A Dynamic Antenna Resource Scheduling Method for Low-Earth Orbit Satellite Communication

Abstract

This invention discloses a dynamic antenna resource scheduling method for low-Earth orbit (LEO) satellite communication, relating to the field of satellite communication technology. It aims to overcome the technical bottlenecks of traditional genetic algorithms and static scheduling methods. Addressing the specific constraints of LEO satellite communication scenarios, it designs a dynamic spatiotemporal quantum coding structure to support automatic reorganization of task units and multi-satellite collaborative parameter embedding. Furthermore, through relative time coding, dynamic grouping, and conflict prediction mechanisms, it solves the problems of short time windows and numerous conflicts in LEO satellite communication. Complementarily, it constructs a fitness function based on a hierarchical game architecture to perform multi-objective optimization of the antenna resource allocation scheme. By dynamically allocating task priorities and energy consumption weights through a Nash equilibrium model, it completes the improvement of the quantum coding and multi-objective game-based genetic algorithm, achieving efficient and robust dynamic antenna resource scheduling for task allocation and conflict optimization.

Description

Technical Field

[0001] This invention relates to the field of satellite communication technology, and in particular to a method for dynamic scheduling of antenna resources for low-Earth orbit satellite communication. Background Technology

[0002] Satellite antenna scheduling is a crucial aspect of satellite communications, aiming to optimize resource allocation, improve communication efficiency, and reduce interference. This is particularly important in low-Earth orbit (LEO) satellite-ground station collaborative communication scenarios, where LEO satellite transit windows are extremely short, typically 5-15 minutes. The high-speed movement and short transit times of LEO satellites lead to the following shortcomings in current satellite antenna resource scheduling methods:

[0003] Traditional static scheduling methods, such as fixed-time-slice polling, allocate resources to each beam position sequentially according to a fixed order and time interval. They rely on a stable and predictable environment and cannot make sufficient resource allocations in a short time to meet rapidly changing resource demands. Another example is the greedy algorithm, which makes the best or optimal choice in each step under the current state. It is prone to getting stuck in local optima, has a slow convergence speed and poor adaptability, and it is difficult to balance task completion rate and energy consumption.

[0004] Traditional genetic algorithm scheduling methods typically employ binary encoding, roulette wheel selection, single-point crossover, and site mutation, with parameters such as crossover rate and mutation rate remaining constant. When dealing with disturbances such as ephemeris updates and antenna failures, they require the entire chromosome to be overturned for time-consuming reconstruction. The rescheduling delay often exceeds the task window tolerance threshold, leading to a high probability of scheduling scheme failure. Summary of the Invention

[0005] The purpose of this invention is to address the problem that traditional satellite antenna scheduling methods in low-Earth orbit satellite communication scenarios have a high probability of failure due to sudden disturbances such as ephemeris updates and antenna failures within extremely short window times. This invention provides a dynamic satellite antenna resource scheduling method based on quantum coding and multi-objective game theory.

[0006] To achieve the objectives of this invention, the technical solution adopted is as follows:

[0007] A method for dynamic scheduling of antenna resources for low-Earth orbit satellite communication includes the following steps:

[0008] S1. Manage antenna static information and real-time status information, acquire antenna resources; parse user task requirements, and acquire task data;

[0009] S2. Based on the antenna resources and the task data, construct a dynamic scheduling model for antenna resources using a genetic algorithm, specifically including the following steps:

[0010] S21. Designing the chromosome structure: A dynamic spatiotemporal quantum hybrid coding system composed of time, antenna, and satellite is adopted to compile the scheduling scheme into a chromosome composed of multiple gene segments, with each gene segment corresponding to a task unit; multiple gene segments are linearly spliced ​​according to the order of scheduling execution, with time as the implicit sequence, to form the chromosome; the structure of the gene segment includes: task ID, satellite ID, ground station ID, antenna ID, start time, duration, antenna elevation angle, antenna azimuth angle, and task priority flag;

[0011] S22. Use relative time encoding and dynamic window grouping to process the time window;

[0012] Relative time encoding specifically includes the following steps: using the starting point of the scheduling cycle as a reference, offset encoding is used to divide the absolute time into fixed units, and the task time is represented by an integer offset;

[0013] The window dynamic grouping includes the following steps: if the task data is a long task, the long task is divided into multiple gene segments, and at least two units of protection time are inserted between adjacent gene segments, so that the task duration of each gene segment is ≤ minimum antenna dwell time + satellite continuous visibility time.

[0014] S3. Obtain the scheduling solution for antenna resource allocation through a dynamic scheduling model;

[0015] S4. Based on the antenna resources and the task data, set physical constraints, verify whether the scheduling solution violates the physical constraints, and then retain the valid scheduling solutions in the scheduling solutions; the physical constraints include: antenna mechanical limitations, coverage limitations, and energy consumption limitations;

[0016] S5. Optimize the effective scheduling solution using the fitness function to generate a dynamic scheduling scheme for antenna resources.

[0017] This invention, based on traditional genetic algorithms, employs a dynamic spatiotemporal quantum hybrid encoding composed of time, antennas, and satellites. Combined with relative time encoding and dynamic window grouping, it constructs a dynamic antenna resource scheduling model. This model not only efficiently expresses the complex constraints of scheduling low-Earth orbit satellites and multiple ground station antenna resources, but also transforms complex time window management into efficient offset operations through relative time encoding and dynamic window grouping. It normalizes absolute time to an integer offset relative to the scheduling starting point, enabling the genetic algorithm to quickly detect and adjust gene segments using integer operations under sudden disturbances such as ephemeris updates and antenna failures. Positioning on the timeline eliminates the need for time-consuming reconstruction of the entire chromosome, thus avoiding overlapping time windows during crossover and mutation operations and achieving efficient conflict elimination. Optimization and conflict management of effective communication periods ensure tasks are completed within the effective timeframe, preventing potential overlap or conflict between time windows of different tasks. Combined with the setting of physical constraints and verification of scheduling solutions, physical conflicts are eliminated, effectively reducing the probability of scheduling scheme failure and improving its effectiveness. Furthermore, the optimization effect of the fitness function balances task completion rate and energy consumption, resulting in a dynamic scheduling scheme that achieves optimization within seconds and a high task completion rate.

[0018] Preferably, in the dynamic antenna resource scheduling method for low-Earth orbit satellite communication described in this invention, step S21 further includes setting the field length of the gene segment to variable-length encoding, specifically including the following steps:

[0019] The field length of the gene segment is encoded using variable length encoding. Bits are then allocated to each field according to the requirements of the task, and extension bits are reserved for adding new constraints.

[0020] As a preferred embodiment of the present invention, by using variable-length encoding for the field length of the gene segment and reserving extension bits, constraints can be added according to the actual needs of the task, supporting automatic reorganization of task units and embedding of multi-satellite collaborative parameters, thereby dynamically adjusting the scheduling scheme, further improving the response capability to sudden disturbances such as ephemeris updates and antenna failures, reducing the probability of obtaining a failed scheduling scheme, and improving the sensitivity of dynamic scheduling of antenna resources.

[0021] Preferably, in the dynamic scheduling method for antenna resources for low-Earth orbit satellite communication described in this invention, step S21 further includes designing deformable gene segments, specifically comprising the following steps:

[0022] The coding length of the gene segment is set so that the coding length dynamically expands / contracts according to the urgency of the task data (low / high).

[0023] As a preferred embodiment of the present invention, by setting a dynamically expandable gene segment encoding length, in the event of sudden disturbances such as ephemeris updates and antenna failures, the gene segments of urgent tasks are compressed into metadata pointers, while the gene segments of non-urgent tasks retain detailed parameters, thereby realizing dynamic priority adjustment, improving the real-time performance and scalability of dynamic antenna resource scheduling, and facilitating the acquisition of a better global scheduling scheme.

[0024] Preferably, in the dynamic antenna resource scheduling method for low-Earth orbit satellite communication described in this invention, step S21 further includes topology optimization of the coding space, specifically comprising the following steps:

[0025] A graph neural network is used to model the chromosome, embedding the relationships between antenna nodes, satellite nodes, and mission nodes into the encoding space of the graph model, thereby constructing a topological graph containing resource nodes, mission nodes, and constraint relationship edges.

[0026] As a preferred embodiment of the present invention, the genetic algorithm is guided to search for high-quality scheduling schemes that conform to real-world constraints more intelligently by combining topology optimization of the coding space, thereby further reducing the failure probability of the scheduling scheme, laying the foundation for multi-objective optimization, enhancing the coordination between the dynamic scheduling model and multi-objective optimization, and improving the applicability of dynamic scheduling of antenna resources.

[0027] Preferably, the dynamic antenna resource scheduling method for low-Earth orbit satellite communication described in this invention, which sets the antenna mechanical constraints and verifies the scheduling solution, specifically includes the following steps:

[0028] Hash mapping collision detection is used to discretize the azimuth and elevation ranges of the antenna into a sparse hash table. According to the antenna specifications, the state of each cell in the sparse hash table is marked as valid or invalid. When a task submits a target angle, the target angle is queried from the sparse hash table to verify whether the target angle is valid or not, thereby verifying whether the scheduling solution is valid or not.

[0029] Compared to traditional methods that directly calculate antenna elevation and azimuth reachability or traverse and check time conflicts to verify the feasibility of scheduling schemes, which require a large amount of real-time computation, the preferred solution of this invention uses hash mapping conflict detection to quickly eliminate infeasible schemes and avoids the generation of invalid scheduling solutions by genetic algorithms, thereby further improving the optimization efficiency of scheduling solutions and meeting the millisecond-level response requirements of low-Earth orbit satellite scheduling.

[0030] Preferably, the dynamic antenna resource scheduling method for low-Earth orbit satellite communication described in this invention, which sets the antenna mechanical constraints and verifies the scheduling solution, specifically includes the following steps:

[0031] Hash mapping collision detection is used to discretize the azimuth and elevation ranges of the antenna into a sparse hash table. According to the antenna specifications, the state of each cell in the sparse hash table is marked as valid or invalid. When a task submits a target angle, the target angle is queried from the sparse hash table to verify whether the target angle is valid or not, thereby verifying whether the scheduling solution is valid or not.

[0032] Preferably, the dynamic antenna resource scheduling method for low-Earth orbit satellite communication described in this application, which sets the coverage limit and verifies the scheduling solution, specifically includes the following steps:

[0033] Geometric reachability verification: Based on the antenna position and satellite orbit parameters, calculate whether the real-time elevation angle EI and azimuth angle Az of the satellite relative to the antenna exceed the threshold range using calculation formula group one; the calculation formula group one includes:

[0034] ;

[0035] ;

[0036] The This represents the northeast offset of the satellite in the ENU coordinate system of the ground station. This represents the northward offset of the satellite in the ENU coordinate system of the ground station. This represents the celestial offset of the satellite in the ENU coordinate system of the ground station.

[0037] If the calculated EI satisfies ,and If so, the geometric reachability verification is passed;

[0038] Signal strength verification: Based on the antenna position and satellite orbit parameters, the minimum receivable elevation angle and minimum receivable EIRP value at different times are pre-calculated, a coverage mapping table is generated, and then it is determined whether the current antenna elevation angle is greater than the minimum receivable elevation angle and whether the satellite EIRP value is greater than the minimum receivable EIRP value; if the current antenna elevation angle is greater than the minimum receivable elevation angle and the satellite EIRP value is greater than the minimum receivable EIRP value, then the verification is successful.

[0039] As a preferred embodiment of the present invention, a lightweight verification of physical constraints is formed through geometric reachability verification and signal strength verification, which transforms complex hardware constraints into efficient queries or calculations, balancing real-time performance and accuracy in satellite antenna scheduling. Combined with a dynamic antenna resource scheduling model based on genetic algorithm improvement, it enables faster and more feasible evaluation and iteration of massive potential scheduling schemes, thereby promoting dynamic multi-objective optimization within a very short time window and further improving the efficiency of obtaining the globally optimal scheduling scheme.

[0040] Preferably, the dynamic antenna resource scheduling method for low-Earth orbit satellite communication described in this invention, which sets the energy consumption limit and verifies the scheduling solution, specifically includes the following steps:

[0041] A steering energy consumption estimation model is established using the algorithm of Formula 9, which is:

[0042] ;

[0043] The This represents the change in the azimuth angle of the antenna. t represents the change in the antenna's elevation angle. delay Let k1 be the azimuth angle turning energy consumption coefficient of the antenna, k2 be the elevation angle turning energy consumption coefficient of the antenna, and k3 be the turning delay time energy consumption coefficient of the antenna. E max This represents the energy consumption threshold for a single turn of the antenna.

[0044] If the steering energy consumption E obtained by the scheduling solution is... switch If the total system energy consumption limit or single antenna energy consumption threshold is within the limit, then the scheduling solution is verified as an executable and valid scheduling solution.

[0045] As a preferred embodiment of the present invention, by establishing a turning energy consumption estimation model, the turning energy consumption of the scheduling solution can be estimated more efficiently, thereby verifying whether the scheduling solution complies with the upper limit of the total system energy consumption or the single antenna energy consumption threshold, ensuring that the final generated scheduling scheme is executable under actual hardware capabilities and power supply conditions, avoiding task interruption or equipment damage due to insufficient energy, further reducing the probability of scheduling scheme failure, and improving the effectiveness and reliability of antenna scheduling.

[0046] Preferably, in the dynamic antenna resource scheduling method for low-Earth orbit satellite communication described in this invention, the fitness function in S5 is a hierarchical game fitness function. Constructing the hierarchical game fitness function includes constructing a multi-objective game layer, specifically including the following steps:

[0047] Using the Nash equilibrium game model, the task completion rate F is calculated through Equation 2. task Maximize the energy consumption benefit F by calculating Equation 3. energy Minimize, and calculate the emergency response benefit F using Equation 4. emergency maximize;

[0048] The second calculation formula is:

[0049] ;

[0050] The W k T represents the weight of task priority. k completedT represents the actual completion time of the k-th task. k required The required duration for the k-th task;

[0051] The calculation formula three is as follows:

[0052] ;

[0053] The E a switch Let E be the turning energy consumption of the a-th antenna. a static E represents the static sustaining power consumption of the a-th antenna; a total Let A be the total system energy consumption threshold, and A be the total number of antennas.

[0054] The fourth calculation formula is:

[0055] ;

[0056] The D emergency For the delay time of the emergency mission, D req λ represents the maximum allowable delay time for the task, and λ is the urgency coefficient.

[0057] As a preferred embodiment of the present invention, a multi-objective fitness function is adopted. Through a hierarchical game architecture and dynamic weight allocation, multiple conflicting indicators can be optimized simultaneously. For example, the scheduling solution can be optimized to obtain a scheduling scheme that meets the requirements of multiple tasks, including task completion rate, energy consumption, and emergency response. This bridges the gap between optimization problems and genetic algorithms, further improving the performance and efficiency of the dynamic resource scheduling model in obtaining globally optimal solutions. It achieves higher task completion rate and lower energy consumption, significantly improving the comprehensiveness and practicality of the scheduling scheme.

[0058] Preferably, the dynamic scheduling method for antenna resources for low-Earth orbit satellite communication described in this invention, in constructing the hierarchical game fitness function, further includes constructing a robust optimization layer, specifically comprising the following steps:

[0059] An anti-interference factor is introduced, and the tolerance R of the scheduling scheme to satellite orbit errors and antenna failures is evaluated through Equation 5. total The calculation formula five is as follows:

[0060] ;

[0061] α, β, and γ are weighting coefficients, R orbit For the satellite's orbital fault tolerance, R hardware For hardware redundancy, R temporal For time-series elasticity;

[0062] The Rorbit The calculation is performed using formula six, which is:

[0063] ;

[0064] The θ dB Let Δθ be the beamwidth of the antenna. max θ represents the maximum angular deviation of the satellite's position. noise k4 is the sensitivity gain coefficient, which serves as the reference value for noise measurement.

[0065] The R hardware The calculation is performed using formula seven, which is:

[0066] ;

[0067] λ is the attenuation coefficient, N satandby(i) N represents the number of spare antennas for the i-th antenna. total N represents the total number of antennas. active t represents the number of antennas in use. switch(i) The switching time between the spare antennas of the i-th antenna;

[0068] The R temporal The calculation is performed using formula eight, which is:

[0069] ;

[0070] The T slack(k) T is the relaxation time for the k-th task. duration(k) The duration of the k-th task.

[0071] As a preferred embodiment of the present invention, by constructing a robust optimization layer, the scheduling scheme is evaluated and optimized, so that the current scheduling scheme has excellent robustness and can effectively cope with common track prediction errors and sudden single antenna failures, and can be approved for execution, thereby further improving the comprehensiveness and robustness of the present invention.

[0072] In summary, due to the adoption of the above technical solution, the beneficial effects of the present invention are:

[0073] By employing relative time coding, dynamic grouping, and conflict prediction mechanisms, the challenges of short time windows and numerous conflicts in low-Earth orbit (LEO) satellite communication are addressed, laying the foundation for multi-objective optimization. Reference time normalization and integer quantization transform complex time window management into efficient offset operations, significantly improving the real-time performance and scalability of dynamic satellite antenna resource scheduling. Furthermore, lightweight verification of physical constraints transforms complex hardware limitations into efficient queries or calculations, resolving the contradiction between real-time performance and accuracy in satellite antenna scheduling and providing underlying support for dynamic multi-objective optimization. Combined with fitness function optimization, the complex "multiple requirements" in satellite resource scheduling are addressed. Finally, by integrating genetic algorithms with quantum coding and conflict prediction technologies, higher task completion rates, lower energy consumption, and stronger fault tolerance are achieved in the scheduling scheme, reducing the possibility of antenna resource scheduling scheme failure and significantly improving the applicability of scheduling schemes with extreme window times in LEO satellite communication. Attached Figure Description

[0074] Figure 1 This is a flowchart illustrating the dynamic scheduling method for antenna resources in low-Earth orbit satellite communication according to the present invention.

[0075] Figure 2 This is a schematic diagram illustrating the specific steps of the component antenna resource dynamic scheduling model of the present invention. Detailed Implementation

[0076] The present invention will now be described in detail with reference to the accompanying drawings.

[0077] To make the objectives, technical solutions, and advantages of this invention clearer, the invention will be further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative and not intended to limit the invention.

[0078] Example 1:

[0079] refer to Figure 1 and Figure 2 As shown, this embodiment discloses a method for dynamic scheduling of antenna resources for low-Earth orbit satellite communication, including the following steps:

[0080] S1. Manage antenna static information and real-time status information, acquire antenna resources; parse user task requirements, and acquire task data;

[0081] S2. Based on antenna resources and task data, construct a dynamic scheduling model for antenna resources using a genetic algorithm. This includes the following steps:

[0082] S21. Design Chromosome Structure: Employ dynamic spatiotemporal quantum hybrid coding composed of time, antenna, and satellite to compile the scheduling scheme into a chromosome composed of multiple gene segments, with each gene segment corresponding to a task unit; multiple gene segments are linearly spliced ​​according to the order of scheduling execution, with time as the implicit sequence, to form the chromosome; the structure of the gene segment includes: task ID, satellite ID, ground station ID, antenna ID, start time, duration, antenna elevation angle, antenna azimuth angle, and task priority flag;

[0083] S22. Use relative time encoding and dynamic window grouping to process the time window;

[0084] Relative time encoding specifically includes the following steps: using the starting point of the scheduling cycle as a reference, offset encoding is used to divide the absolute time into fixed units, and the task time is represented by an integer offset;

[0085] Dynamic grouping of windows includes the following steps: If the task data is a long task, the long task is divided into multiple gene segments, and at least two units of protection time are inserted between adjacent gene segments, so that the task duration of each gene segment is ≤ minimum antenna dwell time + satellite continuous visibility time.

[0086] S3. Obtain the scheduling solution for antenna resource allocation through a dynamic scheduling model;

[0087] S4. Based on antenna resources and mission data, set physical constraints, verify whether the scheduling solution violates the physical constraints, and then retain the valid scheduling solutions in the scheduling solutions; physical constraints include: antenna mechanical limitations, coverage limitations, and energy consumption limitations;

[0088] S5. Use the fitness function to optimize the effective scheduling solution and generate a dynamic scheduling scheme for antenna resources.

[0089] In this invention, low-Earth orbit (LEO) satellites are understood to operate in orbits 160 to 2000 kilometers above the Earth's surface. The dynamic antenna resource scheduling method of this invention is more suitable for application scenarios where LEO satellites have short satellite time windows, but is not limited to LEO satellite applications.

[0090] The present invention manages antenna static and real-time status information by aggregating basic antenna information, including antenna name, frequency band, polarization, type, antenna rotation range, antenna rotation speed, and antenna position. Furthermore, it accesses real-time antenna status information, including antenna health monitoring information, antenna operating status information, and antenna fault information. This allows for the acquisition of antenna resource data using conventional antenna resource management and analysis methods. For example, a scheduling knowledge base and unified data model are constructed to integrate scattered and heterogeneous raw data into standardized information that scheduling algorithms can directly understand and process. By aggregating basic satellite information, including satellite number, satellite name, satellite orbital parameters, and EIRP contour maps of satellite beam coverage, the aggregation of antenna and satellite basic information, combined with the satellite basic information, determines which antenna can physically serve which satellite. Finally, by referring to conventional receiving task analysis methods, the task requirement data is analyzed to obtain task data, such as satellite number, frequency band, and time window information.

[0091] It should be noted that the traditional genetic algorithm model mainly consists of three major modules: encoding, fitness function, and genetic operator. The antenna resource dynamic scheduling method of the present invention is an improvement on the traditional genetic algorithm. Through dynamic spatiotemporal quantum encoding design, it aims to efficiently express the complex constraints of low-orbit satellite multi-ground station antenna resource scheduling. Among them, gene segment, chromosome and dynamic spatiotemporal quantum hybrid encoding can be improved on the encoding of the traditional genetic algorithm, specifically including chromosome structure design and time window processing in steps S21 and S22. For example, the three scheduling resources—antenna, satellite, and time—are mapped as superimposed and evolvable quantum gene segments. This enables flexible representation and global optimization of multi-task parallel conflicts in the scheduling model, breaking through the rigid constraints of traditional coding and laying the foundation for adaptive capabilities for subsequent efficient and dynamic scheduling decisions. Specifically, a three-dimensional dynamic spatiotemporal quantum hybrid coding system of "time-antenna-satellite" is adopted. Each chromosome represents a complete scheduling scheme, and each chromosome is composed of multiple gene segments connected in series. The gene segments use the timeline as the implicit sequence, treating each communication task unit as an independent gene segment, and linearly splicing them according to their scheduling execution order to form a complete chromosome. This chromosome directly represents a complete scheme covering the entire scheduling cycle and containing all task execution sequences, realizing the compilation of gene segments into chromosomes.

[0092] Specifically, in the gene segment encoding design of this invention, S21 further includes setting the field length of the gene segment to variable-length encoding, specifically including the following steps: using variable-length encoding for the field length of the gene segment, then allocating bits to each field according to the requirements of the task, and reserving extension bits for adding new constraints. For example, weather status and modulation mode can be added to the field length; it should be noted that variable-length encoding is understood as encoding the field length to a length that can be allocated bits according to actual needs, relative to a fixed-length encoding design, and can be achieved based on the data parameter encoding architecture.

[0093] Specifically, in this invention, S21 further includes designing deformable gene segments, specifically comprising the following steps: setting the coding length of the gene segment so that the coding length dynamically expands / contracts according to the urgency level of the task data. For example, urgent task gene segments are compressed into source data pointers, and the data structure can be designed as [Task ID][Satellite ID][Priority], while detailed parameters are retained for non-urgent tasks.

[0094] Specifically, in this invention, S21 further includes topology optimization of the coding space, specifically comprising the following steps: using a graph neural network to model the chromosome graph structure, embedding the relationships between antenna nodes, satellite nodes, and mission nodes into the coding space modeled by the graph structure, thereby constructing a topology graph containing resource nodes, mission nodes, and constraint relationship edges. For example, constructing a topology graph containing three types of nodes: resource nodes, identifying antenna attributes such as location, frequency band, and rotation capability; mission nodes, identifying satellite communication requirements such as time window, elevation angle, and priority; and constraint relationship edges, identifying spatiotemporal conflicts, physical reachability, signal quality limitations, and energy consumption limitations.

[0095] It should be noted that the relative time encoding in S22 is understood as a time encoding method. By uniformly normalizing absolute time into an integer offset relative to the scheduling starting point, the true absolute time is mapped to an integer that is easy for the genetic algorithm to process. This allows the improved scheduling model to quickly detect and adjust the position of gene segments on the time axis through simple integer operations, thereby directly avoiding or resolving overlapping time windows in crossover and mutation operations, achieving efficient conflict elimination. Long tasks in the dynamic window grouping are understood as those whose required duration for a communication task exceeds the upper limit of the executable time jointly determined by the "minimum stable dwell time of the antenna" and the "continuous visible time of the satellite," thus triggering dynamic grouping.

[0096] Step S3 of this invention improves the encoding method and encoding data based on steps S21 and S22. Based on the execution steps of the traditional genetic algorithm, it can obtain the scheduling solution for antenna resource allocation. The scheduling solution can be understood as one or more resource allocation schemes that can be obtained through the calculation and execution of the dynamic scheduling model based on the genetic algorithm.

[0097] Step S4 of the present invention, through antenna mechanical constraints, coverage range constraints and energy consumption constraints, can realize the constraint judgment of the resource allocation scheme obtained in the scheduling model. Based on the calculation and execution process of the genetic algorithm, the scheduling scheme represented by each chromosome in the initial population is parallelized and lightweight for feasibility verification, eliminating all individuals that violate any physical constraints, thereby ensuring that the next evolutionary iteration is a scheduling solution without physical conflicts.

[0098] The antenna mechanical limitations described in this invention can be provided to the scheduling model through conventional detection methods. This invention preferably uses hash mapping collision detection to verify the scheduling solution, specifically including the following steps: discretizing the azimuth and elevation ranges of the antenna into a sparse hash table; marking the state of each cell in the sparse hash table as valid or invalid according to the antenna specifications; when a task submits a target angle, querying the sparse hash table to see if the target angle is valid or not, thereby verifying whether the scheduling solution is valid or not.

[0099] The coverage range limitation described in this invention can be achieved through conventional methods. This invention preferably achieves this through geometric reachability verification and signal strength verification, specifically including the following steps: Geometric reachability verification: Based on the antenna position and satellite orbit parameters, calculate whether the real-time elevation angle EI and azimuth angle Az of the satellite relative to the antenna exceed the threshold range using calculation formula group one; calculation formula group one includes:

[0100] ;

[0101] ;

[0102] The This represents the northeast offset of the satellite in the ENU coordinate system of the ground station. This represents the northward offset of the satellite in the ENU coordinate system of the ground station. This represents the celestial offset of the satellite in the ENU coordinate system of the ground station.

[0103] If the calculated EI satisfies ,and If so, the geometric reachability verification is passed;

[0104] Signal strength verification: Based on the antenna position and satellite orbit parameters, the minimum receivable elevation angle and minimum receivable EIRP value at different times are pre-calculated, a coverage mapping table is generated, and then it is determined whether the current antenna elevation angle is greater than the minimum receivable elevation angle and whether the satellite EIRP value is greater than the minimum receivable EIRP value. If the current antenna elevation angle is greater than the minimum receivable elevation angle and the satellite EIRP value is greater than the minimum receivable EIRP value, then the verification passes. Here, the EIRP value is understood as the limit of effective isotropic radiated power.

[0105] The energy consumption limitation described in this invention can be achieved through conventional constraint settings. Preferably, the energy consumption limitation setting and scheduling solution verification of this invention specifically includes the following steps:

[0106] Using the algorithm of Formula 9, a steering energy consumption estimation model is established. Formula 9 is as follows:

[0107] ;

[0108] The This represents the change in the azimuth angle of the antenna. t represents the change in the antenna's elevation angle. delay Let k1 be the azimuth angle turning energy consumption coefficient of the antenna, k2 be the elevation angle turning energy consumption coefficient of the antenna, and k3 be the turning delay time energy consumption coefficient of the antenna. E max This represents the energy consumption threshold for a single turn of the antenna.

[0109] If the steering energy consumption E obtained from the scheduling solution is... switch If the total system energy consumption limit or single antenna energy consumption threshold is within the limit, then the scheduling solution is verified as an executable and effective scheduling solution.

[0110] The fitness value represents the direction and indicator of selection in a genetic algorithm. Genetic algorithms differ from other random, blind searches in that they are directed searches. The fitness function bridges the gap between the optimization problem and the genetic algorithm, and its design directly impacts the performance and efficiency of the genetic algorithm in solving practical problems. Generally, the fitness function is designed based on the objective function of the problem. The fitness function selects individuals by evaluating their fitness within the population. The fitness function of this invention can be implemented using traditional fitness functions based on the coding design. Preferably, a hierarchical game fitness function can be used. Constructing a hierarchical game fitness function includes building a multi-objective game layer, specifically including the following steps:

[0111] Using the Nash equilibrium game model, the task completion rate F is calculated through Equation 2. task Maximize the energy consumption benefit F by calculating Equation 3. energy Minimize, and calculate the emergency response benefit F using Equation 4. emergency maximize;

[0112] The second calculation formula is:

[0113] ;

[0114] The W k T represents the weight of task priority. k completed T represents the actual completion time of the k-th task. k required The required duration for the k-th task;

[0115] Furthermore, the scheduling scheme should be used to minimize the actual completion time T of each task. k completed Approximate to or equal to its required duration T k required and for high-priority tasks W k Allocate sufficient resources to achieve F task maximize.

[0116] The calculation formula three is as follows:

[0117] ;

[0118] The E a switch Let E be the turning energy consumption of the a-th antenna, which is usually proportional to the turning angle. a static Let E be the static sustaining power consumption of the a-th antenna, typically the product of power consumption per unit time and mission duration; a total The total system energy consumption threshold can be dynamically adjusted based on the power supply capacity, where A is the total number of antennas.

[0119] Furthermore, by optimizing task orchestration to reduce the total number of turns and the antenna turning angle Δθ, the sum of total turning and static energy consumption is reduced, making it much smaller than the system's total energy consumption threshold E. max total To achieve F energy Minimize;

[0120] The fourth calculation formula is:

[0121] ;

[0122] The D emergency For the delay time of the emergency mission, D reqλ represents the maximum allowable delay time for a task, referring to the maximum tolerable delay between the theoretically optimal start time and the actual execution time. Exceeding this time may cause the task to fail or its performance to degrade significantly; λ is the urgency coefficient, for example, the urgency coefficient for high-priority tasks is 0.5, and the urgency coefficient for ordinary-priority tasks is 0.1.

[0123] This is achieved by prioritizing urgent tasks and ensuring their actual delay time D. emergency Strictly less than the maximum allowable delay time D req This makes the exponent term in the formula approach zero, thus making F... emergency Approaching 1, thus maximizing it.

[0124] It should be noted that the Nash equilibrium game model described herein is understood as an improvement on the Nash equilibrium calculation method in game theory, which can be achieved by combining calculation formulas two to four.

[0125] Specifically, based on the constructed hierarchical game fitness function, S5 constructs the fitness function and also includes constructing a robust optimization layer, specifically including the following steps:

[0126] An anti-interference factor is introduced, and the tolerance R of the scheduling scheme to satellite orbit errors and antenna failures is evaluated through Equation 5. total Formula 5 is:

[0127] ;

[0128] α, β, and γ are weighting coefficients, R orbit For the satellite's orbital fault tolerance, R hardware For hardware redundancy, R temporal For time-series elasticity;

[0129] The R orbit Calculation using formula six yields the following formula:

[0130] ;

[0131] The θ dB The beamwidth of the antenna, in degrees, Δθ max θ represents the maximum angular deviation of the satellite's position. noise For measuring the noise reference value, for example, the noise reference value is 0.5°, and k4 is the sensitivity gain coefficient, for example, k4=3;

[0132] The R hardware Calculations are performed using formula seven, which is:

[0133] ;

[0134] λ is the attenuation coefficient, N satandby(i) N represents the number of spare antennas for the i-th antenna. total N represents the total number of antennas. active t represents the number of antennas in use. switch(i) The switching time between the spare antennas of the i-th antenna;

[0135] The R temporal Calculation using formula eight yields the following formula:

[0136] ;

[0137] The T slack(k) T is the relaxation time for the k-th task. duration(k) The duration of the k-th task.

[0138] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions and improvements made within the spirit and principles of the present invention should be included within the protection scope of the present invention.

Claims

1. A method for dynamic scheduling of antenna resources for low-Earth orbit satellite communication, characterized in that, Includes the following steps: S1. Manage antenna static information and real-time status information, acquire antenna resources; parse user task requirements, and acquire task data; S2. Based on the antenna resources and the task data, construct a dynamic scheduling model for antenna resources using a genetic algorithm, specifically including the following steps: S21. Designing the chromosome structure: A dynamic spatiotemporal quantum hybrid coding system composed of time, antenna, and satellite is adopted to compile the scheduling scheme into a chromosome composed of multiple gene segments, with each gene segment corresponding to a task unit; multiple gene segments are linearly spliced ​​according to the order of scheduling execution, with time as the implicit sequence, to form the chromosome; the structure of the gene segment includes: task ID, satellite ID, ground station ID, antenna ID, start time, duration, antenna elevation angle, antenna azimuth angle, and task priority flag; S22. Perform relative time encoding and dynamic window grouping on the time window, specifically including the following steps: Based on the starting point of the scheduling cycle, offset encoding is used to divide the absolute time into fixed units and represent the task time with integer offsets to achieve relative time encoding of the time window; If the task data is a long task, the long task is divided into multiple gene segments, and at least two units of protection time are inserted between adjacent gene segments, so that the task duration of each gene segment is ≤ the minimum antenna dwell time + the satellite's continuous visibility time, thereby realizing dynamic grouping of time windows. S3. Obtain the scheduling solution for antenna resource allocation through a dynamic scheduling model; S4. Based on the antenna resources and the task data, set physical constraints, verify whether the scheduling solution violates the physical constraints, and then retain the valid scheduling solutions in the scheduling solutions; the physical constraints include: antenna mechanical limitations, coverage limitations, and energy consumption limitations; S5. Optimize the effective scheduling solution using the fitness function to generate a dynamic scheduling scheme for antenna resources.

2. The method for dynamic scheduling of antenna resources for low-Earth orbit satellite communication according to claim 1, characterized in that, S21 further includes setting the field length of the gene segment to variable-length encoding, specifically including the following steps: The field length of the gene segment is encoded using variable length encoding. Bits are then allocated to each field according to the requirements of the task, and extension bits are reserved for adding new constraints.

3. The method for dynamic scheduling of antenna resources for low-Earth orbit satellite communication according to claim 1, characterized in that, S21 further includes designing deformable gene segments, specifically including the following steps: The coding length of the gene segment is set so that the coding length dynamically expands / contracts according to the urgency of the task data (low / high).

4. The method for dynamic scheduling of antenna resources for low-Earth orbit satellite communication according to claim 1, characterized in that, S21 further includes topology optimization of the coding space, specifically including the following steps: A graph neural network is used to model the chromosome, embedding the relationships between antenna nodes, satellite nodes, and mission nodes into the encoding space of the graph model, thereby constructing a topological graph containing resource nodes, mission nodes, and constraint relationship edges.

5. The method for dynamic scheduling of antenna resources for low-Earth orbit satellite communication according to claim 1, characterized in that, Setting the antenna mechanical limitations and verifying the scheduling solution specifically includes the following steps: Hash mapping collision detection is used to discretize the azimuth and elevation ranges of the antenna into a sparse hash table. According to the antenna specifications, the state of each cell in the sparse hash table is marked as valid or invalid. When a task submits a target angle, the target angle is queried from the sparse hash table to verify whether the target angle is valid or not, thereby verifying whether the scheduling solution is valid or not.

6. The method for dynamic scheduling of antenna resources for low-Earth orbit satellite communication according to claim 1, characterized in that, Setting the coverage limit and verifying the scheduling solution specifically includes the following steps: Geometric reachability verification: Based on the antenna position and satellite orbit parameters, calculate whether the real-time elevation angle EI and azimuth angle Az of the satellite relative to the antenna exceed the threshold range using calculation formula group one; the calculation formula group one includes: ； ； The This represents the eastward offset of the satellite in the ENU coordinate system of the ground station. This represents the northward offset of the satellite in the ENU coordinate system of the ground station. This represents the celestial offset of the satellite in the ENU coordinate system of the ground station. If the calculated EI satisfies ,and If so, the geometric reachability verification is passed; Signal strength verification: Based on the antenna position and satellite orbit parameters, the minimum receivable elevation angle and minimum receivable EIRP value at different times are pre-calculated, a coverage mapping table is generated, and then it is determined whether the current antenna elevation angle is greater than the minimum receivable elevation angle and whether the satellite EIRP value is greater than the minimum receivable EIRP value; if the current antenna elevation angle is greater than the minimum receivable elevation angle and the satellite EIRP value is greater than the minimum receivable EIRP value, then the verification is successful.

7. The method for dynamic scheduling of antenna resources for low-Earth orbit satellite communication according to any one of claims 1-6, characterized in that, The fitness function in S5 is a hierarchical game fitness function. Constructing the hierarchical game fitness function includes constructing a multi-objective game layer, specifically including the following steps: Using Nash equilibrium game model, the task completion rate F task is maximized by calculation formula three, the steering energy consumption benefit F energy is minimized by calculation formula four, and the emergency task response benefit F emergency is maximized by calculation formula four. The second calculation formula is: ； The W k T represents the weight of task priority. k completed T represents the actual completion time of the k-th task. k required The required duration for the k-th task; N is the total number of tasks; The calculation formula three is as follows: ； The E a switch Let E be the turning energy consumption of the a-th antenna. a static E represents the static sustaining power consumption of the a-th antenna. max total Let A be the total system energy consumption threshold, and A be the total number of antennas. The fourth calculation formula is: ； The D emergency For the delay time of the emergency mission, D req λ represents the maximum allowable delay time for the task, and λ is the urgency coefficient.

8. The method for dynamic scheduling of antenna resources for low-Earth orbit satellite communication according to claim 7, characterized in that, Constructing the hierarchical game fitness function further includes constructing a robust optimization layer, specifically including the following steps: An anti-interference factor is introduced, and the tolerance R of the scheduling scheme to satellite orbit errors and antenna failures is evaluated through Equation 5. total The calculation formula five is: ； α, β, and γ are weighting coefficients, R orbit For the satellite's orbital fault tolerance, R hardware For hardware redundancy, R temporal For time-series elasticity; The R orbit The calculation is performed using formula six, which is: ； The θ dB Let Δθ be the beamwidth of the antenna. max θ represents the maximum angular deviation of the satellite's position. noise k4 is the sensitivity gain coefficient, which serves as the reference value for noise measurement. The R hardware The calculation is performed using formula seven, which is: ； λ is the attenuation coefficient, N satandby(i) N represents the number of spare antennas for the i-th antenna. total N represents the total number of antennas. active t represents the number of antennas in use. switch(i) The switching time between the spare antennas of the i-th antenna; The R temporal The calculation is performed using formula eight, which is: ； The T slack(k) T is the relaxation time for the k-th task. duration(k) The duration of the k-th task.

Images