffic according to the proportion of idle resources of neighboring nodes. For example, the idle resource ratio of communication tower B is 300 / (300+200)=0.6, and the idle resource ratio of communication tower C is 200 / (300+200)=0.4, then communication tower A allocates 200×0.6=120 traffic to communication tower B, and allocates 200×0.4=80 traffic to communication tower C. Finally, monitor the traffic load after allocation to ensure that the expected effect is achieved. For example, if the traffic load after allocation is still higher than the local processing capacity, the allocation ratio can be further adjusted. In this way, the traffic pressure in high-load areas can be effectively alleviated, and the stability and efficiency of the network can be improved.

[0071] In the communication structure of a communication tower group, a faulty communication tower may cause network interruption and performance degradation. In order to ensure the stable operation of the network, it is necessary to determine whether there is a faulty communication tower in the communication structure based on the tower operation status. The tower operation status usually includes normal, warning, and fault states, which can be monitored in real time through the LSTM lightweight model. For example, if the tower operation status output by the LSTM lightweight model is faulty, then the communication tower is marked as a faulty node. In this way, the faulty communication tower can be discovered in time and appropriate measures can be taken to deal with it.

[0072] In a specific implementation, the process of determining a faulty communication tower can be divided into the following steps: First, monitor the operating status of each communication tower in the communication structure in real time. For example, suppose there are 10 communication towers, and the operating status of each tower is as follows: Tower A is normal, Tower B is warning, Tower C is faulty, and so on. Secondly, determine whether there is a faulty communication tower based on the operating status. For example, if the operating status of Tower C is faulty, then the communication tower is marked as a faulty node. Then, record the location and fault type of the faulty communication tower. For example, suppose Tower C is located in area A and the fault type is a hardware fault, then record this information. Finally, notify the operation and maintenance personnel to handle it. For example, send a fault alarm to the operation and maintenance personnel, and provide the location and fault type information of the faulty communication tower. In this way, the faulty communication tower can be discovered in time to ensure the stable operation of the network.

[0073] After determining the faulty communication tower, it is necessary to remove the faulty communication tower from the communication structure, and use the communication path planning strategy to re-plan the optimal communication path to bypass the faulty communication tower. The specific implementation method includes: first, remove the faulty communication tower from the communication structure. For example, assuming that the faulty communication tower is Tower C, then remove it from the communication structure. Secondly, re-plan the optimal communication path using the communication path planning strategy. For example, a new minimum spanning tree is generated using the minimum spanning tree algorithm, and the path connecting the high-load node is marked as the optimal communication path. Then, according to the new optimal communication path, a new communication structure is generated. For example, the new communication structure includes a trunk link and a redundant link, the trunk link is used for high-priority data transmission, and the redundant link is used for load balancing. Finally, monitor the operating status of the new communication structure to ensure that the expected effect is achieved. For example, if the new communication structure still has performance problems, the path planning strategy can be further adjusted. In this way, the faulty communication tower can be bypassed to ensure the stable operation of the network.

[0074] In this embodiment, the central computing node can determine and implement effective communication optimization strategies based on the communication structure of the communication tower group, significantly improving the stability and efficiency of the communication network. First, the transmission power adjustment formula is used to dynamically adjust the wireless signal transmission power of the communication tower on the trunk link to ensure that the signal coverage strength reaches the target value and improve the stability and efficiency of data transmission. Secondly, the high-load area and the low-load area are determined according to the communication structure to provide a basis for load balancing and resource optimization. Then, the excess traffic in the high-load area is allocated to the communication tower in the low-load area according to the idle resource ratio of the adjacent nodes, which alleviates the traffic pressure in the high-load area and improves the stability and efficiency of the network. Next, the faulty communication tower is determined according to the tower operation status, and the fault is discovered and handled in time to ensure the stable operation of the network. Finally, the faulty communication tower is removed from the communication structure, and the optimal communication path is replanned to bypass the faulty node to ensure the stable operation of the network. The implementation of these steps significantly improves the data processing capability and network performance of the communication tower group and ensures the stability and efficiency of the communication network.

[0075] In one implementation of this embodiment, the following steps are also included: S810: Using a preset fitness function, determine the fitness of each edge computing node, and use the edge computing node with the highest fitness as the optimal node, and use all edge computing nodes except the optimal node as backup nodes; S820: adopt a first processing strategy to allocate a first preset value of system bandwidth and a second preset value of computing resources to the optimal node; S830, adopting the second processing strategy to set the resource occupation upper limit of each backup node to a third preset value; S840. Encapsulate the first processing strategy and the second processing strategy through the regional coordination gateway to obtain a JSON instruction, and send the JSON instruction to the corresponding edge computing node.

[0076] In the edge computing system, the fitness of the edge computing node reflects its performance and resource utilization efficiency in the current network environment. In order to determine the fitness of each edge computing node, a preset fitness function can be used. The fitness function is usually designed based on multiple performance indicators, such as computing power, network bandwidth, load conditions, energy consumption, etc. For example, assuming that the fitness function is: fitness = computing power × 0.4 + network bandwidth × 0.3 + (1-load rate) × 0.2 + (1-energy consumption rate) × 0.1, where the computing power is 1000MIPS, the network bandwidth is 100Mbps, the load rate is 70%, and the energy consumption rate is 50%, then fitness = 1000 × 0.4 + 100 × 0.3 + (1-0.7) × 0.2 + (1-0.5) × 0.1 = 400 + 30 + 0.06 + 0.05 = 430.11. In this way, the fitness of each edge computing node can be calculated, and the node with the highest fitness is used as the optimal node, and the other nodes are used as backup nodes.

[0077] In the specific implementation, the process of determining fitness can be divided into the following steps: First, collect the performance indicator data of each edge computing node. For example, suppose there are 10 edge computing nodes, and the computing power, network bandwidth, load rate and energy consumption rate data of each node are as follows: Node A is 1000MIPS, 100Mbps, 70%, 50%; Node B is 800MIPS, 80Mbps, 60%, 40%; Node C is 1200MIPS, 120Mbps, 80%, 60%; and so on. Secondly, according to the fitness function, calculate the fitness of each node. For example, the fitness of node A is 430.11, the fitness of node B is 344.08, and the fitness of node C is 516.13. Then, compare the fitness of each node to determine the optimal node and backup node. For example, node C has the highest fitness and is marked as the optimal node; node A and node B have lower fitness and are marked as backup nodes.

[0078] After determining the optimal node, it is necessary to allocate more system bandwidth and computing resources to it to ensure that it can efficiently process high-priority tasks. The first processing strategy usually includes allocating a system bandwidth of a first preset value and computing resources of a second preset value to the optimal node. For example, assuming that the system bandwidth of the first preset value is 200Mbps and the computing resources of the second preset value are 2000MIPS, then the optimal node will be allocated a system bandwidth of 200Mbps and a computing resource of 2000MIPS. In this way, it can be ensured that the optimal node has sufficient resources to efficiently process high-priority tasks and improve the overall performance of the network.

[0079] In a specific implementation, the process of allocating system bandwidth and computing resources can be divided into the following steps: First, determine the current system bandwidth and computing resources of the optimal node. For example, assume that the current system bandwidth of the optimal node is 100Mbps and the computing resources are 1000MIPS. Secondly, according to the first processing strategy, allocate a system bandwidth of a first preset value and computing resources of a second preset value to the optimal node. For example, increase the system bandwidth of the optimal node from 100Mbps to 200Mbps, and increase the computing resources from 1000MIPS to 2000MIPS. Then, monitor the resource usage of the optimal node to ensure that the expected effect is achieved. For example, if the resource utilization rate of the optimal node is still high, the system bandwidth and computing resources can be further increased. In this way, it can be ensured that the optimal node has sufficient resources to efficiently process high-priority tasks and improve the overall performance of the network.

[0080] After the backup node is determined, it is necessary to set a resource occupancy upper limit for it to ensure that it does not over-occupy resources and affect the performance of the optimal node. The second processing strategy usually includes setting a resource occupancy upper limit of a third preset value for each backup node. For example, assuming that the resource occupancy upper limit of the third preset value is 50%, then the resource occupancy upper limit of each backup node will be set to 50%. In this way, it can be ensured that the backup node can share part of the load when needed, while not over-occupying resources and affecting the performance of the optimal node.

[0081] In a specific implementation, the process of setting a resource occupancy upper limit can be divided into the following steps: First, determine the current resource occupancy of each backup node. For example, assume that the current resource occupancy rate of backup node A is 60%, and the current resource occupancy rate of backup node B is 40%. Secondly, according to the second processing strategy, set a resource occupancy upper limit of a third preset value for each backup node. For example, set the resource occupancy upper limit of backup node A to 50%, and set the resource occupancy upper limit of backup node B to 50%. Then, monitor the resource occupancy of the backup nodes to ensure that the resource occupancy upper limit is not exceeded. For example, if the resource occupancy rate of backup node A exceeds 50%, its resource usage can be restricted to ensure that it does not exceed the upper limit. In this way, it can be ensured that the backup nodes can share part of the load when needed, and at the same time will not occupy too much resources and affect the performance of the optimal node.

[0082] After determining the first processing strategy and the second processing strategy, these strategies need to be encapsulated as JSON instructions through the regional coordination gateway and sent to the corresponding edge computing nodes. JSON is a lightweight data exchange format that is easy to read and write, and is also easy for machines to parse and generate. By encapsulating and sending JSON instructions through the regional coordination gateway, the accurate transmission and execution of the strategy can be ensured. For example, assuming that the first processing strategy allocates 200Mbps system bandwidth and 2000MIPS computing resources to the optimal node, and the second processing strategy sets a 50% resource occupancy limit for the backup node, these strategies can be encapsulated as JSON instructions in the following format: {"optimal node":{"system bandwidth":"200Mbps","computing resources":"2000MIPS"},"backup node":{"resource occupancy limit":"50%"}}. These JSON instructions are sent to the corresponding edge computing nodes through the regional coordination gateway to ensure the accurate execution of the strategy.

[0083] In the specific implementation, the process of encapsulating and issuing JSON instructions can be divided into the following steps: First, encapsulate the first processing strategy and the second processing strategy into JSON format. For example, the system bandwidth and computing resource allocation strategy of the optimal node, as well as the resource occupancy upper limit setting strategy of the backup node, are encapsulated into JSON instructions. Secondly, the JSON instructions are issued to the corresponding edge computing nodes through the regional coordination gateway. For example, the JSON instructions are sent to the optimal node and the backup node. Then, the edge computing node receives and parses the JSON instructions and executes the corresponding strategies. For example, after the optimal node receives the JSON instruction, it increases the system bandwidth to 200Mbps and the computing resources to 2000MIPS; after the backup node receives the JSON instruction, it sets the resource occupancy upper limit to 50%. Finally, monitor the execution of the edge computing nodes to ensure the accurate execution of the strategy. For example, if the system bandwidth of the optimal node does not reach 200Mbps, the JSON instruction can be reissued to ensure the accurate execution of the strategy. In this way, the accurate transmission and execution of the strategy can be ensured, and the overall performance of the network can be improved.

[0084] In another embodiment, the computing tasks can also be dynamically allocated to the optimal edge computing node based on the improved particle swarm optimization model, and a redundant backup mechanism between edge nodes can be established with the optimal edge computing node as the center. Specifically, the position vector of each particle in the particle swarm is first defined as , respectively represent the node computing capacity, bandwidth surplus and load rate; The fitness function is designed as: ; in , and is the weight coefficient; Secondly, the edge computing node with the highest fitness is selected as the optimal node by iteratively updating the particle velocity; A redundant backup mechanism is established with the optimal edge computing node as the center, including: 1. Synchronizing the computing task copies to the two adjacent edge computing nodes in real time; 2. Setting up a heartbeat monitoring mechanism. If the optimal node responds to a timeout, switch to the backup node within 50ms. The metadata and key feature parameters generated during the processing are processed with differential privacy protection to obtain the processed data, and the global resources are coordinated based on the processed data to determine the first processing strategy of the optimal edge computing node and the second processing strategy of the edge nodes other than the optimal edge computing node.

[0085] The metadata and key feature parameters generated during the processing can also be processed with differential privacy protection, including: 1. Hash desensitization of the device ID in the metadata; 2. Adding traffic feature parameters that meet the privacy protection requirements; Laplace noise.

[0086] Finally, the first processing strategy and the second processing strategy are encapsulated as JSON instructions through the regional coordination gateway and sent to the corresponding edge computing node.

[0087] This implementation adopts a preset fitness function to determine the fitness of each edge computing node, and uses the node with the highest fitness as the optimal node, and other nodes as backup nodes, to provide a basis for subsequent resource allocation and optimization. Secondly, the first processing strategy is adopted to allocate more system bandwidth and computing resources to the optimal node to ensure that it can efficiently handle high-priority tasks and improve the overall performance of the network. Then, the second processing strategy is adopted to set a resource occupancy limit for the backup node to ensure that it can share part of the load when needed, while not excessively occupying resources and affecting the performance of the optimal node. Finally, the regional coordination gateway is encapsulated and issued JSON instructions to ensure the accurate transmission and execution of the policy. The implementation of these steps significantly improved the resource utilization efficiency and network performance of the edge computing system, and ensured the stability and efficiency of the network.

[0088] An embodiment of the present application also provides a machine-readable storage medium, on which instructions are stored, and the instructions are used to enable a machine to execute the above-mentioned communication tower data processing method based on edge computing.

[0089] Figure 3 A schematic diagram of the structure of an edge system provided in an embodiment of the present application is shown. Figure 3 As shown, the embodiment of the present application also provides an edge computing system, including: n edge computing nodes, where each edge computing node is deployed inside a communication tower, and the number of communication towers is n, where n is an integer greater than 2; Regional coordination gateway, connected to each edge computing node; The central computing node is connected to the regional coordination gateway.

[0090] Those skilled in the art will appreciate that the embodiments of the present application may be provided as methods, systems or computer program products. Therefore, the present application may take the form of a complete hardware embodiment, a complete software embodiment or an embodiment combining software and hardware. Moreover, the present application may take the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program codes.

[0091] The present application is described with reference to the flowcharts and / or block diagrams of the methods, devices (systems), and computer program products according to the embodiments of the present application. It should be understood that each process and / or block in the flowchart and / or block diagram and the combination of the processes and / or blocks in the flowchart and / or block diagram can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing device to generate a machine, so that the instructions executed by the processor of the computer or other programmable data processing device generate instructions for implementing the processes in the flowchart and / or block diagram. Figure 1 A process or multiple processes and / or boxes Figure 1 A device that provides the functions specified in a block or multiple blocks.

[0092] These computer program instructions may also be stored in a computer-readable memory capable of directing a computer or other programmable data processing device to operate in a specific manner, so that the instructions stored in the computer-readable memory produce an article of manufacture comprising an instruction device, which implements the process Figure 1 A process or multiple processes and / or boxes Figure 1 A function specified in one or more boxes.

[0093] These computer program instructions can also be loaded onto a computer or other programmable data processing device so that a series of operating steps are executed on the computer or other programmable device to produce a computer-implemented process, thereby providing instructions for implementing the process. Figure 1 A process or multiple processes and / or boxes Figure 1 A step that specifies a function in one or more boxes.

[0094] In a typical configuration, a computing device includes one or more processors (CPU), input / output interfaces, network interfaces, and memory.

[0095] The memory may include non-permanent memory in a computer-readable medium, random access memory (RAM) and / or non-volatile memory in the form of read-only memory (ROM) or flash RAM. The memory is an example of a computer-readable medium.

[0096] Computer readable media include permanent and non-permanent, removable and non-removable media that can be implemented by any method or technology to store information. Information can be computer readable instructions, data structures, program modules or other data. Examples of computer storage media include, but are not limited to, phase change memory (PRAM), static random access memory (SRAM), dynamic random access memory (DRAM), other types of random access memory (RAM), read-only memory (ROM), electrically erasable programmable read-only memory (EEPROM), flash memory or other memory technology, compact disk read-only memory (CD-ROM), digital versatile disk (DVD) or other optical storage, magnetic cassettes, magnetic tape disk storage or other magnetic storage devices or any other non-transmission media that can be used to store information that can be accessed by a computing device. As defined herein, computer readable media does not include temporary computer readable media (transitory media), such as modulated data signals and carrier waves.

[0097] It should also be noted that the terms "include", "comprises" or any other variations thereof are intended to cover non-exclusive inclusion, so that a process, method, commodity or device including a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, commodity or device. In the absence of more restrictions, the elements defined by the sentence "comprises a ..." do not exclude the existence of other identical elements in the process, method, commodity or device including the elements.

[0098] The above are only embodiments of the present application and are not intended to limit the present application. For those skilled in the art, the present application may have various changes and variations. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present application should be included within the scope of the claims of the present application.

Claims

1. A communication tower data processing method based on edge computing, characterized in that: Applied to an edge computing system, the edge computing system includes n edge computing nodes, a regional coordination gateway and a central computing node, each of the edge computing nodes is deployed inside a communication tower, the number of the communication towers is n, n is an integer greater than 2, and the method includes: For any one of the edge computing nodes, in response to receiving the communication data, determining the current state of the corresponding communication tower based on the communication data, wherein the current state includes the wireless signal state, the node traffic state and the tower operation state; The current state is sent to the central computing node through the regional coordination gateway, and the central computing node receives the current state of each of the communication towers and determines the communication structure of the communication tower group based on the current state of all the communication towers; The central computing node determines the communication optimization strategy of the communication tower group based on the communication structure of the communication tower group, and sends the communication optimization strategy to the corresponding edge computing node through the regional coordination gateway to optimize the communication of the communication tower group.

2. The method according to claim 1, characterized in that The communication data includes base station wireless signal data, node traffic data and communication tower sensor data. Each edge computing node is deployed with an LSTM lightweight model. The current state of the corresponding communication tower is determined based on the communication data, including: In the case where the communication data is the base station wireless signal data, a historical base station load data set and a real-time user distribution heat map are obtained, and a time series analysis is performed on the historical base station load data set to extract periodic load characteristics; Marking an area in the real-time user distribution heat map where the user density exceeds a preset threshold as a high-load area; Based on the periodic load characteristics and the high load area, generating a base station signal coverage strength matrix; Determining a wireless signal state according to a difference between the signal coverage strength matrix and a preset signal quality standard; In the case where the communication data is the node traffic data, the number of data packets in a unit time is counted, and if the number of data packets exceeds a preset traffic threshold, the node traffic state is determined to be a high load state, and if the number of data packets does not exceed the preset traffic threshold, the node traffic state is determined to be a low load state; In the case where the communication data is the communication tower sensor data, the LSTM lightweight model is used to process the communication tower sensor data to obtain the tower operation status.

3. The method according to claim 2, characterized in that If the node flow state of any of the communication towers is in a high load state, after determining the node flow state, the method further includes: Using a preset NFV instance to offload local traffic of an edge computing node corresponding to the communication tower; Acquire the network bandwidth of the edge computing node, and perform dynamic bit rate conversion on the video streaming media of the edge computing node based on the network bandwidth; Priority marking is performed on the communication tower sensor data received by the edge computing node.

4. The method according to claim 3, characterized in that The dynamically converting the video streaming media of the edge computing node based on the network bandwidth includes: If the network bandwidth drops to a preset bandwidth threshold, a preset bit rate calculation formula is used to calculate a reduction value of the video streaming media according to the network bandwidth, and the encoding bit rate of the video streaming media is reduced based on the reduction value.

5. The method according to claim 3, characterized in that: The adopting of the LSTM lightweight model to process the communication tower sensor data to obtain the tower operation status includes: When the node traffic state of any one of the communication towers is in a high load state, the priority-marked communication tower sensor data is input into the LSTM lightweight model to obtain the tower operation state of the communication tower; Among them, the LSTM lightweight model includes an input layer, a first hidden layer, a second hidden layer and an output layer. The input layer is used to receive the communication tower sensor data. The first hidden layer is a sparsely connected structure and includes 32 LSTM units. The second hidden layer is connected to the first hidden layer through a gated jump connection and includes 16 LSTM units. The output layer is a fully connected layer for outputting the operating status of the tower.

6. The method according to claim 1, characterized in that The determining of the communication structure of the communication tower group based on the current status of all the communication towers comprises: According to the wireless signal status and node traffic status of each communication tower, a network topology diagram is constructed with signal coverage strength as weight and traffic load as node attribute, wherein the communication tower corresponding to the traffic load greater than the preset load threshold is regarded as a high-load node, and the communication tower corresponding to the traffic load not greater than the preset load threshold is regarded as a low-load node; According to the network topology diagram, a communication structure of a communication tower group is determined by adopting a communication path planning strategy; Wherein, the communication path planning strategy includes: Using a preset minimum spanning tree algorithm to generate a minimum spanning tree of the network topology graph, and marking the path connecting the high-load node in the minimum spanning tree as the optimal communication path; According to the optimal communication path, a communication structure of the communication tower group is generated, wherein the communication structure includes a trunk link and a redundant link, wherein the trunk link is used for high priority data transmission, and the redundant link is used for load balancing.

7. The method according to claim 6, characterized in that The central computing node determines a communication optimization strategy for the communication tower group based on the communication structure of the communication tower group, including: For the communication tower on the backbone link, a transmission power adjustment formula is used to adjust the wireless signal transmission power of the communication tower; determining a high load area and a low load area according to the communication structure; For any communication tower located in the high-load area, excess traffic exceeding the local processing capacity is distributed to adjacent communication towers located in the low-load area according to the idle resource ratio of adjacent nodes; Determining whether there is a faulty communication tower in the communication structure according to the operating status of the tower; If there is a faulty communication tower, the faulty communication tower is removed from the communication structure, and the communication path planning strategy is used to replan the optimal communication path to bypass the faulty communication tower.

8. The method according to claim 1, characterized in that The method further comprises: Using a preset fitness function, determine the fitness of each edge computing node, and use the edge computing node with the highest fitness as the optimal node, and use all edge computing nodes except the optimal node as backup nodes; Adopting a first processing strategy to allocate a first preset value of system bandwidth and a second preset value of computing resources to the optimal node; Adopting the second processing strategy, setting the resource occupancy upper limit of each backup node to a third preset value; The first processing strategy and the second processing strategy are encapsulated through the regional coordination gateway to obtain a JSON instruction, and the JSON instruction is sent to the corresponding edge computing node.

9. A machine-readable storage medium, characterized in that: The machine-readable storage medium stores instructions for enabling a machine to execute a communication tower data processing method based on edge computing according to any one of claims 1 to 8.

10. An edge computing system, characterized in that: The communication tower data processing method based on edge computing applied to any one of claims 1 to 8 comprises: n edge computing nodes, wherein each of the edge computing nodes is deployed inside a communication tower, the number of the communication towers is n, and n is an integer greater than 2; A regional coordination gateway connected to each of the edge computing nodes; The central computing node is connected to the regional coordination gateway.

Images