LEO satellite dynamic beam management and resource adaptation method

Abstract

The invention discloses an LEO satellite dynamic beam management and resource adaptation method, and relates to the field of multi-beam low-orbit satellite communication. In a communication scene of a low-orbit constellation and a user, a satellite calculates the minimum number of opened beams according to the position of the user by adopting a beam closing algorithm based on an improved ant colony under the condition of full coverage of the low-orbit constellation, and is bound with the user for downlink communication after being opened according to the number, and the user sends a service downloading request; calculating an information transmission rate; and the satellites carry out power distribution based on a continuous convex approximation dynamic power distribution algorithm according to the information transmission rate and the request of the user, and optimization is carried out by taking maximization of the system throughput as a target under the condition that the power is limited. And finally, in combination with the actual opening number of the beams and the allocated power value, the user selects a service satellite and allocates communication resources so as to transmit services. According to the invention, the beam resources are saved, and the performance is improved in the aspect of optimizing the communication capacity.

Description

Technical Field

[0001] The present invention relates to the field of multi-beam low-orbit satellite communications, and in particular to a method for dynamic beam management and resource adaptation of a LEO satellite. Background Art

[0002] The existing traditional resource allocation strategy is not applicable to the user service needs that are non-uniformly distributed in time and space, and does not take into account the problem of beam overlapping coverage in high-latitude areas of near-polar orbit constellations.

[0003] The large-scale commercial use of 5G technology has promoted the research of 6G communication technologies. As one of the candidate technologies for 6G, satellite communication has many advantages: 1) wide coverage, which can provide communication services for remote areas such as mountainous areas and offshore areas; 2) high reliability and not easily affected by natural disasters; 3) flexible networking and large communication capacity.

[0004] Globally, low-orbit satellite communication systems are playing an increasingly important role. As a supplement to traditional terrestrial communications, satellite communications can serve millions of users around the world at the same time, and can maintain communication quality even if the users are thousands of kilometers apart. In addition, when terrestrial communication networks are damaged by natural disasters such as earthquakes and tsunamis, satellite communications can also be used as one of the communication options to ensure the normal processing of user services.

[0005] However, the development of low-orbit satellite communication systems has also brought some problems: the overlapping coverage of beams in high-latitude areas of near-polar orbit low-orbit constellations and the dynamic changes in network topology may cause serious interference to the same-frequency cells and even cause a large amount of beam resource waste; the traditional fixed resource allocation method cannot adapt to the characteristics of uneven business distribution, making it impossible to fully utilize system resources.

[0006] Taking into account the allocation of multi-source resources, the impact of co-frequency interference, and uneven business distribution, beam management and multi-dimensional resource on-demand adaptation solutions are formulated to improve user service quality and system resource utilization, which is conducive to promoting the development of satellite communications and is also of certain significance to the construction of an integrated air-space-ground information network.

[0007] This paper studies how to design dynamic beam closing and power allocation algorithms for multi-beam low-orbit constellation downlink communication scenarios. Summary of the invention

[0008] The present invention designs a LEO satellite dynamic beam management and resource adaptation method, which can dynamically switch beams according to the user coverage beam multiplicity and actual business needs, and adjust power resource allocation; thereby improving user service quality and system resource utilization.

[0009] The LEO satellite dynamic beam management and resource adaptation method comprises the following specific steps:

[0010] Step 1: Build a communication scenario including low-orbit constellations and users. The satellite acts as an edge node and provides services to ground terminal users through downlinks.

[0011] The number of satellites in the low-orbit satellite constellation is s a ; The number of beams emitted by each satellite is k a ; Each beam center in the low-orbit constellation coverage area is distributed with a terminal user; the total number of users n a =s a *k a .

[0012] Step 2: The satellite uses an improved ant colony-based beam closing algorithm based on the user's location, and optimizes with the goal of minimizing the number of open beams when the low-orbit constellation is fully covered.

[0013] The optimization objective function is:

[0014]

[0015] Among them, x s,k ∈{0,1} represents the two-dimensional state representation of beam k of satellite s, x s,k =0 means beam is off, x s,k =1 means the beam is on. s,k,i ∈{0,1} represents the two-dimensional state representation of whether user i is covered by beam k of satellite s, β s,k,i =1 means user i is covered by beam k under satellite s, β s,k,i =0 means that user i is not covered by beam k under satellite s.

[0016] The improved ant colony beam closing algorithm is as follows:

[0017] 1) The set of users that can be covered by each beam of each satellite is formed into a cluster; initially, m ant variables are placed on all beams of all satellites in a round-robin manner;

[0018]

[0019] 2) All m ants calculate in parallel. For ant ω in the current round, when it is placed on beam k, the beam k is added to the solution set D corresponding to the ant ω , and remove the users covered by the beam from the user set n in the scene a Delete from;

[0020] 3) Ant ω based on state transition probability Select the next beam;

[0021] The selection rule is: set the beam solution of ant ω to Dω The users covered by all beams in the list and the users covered by the selected beam are combined to obtain the beam with the largest number of union elements, which is the next beam to be selected.

[0022] Its value is calculated as follows:

[0023]

[0024] in, represents the pheromone concentration left by the beam solution of ant ω at the current moment; S j represents the beam solution set that ant ω will select next time. a represents the relative importance of the residual information on each beam solution set in guiding the ant colony search process. The larger its value, the greater the possibility that the ant will select the beam solution set. β represents the relative importance of heuristic information in guiding the ant colony search process; h is the set of beam solution sets that ant ω has not yet reached in this cycle; represents the sum of the pheromone concentrations left by the ants on the set of beam solutions that have not been reached;

[0025] 4) The user covered by the next beam selected by ant ω is selected from the user set n a Delete; return to 3) continue based on state transition probability Select the next beam; until all ants have set user set n in the current round cycle. a All users in the are deleted, and the current cycle ends;

[0026] 5) From the current round of cycles, select the solution set with the least number of elements in the solution sets of each ant as the optimal solution set D for this round.

[0027] 6) Use the optimal solution set D of this round to update the pheromone concentration;

[0028] The formula for calculating pheromone concentration is:

[0029]

[0030] Among them, ρ represents the volatility of pheromone; σ represents the penalty coefficient that changes with the number of iterations and is Gaussian distributed; [τ min ,τ max ] indicates the range of pheromone concentration; Indicates that ant ω stays in the beam solution set S at time t+1 j The pheromone concentration on.

[0031] 7) Return to 2) and restart the next cycle until the final number of iterations is reached, and retain the last optimal solution set D of this round, which is the minimum beam set that satisfies the full coverage of low-orbit satellites.

[0032] Step 3: At the current moment, according to the optimized minimum number of beams, the beam is turned on and bound to the user for downlink communication. The user sends a service download request, the satellite receives and analyzes the user's position and service request, and calculates the channel gain of each user and the received signal-to-interference-noise ratio;

[0033] For users Channel gain on beam k of satellite s

[0034]

[0035] in is the downlink path loss, is the gain of the satellite transmitting antenna, η is the off-axis angle; is the antenna gain received by the user.

[0036] Then, calculate the user The signal-to-interference-to-noise ratio received on beam k of satellite s is

[0037]

[0038] Assign satellite s to users The power on beam k, Indicates that except for the users under beam k of satellite s, other satellites Beam The set of users occupying the same frequency. For users On its satellite Beam For users the interference caused; For satellite Assign to user In the beam The power on For users Relative to the interference satellite Beam N0 is the Gaussian white noise power on a single channel.

[0039] Step 4: For users After allocating downlink channel resources, the Shannon formula is used to calculate the user Channel capacity on satellite s beam k The information transmission rate is thus obtained

[0040] The channel capacity is:

[0041]

[0042] B i Indicates the bandwidth allocated to the current user;

[0043] The information transmission rate is:

[0044]

[0045] Step 5: The satellite allocates power based on the dynamic power allocation algorithm of continuous convex approximation according to the information transmission rate and user requests, and optimizes the system throughput under power constraints.

[0046] The optimization objective function is:

[0047]

[0048] K represents the set of all beams of low-orbit satellite s; P tot Assign satellite s to users Maximum power in all beams;

[0049] The dynamic power algorithm of continuous convex approximation is as follows:

[0050] 1) Convert the objective function l2 into a special convex function difference form:

[0051] f(P)=g(P)-h(P)

[0052] Among them, g(P) and h(P) are respectively expressed as:

[0053]

[0054] Where P = {p s,k,i |1≤i≤n a} is the power set allocated to each downlink user. After being converted into convex function difference form, g(P) is a concave function, -h(P) is a convex function, and f(P) is still a non-convex function.

[0055] 2) For time slot t, the concave function h(P) is further expanded by a first-order Taylor expansion, and the objective function l2 of the s-th iteration is approximately the lower bound of the original problem of maximization:

[0056] The initial value of s is 1.

[0057] The first-order Taylor expansion is:

[0058]

[0059] Where P[s] is the power vector allocated to the downlink subchannel at the sth convex approximation iteration, is the derivative of h(P) at P[s].

[0060] definition Available is the lower bound of g(P)-h(P). Therefore, in the sth iteration, the objective function l2 is approximately the lower bound of the original problem:

[0061]

[0062] 3) Use the convex optimization toolkit CVX to solve the lower bound of the original maximization problem and obtain the optimal power P[s] at the sth iteration;

[0063] 4) The power P[s] is used as the initial value of each downlink channel power, and the next iteration is performed until the convergence condition is met to obtain the optimal power allocation value of time slot t.

[0064] 5) Substitute the optimal power allocation value P[s] of time slot t into the information transmission rate formula to obtain the total downlink transmission rate of the system in the current time slot t And determine the total rate Total downlink transmission rate of the system compared to the previous time slot Is the difference between them less than the threshold ε? If so, output the optimal power allocation value P[s]; otherwise, return to 2) and recalculate the lower bound of the original maximization problem for the next time slot.

[0065] Step 6: Based on the actual number of beams opened and the allocated power value, the user selects the service satellite and allocates communication resources to transmit the service.

[0066] The advantages of the present invention are:

[0067] 1) A dynamic beam management and resource adaptation method for LEO satellites is proposed. A beam closing algorithm based on an improved ant colony and a power allocation algorithm based on continuous convex approximation are designed for low-orbit constellations, which realizes the dynamic management of multi-beam satellite communication resources. It has better performance in the downlink communication scenario of multi-beam low-orbit satellite systems and can allocate communication resources to users more flexibly.

[0068] 2) A dynamic beam management and resource adaptation method for LEO satellites. Combining the idea of ​​redundant beam closing, a dynamic power allocation algorithm is proposed to perform dynamic beam closing and resource scheduling with the optimization goal of maximizing the total system throughput. Compared with traditional algorithms, it can not only save certain beam resources, but also has better performance in optimizing communication capacity. BRIEF DESCRIPTION OF THE DRAWINGS

[0069] Figure 1 It is a flow chart of a LEO satellite dynamic beam management and resource adaptation method of the present invention;

[0070] Figure 2 It is a schematic diagram of a multi-beam low-orbit constellation downlink communication scenario constructed by the present invention;

[0071] Figure 3 This is a flow chart of downlink communication of a low-orbit satellite in an embodiment of the present invention;

[0072] Figure 4 It is a flow chart of dynamic beam closing based on improved ant colony algorithm of the present invention;

[0073] Figure 5 It is a flow chart of dynamic power allocation based on SCA algorithm of the present invention. DETAILED DESCRIPTION

[0074] The present invention will be further described in detail below with reference to the accompanying drawings and embodiments.

[0075] At present, multi-beam satellite IoT faces the demand for downlink communication of user services. In this scenario, as satellites move to higher latitudes, the distance between satellites gradually decreases, and the overlapping area between beams gradually increases. This aggravates co-frequency interference in high-latitude areas and significantly reduces user QoS. At the same time, due to the obvious non-uniform distribution of user and service demand, the traditional fixed resource allocation strategy will lead to a waste of communication resources.

[0076] Based on this, the present invention proposes a dynamic beam management and resource adaptation technology for LEO Satellite, which divides the multi-beam constellation downlink resource scheduling problem into two processes: dynamic beam closing and power allocation according to the user coverage beam multiplicity and actual business needs.

[0077] like Figure 1 As shown, the specific steps are as follows:

[0078] Step 1: Build a communication scenario including low-orbit constellations and users. The satellite acts as an edge node and provides services to ground terminal users through downlinks.

[0079] The present invention considers the scenario where a low-orbit constellation provides services to ground terminal users through a downlink, such as Figure 2 For the downlink, the low-orbit satellite constellation uses s a Satellites s = {s1,s2,s3...s a}, and k for each satellite a beams k = {1, 2, ..., k a}Achieve global coverage. There is one terminal user distributed at each beam center in the coverage area of ​​the low-orbit constellation; therefore, the total number of users na =s a *k a Satellites serve as edge nodes, and various applications on user terminal devices will request to download data from the satellite side from time to time.

[0080] Step 2: The satellite uses an improved ant colony-based beam closing algorithm based on the user's location, and optimizes with the goal of minimizing the number of open beams when the low-orbit constellation is fully covered.

[0081] The optimization objective function is:

[0082]

[0083] Among them, x s,k ∈{0,1} represents the two-dimensional state representation of beam k of satellite s, x s,k =0 means beam is off, x s,k =1 means the beam is on. s,k,i ∈{0,1} represents the two-dimensional state representation of whether user i is covered by beam k of satellite s, β s,k,i =1 means user i is covered by beam k under satellite s, β s,k,i =0 means that user i is not covered by beam k under satellite s.

[0084] Improve the beam closing algorithm of ant colony, such as Figure 4 As shown, specifically:

[0085] 1) The set of users that can be covered by each beam of each satellite is formed into a cluster; initially, m ant variables are placed on all beams of all satellites in a round-robin manner;

[0086]

[0087] 2) All m ants calculate in parallel. For ant ω in the current round, when it is placed on beam k, the beam k is added to the solution set D corresponding to the ant ω , and remove the users covered by the beam from the user set n in the scene a Delete from;

[0088] 3) Ant ω based on state transition probability Select the next beam;

[0089] The selection rule is: set the beam solution of ant ω to D ω The users covered by all beams in the list and the users covered by the selected beam are combined to obtain the beam with the largest number of union elements, which is the next beam to be selected.

[0090] Its value is calculated as follows:

[0091]

[0092] in, represents the pheromone concentration left by the beam solution of ant ω at the current moment; S j represents the beam solution set that ant ω will select next time. a represents the relative importance of the residual information on each beam solution set in guiding the ant colony search process. The larger its value, the greater the possibility that the ant will select the beam solution set. β represents the relative importance of heuristic information in guiding the ant colony search process; h is the set of beam solution sets that ant ω has not yet reached in this cycle; represents the sum of the pheromone concentrations left by the ants on the set of beam solutions that have not been reached;

[0093] 4) The user covered by the next beam selected by ant ω is selected from the user set n a Delete; return to 3) continue based on state transition probability Select the next beam; until all ants have set user set n in the current round cycle. a All users in the are deleted, and the current cycle ends;

[0094] 5) From the current round of cycles, select the solution set with the least number of elements in the solution sets of each ant as the optimal solution set D for this round.

[0095] 6) Use the optimal solution set D of this round to update the pheromone concentration;

[0096] The formula for calculating pheromone concentration is:

[0097]

[0098] Among them, ρ represents the volatility of pheromone; σ represents the penalty coefficient that changes with the number of iterations and is Gaussian distributed; [τ min ,τ max ] indicates the range of pheromone concentration; Indicates that ant ω stays in the beam solution set S at time t+1 j The pheromone concentration on.

[0099] 7) Return to 2) and restart the next cycle until the final number of iterations is reached, and retain the last optimal solution set D of this round, which is the minimum beam set that satisfies the full coverage of low-orbit satellites.

[0100] Step 3: At the current moment, according to the optimized minimum number of beams, the beam is turned on and bound to the user for downlink communication. The user sends a service download request, the satellite receives and analyzes the user's position and service request, and calculates the channel gain of each user and the received signal-to-interference-noise ratio;

[0101] For users Channel gain on beam k of satellite s

[0102]

[0103] in is the downlink path loss, is the gain of the satellite transmitting antenna, η is the off-axis angle; The antenna gain is the gain received by the user, which is the same in all directions.

[0104] Downlink users In addition to the useful signal, the received signal is also affected by the interference from other co-frequency beams and system noise. Then, the user The signal-to-interference-to-noise ratio received on beam k of satellite s is

[0105]

[0106] For users In satellite Beam Total downlink interference received on Assign satellite s to users The power on beam k, Indicates that except for the users under beam k of satellite s, other satellites Beam The set of users occupying the same frequency. For users On its satellite Beam For users the interference caused; For satellite Assign to user In the beam The power on For users Relative to the interference satellite Beam N0 is the Gaussian white noise power on a single channel.

[0107] Step 4: For users After allocating downlink channel resources, the Shannon formula is used to calculate the user Channel capacity on satellite s beam k The information transmission rate is thus obtained

[0108] The channel capacity is:

[0109]

[0110] B i Indicates the bandwidth allocated to the current user;

[0111] The system information transmission rate during downlink communication is:

[0112]

[0113] Step 5: The satellite allocates power based on the dynamic power allocation algorithm of continuous convex approximation according to the information transmission rate and user requests, and optimizes the system throughput under power constraints.

[0114] The optimization objective function is:

[0115]

[0116] K represents the set of all beams of low-orbit satellite s; P tot Assign satellite s to users Maximum power in all beams;

[0117] Dynamic power algorithms for continuous convex approximations, such as Figure 5 As shown, specifically:

[0118] 1) Convert problem l2 into a special convex function difference (Difference of Convex Functions) form:

[0119] f(P)=g(P)-h(P)

[0120] Among them, g(P) and h(P) are respectively expressed as:

[0121]

[0122] Where P = {p s,k,i |1≤i≤n a} is the power set allocated to each downlink user. After being converted into convex function difference form, g(P) is a concave function, -h(P) is a convex function, and f(P) is still a non-convex function.

[0123] 2) For time slot t, the concave function h(P) is further expanded by a first-order Taylor expansion, and the objective function l2 of the s-th iteration is approximately the lower bound of the original problem of maximization:

[0124] The initial value of s is 1. Since h(P) is a concave function, its first-order Taylor expansion at P[s] has the following relationship with h(P):

[0125]

[0126] Where P[s] is the power vector allocated to the downlink subchannel at the sth convex approximation iteration, is the derivative of h(P) at P[s]. Definition Available is the lower bound of g(P)-h(P). Therefore, in the sth iteration, the optimization problem can be approximated as maximizing the lower bound of the original problem:

[0127]

[0128] 3) Using the convex optimization toolkit CVX to solve, the optimal power allocation solution P[s] at the sth iteration can be obtained;

[0129] 4) The power P[s] is used as the initial value of each downlink channel power, and the next iteration is performed until the convergence condition is met to obtain the optimal power allocation value of time slot t.

[0130] 5) Substitute the optimal power allocation value P[s] of time slot t into the information transmission rate formula in step 4 to obtain the total downlink transmission rate of the system in the current time slot t And determine the total rate Total downlink transmission rate of the system compared to the previous time slot Is the difference between them less than the threshold ε? If so, output the optimal power allocation value P[s]; otherwise, return to 2) and recalculate the lower bound of the original maximization problem for the next time slot.

[0131] Step 6: Based on the actual number of beams opened and the allocated power value, the user selects the service satellite and allocates communication resources to transmit the service.

[0132] Example:

[0133] The downlink communication process of low-orbit satellite is as follows Figure 3 As shown: the user generates a download request and sends it to the service satellite; the service satellite receives the user's location information and service request, and dynamically closes the redundant beams according to the user's location, and allocates power resources to the user according to the user's request. The user then downloads the data transmission service according to the allocated resources.

[0134] Since there is a situation in which a user in the satellite network in high-latitude areas is covered by multiple beams at the same time, it is necessary to close the redundant beams in each time slot while ensuring user satisfaction and seamless coverage, and recalculate the association between each user and the satellite beam. For example, the beam set covering user N1 is N1 = {k1, k2}, the beam set covering user N2 is N2 = {k2, k3}, and the beam set covering user N3 is N3 = {k3}. At this time, you can choose to close beam k1 or k2 to achieve full coverage of the user.

[0135] The present invention uses an ant colony algorithm based on penalty function theory to close the beam. Ants ω (ω = 1, 2, ..., m) determine their own movement direction according to the pheromones on each subset, and the transfer probability It reflects the possibility of ant ω transferring from the current subset to the next subset j at time t. The strategy for selecting the next subset is to first perform a union operation on the subset with the elements in the current solution set obtained by ant ω, and the ant gives priority to the subset with a larger number of elements in the union.

[0136] At the same time, combined with the MMAS algorithm, the penalty function theory is introduced into the pheromone update method of the ant colony algorithm, so that the pheromone update can be dynamically updated as the number of evolutions increases, so that it will not be overly concentrated or dispersed, so as to avoid premature convergence when accelerating convergence.

[0137] For the dynamic power allocation problem, the optimization goal is to maximize the system throughput under power constraints. Due to the co-channel interference between beams, the current problem is a non-convex optimization problem and a closed-form solution cannot be directly obtained. This problem is converted into a convex problem and solved it iteratively using a continuous convex approximation algorithm.

Claims

1. A method for dynamic beam management and resource adaptation of a LEO satellite, characterized in that: The specific steps are as follows: Step 1: Build a communication scenario including low-orbit constellations and users. The satellite acts as an edge node and provides services to ground terminal users through downlinks. Step 2: The satellite uses an improved ant colony-based beam closing algorithm based on the user's location, and optimizes the system with the goal of minimizing the number of open beams when the low-orbit constellation is fully covered; The improved ant colony beam closing algorithm is as follows: 1) The set of users that can be covered by each beam of each satellite is formed into a cluster; initially, m ant variables are placed on all beams of all satellites in a round-robin manner; m<n a ;n a is the total number of users in the communication scenario; 2) All m ants calculate in parallel. For ant ω in the current round, when it is placed on beam k, the beam k is added to the solution set D corresponding to the ant ω , and remove the users covered by the beam from the user set n in the scene a Delete from; 3) Ant ω based on state transition probability Select the next beam; Its value is calculated as follows: in, represents the pheromone concentration left by the beam solution of ant ω at the current moment; S j represents the beam solution set that ant ω will choose next time; a represents the relative importance of the residual information on each beam solution set in guiding the ant colony search process. The larger its value, the greater the possibility that the ant will choose this beam solution set; β represents the relative importance of heuristic information in guiding the ant colony search process; h is the set of beam solutions that ant ω has not reached in this cycle; represents the sum of the pheromone concentrations left by the ants on the set of beam solutions that have not been reached; 4) The user covered by the next beam selected by ant ω is selected from the user set n a Delete; return to 3) continue based on state transition probability Select the next beam; until all ants have set user set n in the current round cycle. a All users in the are deleted, and the current cycle ends; 5) From the current round of cycles, select the solution set with the least number of elements in the solution set of each ant as the optimal solution set D of this round; 6) Use the optimal solution set D of this round to update the pheromone concentration; The formula for calculating pheromone concentration is: Among them, ρ represents the volatility of pheromone; σ represents the penalty coefficient that changes with the number of iterations and presents a Gaussian distribution; [τ min ,τ max ] indicates the range of pheromone concentration; Indicates that ant ω stays in the beam solution set S at time t+1 j pheromone concentration on; 7) Return to 2) and restart the next cycle until the final number of iterations is reached, and retain the optimal solution set D of the last round, which is the minimum beam set that satisfies the full coverage of low-orbit satellites; Step 3: At the current moment, according to the optimized minimum number of beams, the beam is turned on and bound to the user for downlink communication. The user sends a service download request, the satellite receives and analyzes the user's position and service request, and calculates the channel gain of each user and the received signal-to-interference-noise ratio; Step 4: Target users After allocating downlink channel resources, the Shannon formula is used to calculate the user Channel capacity on satellite s beam k The information transmission rate is thus obtained Step 5: The satellite allocates power based on the dynamic power allocation algorithm of continuous convex approximation according to the information transmission rate and the user's request, and optimizes the system throughput to maximize the system throughput under power constraints; The optimization objective function is: B i Indicates the bandwidth allocated to the current user; Assign satellite s to users The power in beam k; For users Channel gain on beam k of satellite s; Indicates that except for the users under beam k of satellite s, other satellites Beam The set of users occupying the same frequency in For satellite Assign to user In the beam The power on For users Relative to the interference satellite Beam The channel gain on the channel, N0 is the Gaussian white noise power on a single channel; K represents the set of all beams of the low-orbit satellite s; P tot Assign satellite s to users Maximum power in all beams; The dynamic power algorithm of continuous convex approximation is as follows: 1) Convert the objective function l2 into a special convex function difference form: f(P)=g(P)-h(P) Among them, g(P) and h(P) are respectively expressed as: Where P = {p s,k,i |1≤i≤n a } is the power set allocated to each downlink user; 2) For time slot t, the concave function h(P) is further expanded by a first-order Taylor expansion, and the objective function l2 of the s-th iteration is approximately the lower bound of the original problem of maximization: The initial value of s is 1; The first-order Taylor expansion is: Where P[s] is the power vector allocated to the downlink subchannel at the sth convex approximation iteration, is the derivative of h(P) at P[s]; definition Available is the lower bound of g(P)-h(P); therefore, in the sth iteration, the objective function l2 is approximately the lower bound of the maximization problem: 3) Use the convex optimization toolkit CVX to solve the lower bound of the original maximization problem and obtain the optimal power P[s] at the sth iteration; 4) Taking the power P[s] as the initial value of each downlink channel power, the next iteration is performed until the convergence condition is met, and the optimal power allocation value of time slot t is obtained; 5) Substitute the optimal power allocation value P[s] of time slot t into the information transmission rate formula to obtain the total downlink transmission rate of the system in the current time slot t And determine the total rate Total downlink transmission rate of the system compared to the previous time slot Is the difference between them less than the threshold ε? If so, output the optimal power allocation value P[s]; otherwise, return to 2) and recalculate the lower bound of the original maximization problem for the next time slot; Step 6: Based on the actual number of beams opened and the allocated power value, the user selects the service satellite and allocates communication resources to transmit the service.

2. A LEO satellite dynamic beam management and resource adaptation method as claimed in claim 1, characterized in that: In step 1, the number of satellites included in the low-orbit satellite constellation is s a ; The number of beams emitted by each satellite is k a ; Each beam center in the low-orbit constellation coverage area is distributed with a terminal user; the total number of users n a =s a *k a .

3. A LEO satellite dynamic beam management and resource adaptation method as claimed in claim 1, characterized in that: In the step 2, the optimization objective function is: Where S represents the set of all low-orbit satellites; K represents the set of all beams of low-orbit satellite s; x s,k ∈{0,1} represents the two-dimensional state representation of beam k of satellite s, x s,k =0 means beam is off, x s,k =1 means beam is on; β s,k,i ∈{0,1} represents the two-dimensional state representation of whether user i is covered by beam k of satellite s, β s,k,i =1 means user i is covered by beam k under satellite s, β s,k,i =0 means that user i is not covered by beam k under satellite s.

4. A LEO satellite dynamic beam management and resource adaptation method as claimed in claim 1, characterized in that: In step 2, the selection rule of the state transition probability is: the beam solution set D of ant ω is ω The users covered by all beams in the list and the users covered by the selected beam are combined to obtain the beam with the largest number of union elements, which is the next beam to be selected.

5. A LEO satellite dynamic beam management and resource adaptation method as claimed in claim 1, characterized in that: In step 3, for the user Channel gain on beam k of satellite s in is the downlink path loss, is the gain of the satellite transmitting antenna, η is the off-axis angle; Antenna gain for user reception; Then, calculate the user The signal-to-interference-noise ratio received on beam k of satellite s is For users On its satellite Beam For users The interference generated.

6. A LEO satellite dynamic beam management and resource adaptation method as claimed in claim 1, characterized in that: The channel capacity of step 4 is: The information transmission rate is:

Images