Intelligent scheduling method and related device for UAV aerial base station based on low-orbit satellite Internet

Abstract

The present invention provides a method and related devices for intelligent scheduling of drone air base stations based on low-orbit satellite Internet, and relates to the technical field of drone scheduling. The target monitoring area is divided into grids; the monitoring requirements of different grids are quantified to generate task grids according to the importance of the line, equipment failure history, terrain and environmental conditions, natural disaster risks and external damage risks; based on the contextual multi-armed bandit algorithm combined with the task grid, the drone air nodes are dynamically scheduled to the optimal task grid through feature mapping and benefit estimation to achieve real-time monitoring and data transmission of the power transmission corridor. Considering the limited energy reserves of drones, the long-term scheduling and deployment problems of drone air nodes in dynamically changing environments are solved, and the equipment energy efficiency and energy utilization rate of drone air nodes are improved.

Description

Technical Field

[0001] The present application relates to the technical field of drone dispatching, and in particular to a method and related devices for intelligent dispatching of drone aerial base stations based on low-orbit satellite Internet. Background Art

[0002] The safety of transmission line corridors is an important guarantee for maintaining the stable operation of the power grid. However, transmission corridors are widely distributed, usually across complex terrains such as mountains, forests and wilderness. These areas are threatened by aging lines, equipment failures, extreme weather, external damage, and natural disasters such as wildfires, landslides, and mudslides. If these problems are not discovered and handled in time, they may lead to serious power outages and large-scale blackouts, causing economic losses and safety hazards. Therefore, establishing an efficient and reliable monitoring system is crucial to the safe operation and emergency response of the power grid.

[0003] A large number of transmission corridors are located in remote areas, and there are generally a large number of communication blind spots. The construction and maintenance costs of traditional ground monitoring and communication networks are high, and the coverage is limited, making it difficult to meet the real-time monitoring needs of transmission line corridors. UAV-assisted communication provides a new solution for comprehensive monitoring of remote transmission line corridors due to its flexible mobile deployment capabilities. Drones can patrol over the corridors, collect data in real time from monitoring sensors deployed along the line, and monitor the status of transmission lines, equipment conditions, the impact of extreme weather, and changes in the surrounding environment. However, the undulating terrain and communication blind spots in remote mountainous areas pose major challenges to establishing communication links and data transmission. Traditional ground communication base stations are difficult to establish in these areas and cannot provide stable remote communication services for drones.

[0004] Therefore, how to solve the long-term scheduling and deployment problem of UAV aerial nodes in a dynamically changing environment and improve the equipment energy efficiency and energy utilization of UAV aerial nodes has become a technical problem that needs to be solved urgently. Summary of the invention

[0005] In order to solve the long-term scheduling and deployment problem of drone aerial nodes in a dynamically changing environment and improve the equipment energy efficiency and energy utilization of drone aerial nodes, the present application provides a drone aerial base station intelligent scheduling method and related devices based on low-orbit satellite Internet.

[0006] In the first aspect, the present application provides a method for intelligent scheduling of drone aerial base stations based on low-orbit satellite Internet, which adopts the following technical solutions:

[0007] An intelligent scheduling method for drone air base stations based on low-orbit satellite Internet, comprising:

[0008] Grid division of the target monitoring area;

[0009] Quantify the monitoring requirements of different grids to generate mission grids based on line importance, equipment failure history, terrain and environmental conditions, natural disaster risks, and external damage risks;

[0010] Based on the contextual multi-armed bandit algorithm combined with the task grid, the drone aerial nodes are dynamically dispatched to the optimal task grid through feature mapping and benefit estimation to achieve real-time monitoring and data transmission of the transmission corridor.

[0011] Optionally, the step of quantifying the monitoring requirements of different grids to generate a task grid according to line importance, equipment failure history, terrain and environmental conditions, natural disaster risks, and external damage risks includes:

[0012] According to the importance of the line , Equipment failure history , terrain and environmental conditions , natural disaster risks and external damage risk Formula quantification through monitoring requirements;

[0013]

[0014] Among them, the weight Assignments are made based on historical data analysis, security priorities, and inspection needs.

[0015] Optional, line importance The formula is:

[0016]

[0017] in, and is the adjustment factor, Used to indicate voltage circuits, Used to represent capacitive circuits;

[0018] Equipment failure history The formula is:

[0019]

[0020] in, and is the adjustment factor, Reflects the impact of each failure on the overall monitoring needs, is the time decay factor, Represents the time after each failure occurs, Indicates the trend of equipment failure, represents the number of all faults that occurred in history, is the scaling factor, which determines the fault tendency;

[0021] Topography and environmental factors The formula is:

[0022]

[0023] For the entire area The integral of is the slope function, is the vegetation density function;

[0024] Natural disaster risks The formula is:

[0025]

[0026] in, Respectively represent Landslide, fire and flood risks in each grid, represents the number of types of natural disasters considered;

[0027] External damage risk The formula is:

[0028]

[0029] in, represents the population density around the transmission line corridor, where represents the population of the region, Indicates the area of ​​the region, represents the normalized traffic density near the transmission line corridor, where represents the traffic flow near the transmission corridor, represents the maximum traffic flow in the area, Represents the industrial activity index, which is a binary indicator variable. When there is industrial activity in the region, ,otherwise ,coefficient Indicates the relative importance of each factor in determining the overall external damage risk.

[0030] Optionally, before the step of dynamically scheduling drone air nodes to the optimal task grid based on the contextual multi-armed bandit algorithm in combination with the task grid through feature mapping and benefit estimation to achieve real-time monitoring and data transmission of the transmission corridor, the method further includes:

[0031] Obtain the characteristic information of the drone air node and the characteristic information of the mission grid, where the characteristic information of the drone air node includes: drone location, remaining power, and arrival at the charging station The characteristic information of the task grid includes: grid location, number of monitoring devices in the grid and network connection information of the monitoring devices.

[0032] Optionally, the implementation of the contextual multi-armed bandit algorithm includes:

[0033] According to the characteristic information of the aerial node of the UAV and the characteristic information of the task grid, a weight matrix is ​​used Mapping to context feature space;

[0034] Estimating linear parameters using ridge regression , calculate the expected return ;

[0035] The grid with the highest profit is selected through the upper confidence bound strategy, and the history record and weight matrix are dynamically updated.

[0036] Optionally, the step of estimating the revenue includes:

[0037] The network connectivity coverage component is based on the matching degree between the number of drone nodes and the grid requirements;

[0038] The remaining power of the drone is used to ensure the remaining power when it returns to the charging station.

[0039] The deployment limit component after charging constrains the deployment position of the drone after charging.

[0040] Optionally, the step of dynamically scheduling drone aerial nodes also includes: updating the benefit vector and training sequence to ensure continuous optimization and adaptive adjustment of the scheduling algorithm.

[0041] In a second aspect, the present application provides a drone air base station intelligent scheduling system based on low-orbit satellite Internet, and the drone air base station intelligent scheduling system based on low-orbit satellite Internet includes:

[0042] A division module is used to divide the target monitoring area into grids;

[0043] Mission grid module, used to quantify monitoring requirements of different grids to generate mission grids based on line importance, equipment failure history, terrain and environmental conditions, natural disaster risks, and external damage risks;

[0044] The drone scheduling module is used to dynamically schedule drone aerial nodes to the optimal task grid based on the contextual multi-armed bandit algorithm combined with the task grid through feature mapping and benefit estimation, so as to realize real-time monitoring and data transmission of the transmission corridor.

[0045] In a third aspect, the present application provides a computer device, comprising: a memory and a processor, wherein the processor executes the method described above when running computer instructions stored in the memory.

[0046] In a fourth aspect, the present application provides a computer-readable storage medium, comprising instructions, which, when executed on a computer, enable the computer to execute the method described above.

[0047] In summary, the present application includes the following beneficial technical effects:

[0048] This application divides the target monitoring area into grids; quantifies the monitoring requirements of different grids to generate task grids based on line importance, equipment failure history, terrain and environmental conditions, natural disaster risks and external damage risks; dynamically dispatches drone aerial nodes to the optimal task grid based on the contextual multi-armed bandit algorithm and the task grid through feature mapping and benefit estimation to achieve real-time monitoring and data transmission of the transmission corridor. Considering the limited energy reserves of drones, the long-term scheduling and deployment problems of drone aerial nodes in dynamically changing environments are solved to improve the equipment energy efficiency and energy utilization of drone aerial nodes. BRIEF DESCRIPTION OF THE DRAWINGS

[0049] Figure 1 It is a schematic diagram of the computer device structure of the hardware operating environment involved in the embodiment of the present application;

[0050] Figure 2 This is a flow chart of the first embodiment of the method for intelligent scheduling of drone aerial base stations based on low-orbit satellite Internet in this application;

[0051] Figure 3 This is a framework diagram of the intelligent scheduling method of drone aerial base stations based on low-orbit satellite Internet in this application;

[0052] Figure 4 It is a structural block diagram of the first embodiment of the drone aerial base station intelligent scheduling system based on low-orbit satellite Internet in this application. DETAILED DESCRIPTION

[0053] In order to make the purpose, technical solution and advantages of the present application more clearly understood, the present application is further described in detail below through the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present application and are not used to limit the present application.

[0054] Reference Figure 1 , Figure 1 A schematic diagram of the computer device structure of the hardware operating environment involved in the embodiment of the present application.

[0055] like Figure 1As shown, the computer device may include: a processor 1001, such as a central processing unit (CPU), a communication bus 1002, a user interface 1003, a network interface 1004, and a memory 1005. Among them, the communication bus 1002 is used to realize the connection and communication between these components. The user interface 1003 may include a display screen (Display), an input unit such as a keyboard (Keyboard), and the optional user interface 1003 may also include a standard wired interface and a wireless interface. The network interface 1004 may optionally include a standard wired interface and a wireless interface (such as a wireless fidelity (Wireless-Fidelity, Wi-Fi) interface). The memory 1005 may be a high-speed random access memory (Random Access Memory, RAM), or a stable non-volatile memory (Non-Volatile Memory, NVM), such as a disk storage. The memory 1005 may also be a storage device independent of the aforementioned processor 1001.

[0056] Those skilled in the art will understand that Figure 1 The structure shown in the figure does not constitute a limitation on the computer device, and may include more or less components than shown in the figure, or combine certain components, or arrange the components differently.

[0057] like Figure 1 As shown, the memory 1005 as a storage medium may include an operating system, a network communication module, a user interface module, and an intelligent scheduling program for drone air base stations based on low-orbit satellite Internet.

[0058] exist Figure 1 In the computer device shown, the network interface 1004 is mainly used for data communication with the network server; the user interface 1003 is mainly used for data interaction with the user; the processor 1001 and the memory 1005 in the present application can be set in the computer device, and the computer device calls the drone air base station intelligent scheduling program based on low-orbit satellite Internet stored in the memory 1005 through the processor 1001, and executes the drone air base station intelligent scheduling method based on low-orbit satellite Internet provided in the embodiment of the present application.

[0059] The present application embodiment provides a method for intelligent scheduling of drone air base stations based on low-orbit satellite Internet, referring to Figure 2 , Figure 2 This is a flow chart of the first embodiment of the intelligent scheduling method for drone aerial base stations based on low-orbit satellite Internet in this application.

[0060] In this embodiment, the intelligent scheduling method of drone air base stations based on low-orbit satellite Internet includes the following steps:

[0061] Step S10: Divide the target monitoring area into grids.

[0062] It is understood that the explanations of professional terms in this embodiment include:

[0063] MAB: abbreviation of Multi-armed Bandit algorithm. The multi-armed bandit model is a classic problem in reinforcement learning. It simulates a bandit with multiple "arms", each arm has a different probability distribution, representing different reward expectations. The goal is to maximize the total reward by constantly trying different arms, while balancing the relationship between "exploration" and "utilization", that is, making decisions between exploring the potential of unknown arms and utilizing the known optimal arm. The multi-armed bandit problem is often used to optimize scenarios such as resource allocation, advertising, and recommendation systems. The multi-armed bandit algorithm is a class of strategies for solving the multi-armed bandit problem, which aims to maximize the cumulative reward while balancing exploration (trying different arms) and utilization (selecting the current optimal arm). Common multi-armed bandit algorithms include - Greedy algorithm (with probability Explore unknown options and use the current best choice the rest of the time), the upper confidence bound (UCB) algorithm (weighs the average reward of each arm and the number of choices, giving priority to pullers with higher uncertainty), and Thompson sampling (assigns a probability to each arm based on Bayesian inference and makes choices based on the probability).

[0064] It should be noted that the transmission line corridors in remote areas are usually located in complex terrain and changing environments, and are susceptible to line aging, equipment failure, extreme weather, external damage and natural disasters. Traditional ground monitoring and manual inspections face challenges such as limited communication coverage, large data transmission delays, and high costs, making it difficult to meet monitoring needs. In order to solve these problems, this embodiment uses drones as aerial communication nodes to collect sensor data and transmit it remotely via satellite Internet. The system architecture is as follows Figure 3 As shown, it contains the following elements:

[0065] Monitoring sensor network: Various sensors including temperature, humidity, wind speed and equipment status are deployed along the transmission line corridor to monitor changes in the line and environment in real time.

[0066] UAV communication node: UAVs serve as aerial communication nodes, patrolling regularly to collect data from ground sensors and take photos of the corridor's surrounding environment. UAVs are equipped with satellite communication terminals and can transmit data directly via satellite Internet. Their flight paths are guided by a quantified monitoring demand formula.

[0067] Satellite Internet: The satellite terminal on the drone establishes a communication link with the satellite and transmits data to the ground control center in real time, ensuring stable communication around the world.

[0068] Ground Control Center: The ground control center receives, processes and analyzes data from drones to monitor the status of the transmission line corridor in real time and adjust the drone flight path and communication strategy as needed.

[0069] Drone nest: The drone nest serves as a functional node for drone take-off and landing, charging, maintenance, and storage. It can also transmit large data such as videos and pictures brought back by the drone back to the ground control center through the ground network, thereby saving bandwidth and traffic costs.

[0070] The data flow in the system starts with monitoring sensors that collect data on the transmission line corridors. This data is transmitted to drone communication nodes, which are responsible for relaying the data to the ground control center via satellite internet. The ground center then processes and analyzes the data to detect potential risks and adjust the drone flight path accordingly.

[0071] The entire inspection area is divided into several grids, and the monitoring requirements of each grid are determined by the monitoring requirement indicators defined later. The ground nodes (such as sensors) randomly generate data based on uncertain communication requirements, making the system environment dynamic. The drones must move between grids to cover areas with different monitoring requirements and establish communication with the ground nodes.

[0072] In this model, multiple drones are deployed in the inspection area and their actions are centrally coordinated by the control center. They work together to ensure maximum coverage efficiency. The control center assigns different areas to each drone and adjusts their flight paths to minimize overlap, thereby improving inspection efficiency and coverage while saving energy. Drones have three states:

[0073] Hovering state: The drone hovers over a grid to collect data.

[0074] Moving state: The drone moves between grids.

[0075] Charging state: The drone returns to the charging grid when the battery is depleted.

[0076] The energy consumption of the drone is closely related to its state and inspection path. Assume that the speed of the drone in the moving state is The power consumption in the hovering state and the moving state are and . is the number of drones, is the task duration, is the task interval time, The time it takes for the drone to move during the mission. The number of times the drone air node is charged, Indicates that the drone is on the charging grid The time required to replenish the battery.

[0077] Step S20: quantify the monitoring requirements of different grids to generate a task grid based on line importance, equipment failure history, terrain and environmental conditions, natural disaster risks, and external damage risks.

[0078] It is understandable that in the inspection of transmission line corridors, monitoring needs depend on multiple static factors, including line importance, equipment failure history, terrain and environmental conditions, natural disaster risks and external damage risks. By quantifying these factors, the deployment of drone inspections can be optimized to ensure priority inspections of key areas and improve monitoring efficiency and safety.

[0079] It should be noted that the steps to quantify the monitoring requirements of different grids to generate task grids based on line importance, equipment failure history, terrain and environmental conditions, natural disaster risks and external damage risks include:

[0080] According to the importance of the line , Equipment failure history , terrain and environmental conditions , natural disaster risks and external damage risk Formula quantification through monitoring requirements;

[0081]

[0082] Among them, the weight Assignments are made based on historical data analysis, security priorities, and inspection needs.

[0083] It is understandable that the importance of lines The formula is:

[0084]

[0085] in, and is the adjustment factor, Used to indicate voltage circuits, Used to represent capacitive circuits;

[0086] Equipment failure history The formula is:

[0087]

[0088] in, and is the adjustment factor, Reflects the impact of each failure on the overall monitoring needs, is the time decay factor, Represents the time after each failure occurs, Indicates the trend of equipment failure, represents the number of all faults that occurred in history, is the scaling factor, which determines the fault tendency;

[0089] Topography and environmental factors The formula is:

[0090]

[0091] For the entire area The integral of is the slope function, is the vegetation density function;

[0092] Natural disaster risks The formula is:

[0093]

[0094] in, Respectively represent Landslide, fire and flood risks in each grid, represents the number of types of natural disasters considered;

[0095] External damage risk The formula is:

[0096]

[0097] in, represents the population density around the transmission line corridor, where represents the population of the region, Indicates the area of ​​the region, represents the normalized traffic density near the transmission line corridor, where represents the traffic flow near the transmission corridor, represents the maximum traffic flow in the area, Represents the industrial activity index, which is a binary indicator variable. When there is industrial activity in the region, ,otherwise ,coefficient Indicates the relative importance of each factor in determining the overall external damage risk. External damage risk includes damage caused by human activities or external environmental factors. Transmission line corridors close to densely populated areas, transportation routes or industrial areas face higher external damage risks. By analyzing land use and human activity data around the corridor, external damage risks can be assessed and monitoring needs can be increased in high-risk areas.

[0098] In specific implementation, in order to ensure that the impact of different factors in the calculation is consistent, all factors should be standardized. Common normalization methods include Min-Max normalization and Z-score normalization. This quantitative formula provides a scientific basis for optimizing the deployment of drone inspections, ensuring more frequent inspections in high-risk areas, thereby improving the efficiency and safety of transmission line corridor monitoring.

[0099] Step S30: Based on the contextual multi-armed bandit algorithm combined with the task grid, the drone aerial nodes are dynamically dispatched to the optimal task grid through feature mapping and benefit estimation to achieve real-time monitoring and data transmission of the transmission corridor.

[0100] It should be noted that before the steps of dynamically dispatching drone air nodes to the optimal task grid based on the contextual multi-armed bandit algorithm combined with the task grid through feature mapping and benefit estimation to realize real-time monitoring and data transmission of the transmission corridor, it also includes:

[0101] Obtain the characteristic information of the drone air node and the characteristic information of the mission grid, where the characteristic information of the drone air node includes: drone location, remaining power, and arrival at the charging station The characteristic information of the task grid includes: grid location, number of monitoring devices in the grid and network connection information of the monitoring devices.

[0102] It can be understood that the implementation of the contextual multi-armed bandit algorithm includes: according to the characteristic information of the drone air node and the characteristic information of the task grid through the weight matrix Mapping to context feature space; estimating linear parameters using ridge regression , calculate the expected return ; The grid with the highest profit is selected through the upper confidence bound strategy, and the historical records and weight matrix are dynamically updated.

[0103] It should be noted that the steps of benefit estimation include: network connection coverage component, which is based on the matching degree between the number of drone nodes and grid demand; remaining power component, which ensures the remaining power of the drone when it returns to the charging station; and post-charging deployment restriction component, which constrains the deployment location of the drone after charging.

[0104] In a specific implementation, the step of dynamically scheduling drone aerial nodes also includes: updating the benefit vector and training sequence to ensure continuous optimization and adaptive adjustment of the scheduling algorithm.

[0105] It should be noted that this embodiment optimizes the scheduling of aerial node positions in drone-assisted communication networks based on the scheduling algorithm of the contextual multi-armed bandit. The drone is regarded as a mobile platform carrying communication nodes, and its task is to fly to different mission grid areas to help collect and transmit data from ground nodes. In order to represent the change of decision rewards under the system state, it is assumed that there is a linear relationship between the decision rewards and the context features formed by the mission grid and the state features of the drone aerial nodes. The aerial nodes and mission grids carried by the drone are represented by feature vectors and initialized before the mission starts.

[0106] In the specific implementation, the drone characteristics include the drone location, remaining battery power, and the time to reach the charging station. The characteristics of the task grid include the grid location, the number of monitoring devices in the grid, and the network connection information of the monitoring devices.

[0107] In the In the round deployment, the selected task grid is represented as ,and The number of user nodes with network connection requirements follows a binomial distribution ,in express As the remaining battery, location, and network connection information of the drone change during the mission, the feature vector needs to be regularly updated and encoded as contextual features to estimate the reward of each decision in real time, thus ensuring efficient real-time deployment.

[0108] The drones carrying the aerial nodes are regarded as the agents in the multi-armed bandit algorithm, and the mission area is divided into multiple grids as “arms”.

[0109] By applying context features, the optimal task grid can be selected through the following steps :

[0110] 1. Initialization: Weight Matrix , historical record discount factor , training sequence , the revenue vector ;

[0111] Feature weight matrix Is a linear transformation matrix whose elements represent the degree of association between the characteristics of the drone's air nodes and the characteristics of the mission grid. is the historical record discount factor, balancing the impact of historical records and current input on the result. Input in round decision context features As a pre-training sequence, combined into a matrix ,in, is the dimension value of the context feature vector. To satisfy the dimension The identity matrix of . The dimension is The zero vector of .

[0112] 2. Input: Historical interaction records , extract the features of drone aerial nodes , Mesh Features , earnings history ;

[0113] Historical interaction records ;

[0114] UAV aerial node characteristics Including drone location, remaining battery, and charging station The distance and the task grid deployed in the last round;

[0115] Mesh Features Including the grid location, the number of monitoring devices in the grid and the network connection information of the monitoring devices;

[0116] Earnings History , ;

[0117] 3. Gradient descent optimization: : ;

[0118] Input data: context vectors of UAV and mesh features from history and These feature vectors are used as input to the logistic regression model to fit the return relationship.

[0119] Linear model assumption: Assume that the features of UAV and grid are expressed by weight matrix Mapping to context feature space:

[0120]

[0121] Objective function: Logistic regression is optimized by maximum likelihood estimation . Assume that in each task grid The benefits The following relations are satisfied:

[0122]

[0123] in is the linear parameter used to predict returns.

[0124] Loss function: The loss function for logistic regression is the negative of the log-likelihood function:

[0125]

[0126] Optimization method: Gradient descent Perform iterative optimization until the loss function converges.

[0127] Update formula: The weight matrix update formula for logistic regression is:

[0128]

[0129] in: is the learning rate; is the gradient of the loss function;

[0130] 4. =1, execute 5-18 until =T;

[0131] 5. Select the drone air nodes in order: , ;

[0132] 6. Apply feature dimensionality reduction to obtain contextual features: ;

[0133] 7. For each drone air node , , execute 8-10;

[0134] 8. Feature Mapping: , get the mapping from user node feature space to grid feature space;

[0135] 9. Feature Dimensionality Reduction: 6-Dimensional Feature Vectors , ;

[0136] right and Dimensionality reduction is performed by kernel function mapping and K-means clustering to obtain the 6-dimensional drone aerial node mapping features. With mesh features ;

[0137] 10. Feature Union: ;

[0138] Mapping features of drone aerial nodes after dimensionality reduction With mesh features Calculate the outer product to obtain the 36-dimensional context feature , which can be used to obtain a linear estimate of expected returns;

[0139] 11. Yes Each grid in , execute 12-13;

[0140] 12. Estimate linear parameters: ;

[0141] Assuming there is a linear relationship between the feature vector and the expected reward, this relationship is used to calculate the reward estimate for each round of decision making, as follows:

[0142]

[0143] in, The expected return and context characteristics are The linear regression solution of the vector parameter of the linear relationship meets the condition .parameter The linear relationship represented by Get the grid in the round decision The expected value of the expected return ;

[0144] The contextual multi-armed bandit algorithm uses ridge regression (RR) to approximate the linear relationship between features and rewards to estimate rewards. The parameter estimation formula is:

[0145]

[0146] 13. Estimation selection grid of income,

[0147] ;

[0148] By defining the upper limit of the estimation error of the estimated linear model, the UCB of the expected benefits of different grid context features can be obtained, and the greedy strategy is applied to select the grid in each round of decision to improve the benefit value. In the Estimated value of expected return in round decision:

[0149]

[0150] 14. Choose the grid with the highest expected return: ;

[0151] 15. Get real benefits: Select the grid with the highest expected benefit to perform the task, i.e. , get real benefits ;

[0152] The return quantification function is divided into three components, including the following:

[0153] (1) Network connection status: In the round decision, assuming that the drone air node reaches the grid Execute the task. The number of ground nodes with network requirements and the maximum number of node connections of drone aerial nodes By comparison, we can get the first component of the decision benefit :

[0154]

[0155] in, Indicated in In the round decision, the grid This component is used to ensure network coverage. After a certain round of deployment, if all nodes in the deployed grid cannot be served by drone aerial nodes, additional drone aerial nodes should be deployed to provide services, that is, increase ;

[0156] (2) UAV power remaining: The power consumption of the drone air node comes from the mobile business energy consumption CPoM and the communication business energy consumption CPoC. Maneuver to The CPoM and CPoC consumed in the process will be calculated in the process of feature update, so that the UAV can reach Remaining power after Based on the above assumptions, we can get the second component of decision benefit: :

[0157]

[0158] in, For Grid arrive For the long-term deployment of drone air nodes, it is necessary to ensure that drone air nodes are After the task is completed, there is enough power to return to the charging grid This component ensures that the drone aerial node can perform tasks in the grid that can meet the requirements of long-term deployment, which improves the continuity of the task and conforms to the actual situation;

[0159] (3) Deployment location of the drone aerial node after charging: After charging is completed, the remaining power is the maximum battery capacity. In order to balance other drones with less remaining power, aerial nodes need to be deployed to a distance of The requirement of a closer grid stipulates that the aerial nodes that have completed charging cannot be deployed to the charging grid. , and after charging, the aerial nodes should be deployed away from grid, so that it can be deployed to a distance after its power drops. Closer to the grid, more convenient for charging. Based on the above assumptions, we can get the third component of decision benefit: :

[0160]

[0161] After obtaining the three components of decision benefits, , , Perform an "and" operation to obtain the actual benefit of this round of decision making That is, only when all three conditions are met can positive feedback be obtained;

[0162] 16. Update the profit vector: ;

[0163] 17. Update training sequence: ;

[0164] 18. Update time: t=t+1;

[0165] 19. Output drone air nodes , Task Grid .

[0166] It can be understood that this embodiment first divides the target monitoring area into grids; then quantifies the monitoring requirements of different grids according to the importance of the line, equipment failure history, terrain and environmental conditions, natural disaster risks and external damage risks; finally, based on the contextual multi-armed bandit algorithm, it is assumed that there is a linear relationship between the decision benefit and the contextual features composed of the state features of the task grid and the drone air node. In the algorithm framework of the multi-armed bandit, the air relay node carried by the drone is regarded as the gambler in the multi-armed bandit algorithm, and the grid divided by the task area is regarded as an arm. In a data-driven way, both the drone air node and the task grid are represented by feature vectors, and the environmental state represented by the drone air node and the task grid is encoded as a contextual feature through feature weight mapping, which contains dynamically changing environmental information. Single-step reinforcement learning is applied to estimate the decision benefit of this round in real time to meet the requirements of real-time and efficient scheduling, which can effectively improve the average benefit, network coverage and energy efficiency of the dynamic scheduling of drone air nodes under the transmission line corridor monitoring system based on low-orbit satellite Internet.

[0167] It can be understood that the present embodiment combines drone-assisted communication with satellite Internet, making the inspection of the transmission line corridor more flexible, with real-time data transmission, overcoming the communication limitations in remote areas and providing reliable support for its safe operation. It should be noted that in this system, the drone is equipped with a satellite terminal instead of deploying the satellite terminal directly on the ground, mainly based on the following key considerations:

[0168] Flexibility and mobility: In remote areas with complex terrain and dynamic communication needs, fixed ground satellite terminals are not only expensive to deploy, but also difficult to adapt to changes in inspection needs in different areas. UAVs equipped with satellite terminals can flexibly adjust flight paths, quickly cover different areas, and achieve higher inspection efficiency.

[0169] Overcoming terrain obstacles: Ground satellite terminals are easily blocked by obstacles in mountainous areas and forests, resulting in signal attenuation. However, drones fly at higher altitudes and can bypass terrain and vegetation obstructions, establish line-of-sight communication links, and ensure stable signal transmission.

[0170] Reduce construction and maintenance costs: Deploying ground satellite terminals in remote areas involves huge challenges in transportation, installation and maintenance. Drones carrying satellite terminals can return to the base regularly for maintenance and upgrades, reducing the high cost of building fixed facilities.

[0171] Therefore, the use of satellite terminals carried by drones provides a more flexible, economical and stable communication service, providing efficient data transmission support for the inspection of transmission line corridors.

[0172] The proposed intelligent scheduling method for UAV aerial base stations meets the requirements of real-time and efficient scheduling, and effectively improves the average revenue, network coverage and energy efficiency of dynamic scheduling of UAV aerial nodes in the transmission line corridor monitoring system based on low-orbit satellite Internet.

[0173] This embodiment divides the target monitoring area into grids; quantifies the monitoring requirements of different grids to generate task grids based on line importance, equipment failure history, terrain and environmental conditions, natural disaster risks, and external damage risks; dynamically dispatches drone air nodes to the optimal task grid based on the contextual multi-armed bandit algorithm and feature mapping and benefit estimation to achieve real-time monitoring and data transmission of the transmission corridor. Considering the limited energy reserves of drones, the long-term scheduling and deployment problems of drone air nodes in dynamically changing environments are solved to improve the equipment energy efficiency and energy utilization of drone air nodes.

[0174] In addition, an embodiment of the present application also proposes a computer-readable storage medium, on which is stored a program for intelligent scheduling of drone air base stations based on low-orbit satellite Internet. When the program for intelligent scheduling of drone air base stations based on low-orbit satellite Internet is executed by a processor, the steps of the method for intelligent scheduling of drone air base stations based on low-orbit satellite Internet as described above are implemented.

[0175] Reference Figure 4 , Figure 4 This is a structural block diagram of the first embodiment of the drone aerial base station intelligent scheduling system based on low-orbit satellite Internet in this application.

[0176] like Figure 4 As shown, the UAV aerial base station intelligent scheduling system based on low-orbit satellite Internet proposed in the embodiment of the present application includes:

[0177] A division module 10 is used to divide the target monitoring area into grids;

[0178] A task grid module 20, for quantifying the monitoring requirements of different grids to generate a task grid according to line importance, equipment failure history, terrain and environmental conditions, natural disaster risk and external damage risk;

[0179] The drone scheduling module 30 is used to dynamically schedule drone air nodes to the optimal task grid based on the contextual multi-armed bandit algorithm combined with the task grid through feature mapping and benefit estimation, so as to realize real-time monitoring and data transmission of the transmission corridor;

[0180] It should be understood that the above is only an example and does not constitute any limitation on the technical solution of the present application. In specific applications, technicians in this field can make settings as needed, and the present application does not impose any limitation on this.

[0181] This embodiment divides the target monitoring area into grids; quantifies the monitoring requirements of different grids to generate task grids based on line importance, equipment failure history, terrain and environmental conditions, natural disaster risks and external damage risks; dynamically dispatches drone aerial nodes to the optimal task grid based on the contextual multi-armed bandit algorithm in combination with the task grid through feature mapping and benefit estimation to achieve real-time monitoring and data transmission of the transmission corridor. Considering the limited energy reserves of drones, the long-term scheduling and deployment problems of drone aerial nodes in dynamically changing environments are solved to improve the equipment energy efficiency and energy utilization of drone aerial nodes.

[0182] It should be noted that the workflow described above is merely illustrative and does not limit the scope of protection of the present application. In practical applications, technicians in this field can select part or all of it according to actual needs to achieve the purpose of the present embodiment, and no limitation is made here.

[0183] In addition, for technical details not described in detail in this embodiment, please refer to the method for intelligent scheduling of drone aerial base stations based on low-orbit satellite Internet provided in any embodiment of the present application, which will not be repeated here.

[0184] In addition, it should be noted that, in this article, the terms "include", "comprises" or any other variations thereof are intended to cover non-exclusive inclusion, so that a process, method, article or system including a series of elements includes not only those elements, but also includes other elements not explicitly listed, or also includes elements inherent to such process, method, article or system. In the absence of further restrictions, an element defined by the sentence "comprises a ..." does not exclude the existence of other identical elements in the process, method, article or system including the element.

[0185] The serial numbers of the above-mentioned embodiments of the present application are for description only and do not represent the advantages or disadvantages of the embodiments.

[0186] Through the description of the above implementation methods, those skilled in the art can clearly understand that the above-mentioned embodiment methods can be implemented by means of software plus a necessary general hardware platform, and of course by hardware, but in many cases the former is a better implementation method. Based on such an understanding, the technical solution of the present application, or the part that contributes to the prior art, can be embodied in the form of a software product, which is stored in a storage medium (such as a read-only memory (ROM) / RAM, a magnetic disk, or an optical disk), and includes a number of instructions for a terminal device (which can be a mobile phone, a computer, a server, or a network device, etc.) to execute the methods of each embodiment of the present application.

[0187] The above are only preferred embodiments of the present application, and are not intended to limit the patent scope of the present application. Any equivalent structure or equivalent process transformation made using the contents of the present application specification and drawings, or directly or indirectly applied in other related technical fields, are also included in the patent protection scope of the present application.

Claims

1. An intelligent scheduling method for drone air base stations based on low-orbit satellite Internet, characterized in that: include: Grid division of the target monitoring area; Quantify the monitoring requirements of different grids to generate mission grids based on line importance, equipment failure history, terrain and environmental conditions, natural disaster risks, and external damage risks; Based on the contextual multi-armed bandit algorithm combined with the task grid, the drone air nodes are dynamically dispatched to the optimal task grid through feature mapping and benefit estimation to achieve real-time monitoring and data transmission of the transmission corridor; Among them, before the step of dynamically scheduling drone air nodes to the optimal task grid based on the contextual multi-armed bandit algorithm combined with the task grid through feature mapping and benefit estimation to realize real-time monitoring and data transmission of the transmission corridor, it also includes: Obtain the characteristic information of the drone air node and the characteristic information of the mission grid, where the characteristic information of the drone air node includes: drone location, remaining power, and arrival at the charging station The distance and the task grid deployed in the previous round. The characteristic information of the task grid includes: the grid location, the number of monitoring devices in the grid and the network connection information of the monitoring devices; The implementation of the contextual multi-armed bandit algorithm includes: According to the characteristic information of the aerial node of the UAV and the characteristic information of the task grid, a weight matrix is ​​used Mapping to context feature space; Estimating linear parameters using ridge regression , calculate the expected return ; The grid with the highest profit is selected through the upper confidence bound strategy, and the history record and weight matrix are dynamically updated.

2. The intelligent scheduling method for UAV aerial base stations based on low-orbit satellite Internet according to claim 1 is characterized in that: The step of quantifying the monitoring requirements of different grids to generate a task grid according to line importance, equipment failure history, terrain and environmental conditions, natural disaster risks and external damage risks includes: According to the importance of the line , Equipment failure history , terrain and environmental conditions , natural disaster risks and external damage risk Formula quantification through monitoring requirements; Among them, the weight Assignments are made based on historical data analysis, security priorities, and inspection needs.

3. The intelligent scheduling method for UAV aerial base stations based on low-orbit satellite Internet according to claim 2 is characterized in that: Line importance The formula is: in, and is the adjustment factor, Used to indicate voltage circuits, Used to represent capacitive circuits; Equipment failure history The formula is: in, and is the adjustment factor, Reflects the impact of each failure on the overall monitoring needs, is the time decay factor, Represents the time after each failure occurs, Indicates the trend of equipment failure, represents the number of all faults that occurred in history, is the scaling factor, which determines the fault tendency; Topography and environmental factors The formula is: For the entire area The integral of is the slope function, is the vegetation density function; Natural disaster risks The formula is: in, Respectively represent Landslide, fire and flood risks in each grid, represents the number of types of natural disasters considered; External damage risk The formula is: in, represents the population density around the transmission line corridor, where represents the population of the region, Indicates the area of ​​the region, represents the normalized traffic density near the transmission line corridor, where represents the traffic flow near the transmission corridor, represents the maximum traffic flow in the area, Represents the industrial activity index, which is a binary indicator variable. When there is industrial activity in the region, ,otherwise ,coefficient Indicates the relative importance of each factor in determining the overall external damage risk.

4. The method for intelligent scheduling of UAV aerial base stations based on low-orbit satellite Internet according to claim 1 is characterized in that: The steps of the revenue estimation include: The network connectivity coverage component is based on the matching degree between the number of drone nodes and the grid requirements; The remaining power of the drone is used to ensure the remaining power when it returns to the charging station. The deployment limit component after charging constrains the deployment position of the drone after charging.

5. The method for intelligent scheduling of UAV aerial base stations based on low-orbit satellite Internet according to claim 1 is characterized in that: The step of dynamically scheduling the drone aerial nodes also includes: updating the benefit vector and the training sequence to ensure continuous optimization and adaptive adjustment of the scheduling algorithm.

6. An intelligent dispatching system for drone air base stations based on low-orbit satellite Internet, characterized in that: Executing the method according to claim 1, the UAV aerial base station intelligent scheduling system based on low-orbit satellite Internet includes: A division module is used to divide the target monitoring area into grids; Mission grid module, used to quantify monitoring requirements of different grids to generate mission grids based on line importance, equipment failure history, terrain and environmental conditions, natural disaster risks, and external damage risks; The drone scheduling module is used to dynamically schedule drone aerial nodes to the optimal task grid based on the contextual multi-armed bandit algorithm combined with the task grid through feature mapping and benefit estimation, so as to realize real-time monitoring and data transmission of the transmission corridor.

7. A computer device, characterized in that: The device comprises: a memory and a processor, wherein the processor executes the method according to any one of claims 1 to 5 when running computer instructions stored in the memory.

8. A computer-readable storage medium, characterized in that: The method comprises instructions, which, when executed on a computer, cause the computer to execute the method according to any one of claims 1 to 5.

Images