A communication and remote sensing integrated satellite constellation clustering method based on service prediction

Abstract

The application is suitable for the technical field of satellite network communication, and provides a communication and remote sensing integrated satellite constellation clustering method based on service prediction, which comprises the following steps: constructing a communication and remote sensing integrated satellite model and an integrated satellite constellation model composed of integrated satellites; predicting the service traffic in future time slots by using the historical service data in the service area of each ground station, and determining the function selection of the integrated satellites in future time slots; constructing an integrated satellite constellation clustering model based on Stackelberg game with the target of maximizing the system utility; introducing a random best response strategy, solving the Stackelberg game model, and completing the clustering of the communication and remote sensing integrated satellite constellation. The application can effectively predict the heterogeneous service traffic in the region, pre-divide the functions of the integrated satellites, and establish a clustering mode with high stability and low routing distance, so as to reduce the management complexity of the large constellation and improve the overall performance of the satellite communication system.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of satellite network communication, and in particular to a communication and remote sensing integrated satellite constellation clustering method based on service prediction. BACKGROUND

[0002] Low-orbit remote sensing satellites provide effective support for various earth observation services. With the continuous development of the economy and society, people's demand for timeliness and resolution of remote sensing services is increasing. In recent years, low-orbit communication satellites have also attracted much attention, and domestic and foreign organizations have successively launched broadband satellites to build space-based Internet to assist ground communication. However, the current operation mode of remote sensing satellites results in short satellite coverage service time, and the satellite-ground resource transmission is not timely, which is difficult to meet the demand for remote sensing data acquisition with large data volume and high timeliness. In order to realize low-latency data downlink without sacrificing the service quality of communication services, and to establish a large-scale constellation group while saving limited payload resources, orbital resources and frequency resources, communication and remote sensing integrated satellites have become the trend. The communication and remote sensing integrated satellite takes a general satellite platform as the main body, integrates communication payloads and remote sensing payloads, and can not only complete high-coverage, high-service-time, and high-timeliness earth observation tasks, but also can perform inter-satellite communication and satellite-ground communication, and establish an inter-satellite Internet. In a large-scale constellation, in order to realize efficient backhaul, inter-satellite links are often used to transmit data back to the ground satellite and back to the ground. In the context of a large-scale constellation, the overhead of routing calculation and constellation management is huge, and a large number of opportunistic routing requires a large amount of space to store routing tables, which is limited by on-board resources and energy, so the research on satellite clustering is very necessary. Satellite clustering is to divide the satellites in the constellation into several mutually disjoint clusters. In each cluster, a node is selected as the cluster head, and the other nodes are cluster members. The cluster head is responsible for intra-cluster management, i.e., routing scheduling and resource allocation, which can effectively reduce the management overhead and reduce the routing distance. Different choices of cluster members and cluster heads will result in different network management overhead and backhaul effects. Therefore, a communication and remote sensing integrated satellite constellation clustering method based on service prediction is proposed. SUMMARY

[0003] The present application aims to provide a communication and remote sensing integrated satellite constellation clustering method based on service prediction, which aims to solve the problems in the background technology.

[0004] A communication and remote sensing integrated satellite constellation clustering method based on service prediction, the method comprising:

[0005] Step S1, constructing a communication and remote sensing integrated satellite model comprising a universalized satellite platform, a communication payload and a remote sensing payload, and a communication and remote sensing integrated satellite constellation model comprising the communication and remote sensing integrated satellites;

[0006] Step S2, predicting future time slot traffic data using historical traffic data in each ground station service area, and determining the function selection of the integrated satellite in each future time slot based on the prediction of the traffic data;

[0007] Step S3, constructing a communication and remote sensing integrated satellite constellation clustering model based on a Stackelberg game with the goal of maximizing system utility;

[0008] Step S4, introducing a random best response strategy to solve the Stackelberg game model and complete satellite clustering of the communication and remote sensing integrated satellite constellation.

[0009] Preferably, in step S1, the communication and remote sensing integrated satellite model and the communication and remote sensing integrated satellite constellation model are constructed as follows:

[0010] First, a universalized satellite platform is constructed. Based on mainstream satellite platforms, the electrical, software, network and other interface forms are changed, and open and standardized architectures are adopted at the interfaces, rather than designing a dedicated platform for communication or remote sensing functions, so that the communication payload and the remote sensing payload can be plug-and-play on the universalized satellite platform. When new payloads need to be carried or payloads need to be replaced due to damage, the platform does not need to be redesigned. The effective payload carried on the universalized satellite platform is designed to complete the design of the remote sensing payload mainly composed of high-resolution optical cameras and the communication payload mainly composed of phased array antennas under the conditions of low cost and replaceability, thereby realizing the communication function and the remote sensing function. Laser transmitters and receivers are carried on the satellite to enable each satellite to establish inter-satellite links. Then, based on the Walker constellation orbit, a communication and remote sensing integrated satellite constellation is constructed with the communication and remote sensing integrated satellites as the constituent elements.

[0011] Preferably, the content of step S2 includes:

[0012] Step S21, collecting historical traffic data in each ground station service area, including remote sensing traffic demand and communication traffic demand, and performing data preprocessing;

[0013] Step S22, predicting future traffic data using an LSTM neural network prediction algorithm;

[0014] Step S23, evaluating the prediction results using two indicators, mean absolute error and root mean square error;

[0015] Step S24: Based on the proportion of business demand in this area, set the satellite functions within and outside the visible range of the ground station.

[0016] Preferably, step S24 includes the following:

[0017] Based on the LSTM neural network prediction algorithm, the communication service demand and remote sensing service demand of each ground station have been predicted. The average time that each satellite spends within the service area of ​​the corresponding ground station is taken as a time slot. For each ground station, the predicted communication service demand and the predicted remote sensing service demand are compared, and the service with the larger predicted value is selected as the mainstream service in this time slot. Satellites within the line of sight of the ground station are set as the corresponding mainstream service functions. For satellites not within the line of sight of any ground station, they are uniformly set to communication mode and remote sensing mode.

[0018] Preferably, step S3 includes the following:

[0019] Step S31: Using the virtual spatiotemporal projection method, project domestic remote sensing ground stations onto the globe;

[0020] Step S32: Establish a cluster model for an integrated communication and remote sensing satellite constellation;

[0021] Step S33: Perform deterministic modeling and strategy selection modeling for leaders and followers;

[0022] Step S34: Design the reward function for the Stackelberg game model.

[0023] Preferably, the principle of the virtual spatiotemporal projection method in step S31 is as follows:

[0024] Based on the orbital characteristics of the Walker constellation, the satellite whose nadir point is closest to each ground station is found, and the virtual image of the ground station is translated and copied along the orbit of the corresponding satellite to other geographical locations. The number of projections can be reasonably set according to the constellation management requirements; the ground station includes the projected virtual image of the ground station.

[0025] Preferably, the content of the integrated communication and remote sensing satellite constellation cluster model in step S32 is as follows:

[0026] In an integrated satellite constellation, each integrated satellite is divided into several non-overlapping clusters. Within each cluster, a node is selected as the cluster head, and the other nodes become cluster members. The cluster head is responsible for cluster management, namely routing scheduling and resource allocation. Indicates the assembly at the ground station. ,in For the first A ground station, a cluster of satellites consisting of express, wherein is the i-th cluster head satellite, the cluster member satellite set is represented by , wherein, , wherein is the i-th cluster head satellite, cluster managed by the i-th cluster head satellite, is the j-th cluster member satellite within the cluster, represents the maximum number of cluster members within the cluster, is the j-th cluster member satellite within the cluster; the ground station corresponds to the cluster head one by one, the ground station has a condition of backhaul to the ground with the cluster head as the main, i.e. transmission between satellites within the visual range of the ground station, the cluster head selects cluster member satellites among other satellites according to the corresponding condition, and the cluster member satellites select the satellite backhaul to the ground station. Preferably, the contents of the deterministic modeling and policy selection modeling in the step S33 include:

[0027] The ground station is the leader and the cluster head satellite is the follower; within the visual range of the ground station, the ground station selects a satellite as the cluster head, after determining the cluster head, the cluster head searches the nearby satellites as cluster members in ascending order of geometric distance to work together to complete the backhaul of on-board data.

[0028] Preferably, the design method of the reward function in the step S34 is as follows:

[0029] The star-ground backhaul effect after the cluster head selection is given priority, and the influence of inter-satellite stability and intra-cluster routing distance on the clustering utility is comprehensively considered, a system utility function is defined, and the system utility function is taken as the reward function of the Stackelberg game, and the expression is:

[0030] ;

[0031] ;

[0032] In the formula, represents the system utility, represents the transmission rate between the ground station and the cluster head satellite, represents the transmission rate between the ground station and the cluster member satellite, wherein the transmission rate between the ground station and the non-visible cluster member satellite is zero, is the inter-satellite stability factor, is the intra-cluster distance factor, is the sum of the remote sensing business demand and the communication business demand of the ground station;

[0033] The transmission rate between the ground station and the cluster head satellite is expressed as:

[0034] ; ​​

[0035] wherein, denotes the bandwidth of the link between the ground station and the cluster head satellite, denotes the signal-to-noise ratio of the signal transmitted in this link;

[0036] The transmission rate between the ground station and the cluster member satellites is:

[0037] ;

[0038] wherein, denotes the bandwidth of the link between the ground station and the cluster member satellites, denotes the signal-to-noise ratio of the signal transmitted in this link;

[0039] The expression of the inter-satellite stability factor is:

[0040] ;

[0041] wherein, is the received power between the cluster head and each cluster member, determined by the distribution of the deviation angle between the cluster head and the cluster member, is the maximum received power of the receiver for establishing the inter-satellite link;

[0042] The expression of the intra-cluster distance factor is:

[0043] , ;

[0044] wherein, is the routing distance between the cluster member and the satellite for backhauling, obtained by the shortest path algorithm, denotes the satellite for backhauling.

[0045] Preferably, the content of the step S4 comprises:

[0046] The random best response strategy is introduced, and the strategy update probability is first calculated, and the expression is:

[0047] ;

[0048] wherein, is the strategy update probability, is the system utility of the last time slot, is the normalization operation, indicating that the clustering reward function of the last time slot will act on the strategy update probability of the current time slot;

[0049] Under the premise of known star-ground visibility and obtained ground station task demand through business prediction, if there is visibility between the ground station and several satellites with the ability to face the earth in a region, the ground station selects the satellite with the maximum estimated transmission rate as the cluster head; after the preliminary decision of several satellites as cluster heads, the cluster head satellite executes the cluster member selection strategy, and obtains the inter-satellite stability factor and the routing distance factor, and calculates the clustering reward function; in the process of strategy updating, only the cluster head selection of one ground station is changed in each cycle, the satellite with suboptimal channel state is selected as the cluster head, the cluster member selection strategy is executed, the clustering reward function after the change is estimated, if the new function value is greater than the previous optimal, a random number between 0 and 1 is generated and compared with the strategy updating probability, if the random number is less than the strategy updating probability, the cluster head selection strategy is updated, and after the update, each cluster head re-executes the cluster member selection strategy; if the new function value is less than or equal to the previous optimal or the random number is greater than or equal to the strategy updating probability, the cluster head selection strategy is not updated.

[0050] The specific content of the cluster member selection strategy is:

[0051] After determining the cluster head, the cluster members are selected in the descending order of the received power after the cluster head and other satellites establish a link, until the received power between the cluster head and each cluster member is lower than the threshold value; in each cluster, the satellite covering the ground station by the beam has the backhaul capability, and the minimum routing distance between each cluster member satellite and the satellite facing the earth in the cluster is calculated.

[0052] The beneficial effects realized by the application are:

[0053] The application can effectively predict the heterogeneous business traffic in the region, pre-divide the functions of the integrated satellite, and establish a clustering mode with high stability and low routing distance, thereby reducing the management complexity of large constellation and improving the overall performance of the satellite communication system. BRIEF DESCRIPTION OF DRAWINGS

[0054] Fig. 1 It is a flow chart of a communication remote sensing integrated satellite constellation clustering method based on business prediction.

[0055] Fig. 2 It is a schematic diagram of a communication remote sensing integrated satellite constellation clustering model in the application.

[0056] Fig. 3 It is a single cluster model schematic diagram of a communication remote sensing integrated satellite constellation in the application.

[0057] Fig. 4 It is a flow chart of the solution method of the Stackelberg game in the application. DETAILED DESCRIPTION

[0058] With reference to the drawings of the embodiments of the present application, the technical solutions in the embodiments of the present application will be described clearly and completely. Obviously, the described embodiments are only a part of the embodiments of the present application, rather than all the embodiments of the present application. Based on the embodiments of the present application, all other embodiments obtained by those of ordinary skill in the art without creative effort belong to the scope of the present application.

[0059] Embodiment 1, please refer to Figs. 1 to 4 The embodiment of the present application provides a communication and remote sensing integrated satellite constellation clustering method based on service prediction, and the method comprises the following steps:

[0060] Step S1, constructing a communication and remote sensing integrated satellite model comprising a universalized satellite platform, a communication load and a remote sensing load, and a communication and remote sensing integrated satellite constellation model composed of the communication and remote sensing integrated satellites;

[0061] In step S1, the communication and remote sensing integrated satellite model and the communication and remote sensing integrated satellite constellation model are constructed as follows:

[0062] Firstly, the universalized satellite platform is constructed. On the basis of mainstream satellite platforms, the electrical, software, network and other interface forms are changed, and open and standardized architectures are adopted at the interfaces, instead of designing special platforms for communication or remote sensing functions, so that the communication load and the remote sensing load can realize plug and play on the universalized satellite platform. When new loads need to be carried or loads need to be replaced due to damage, the platform does not need to be redesigned. The effective loads carried on the universalized satellite platform are designed under the conditions of low cost and replaceability, so as to complete the design of the remote sensing load mainly composed of high-resolution optical cameras and the communication load mainly composed of phased array antennas, thereby realizing the communication function and the remote sensing function. Laser transmitters and receivers are carried on the satellites, so that each satellite has the ability to establish inter-satellite links. Then, based on the Walker constellation orbit, the communication and remote sensing integrated satellite constellation is constructed by taking the communication and remote sensing integrated satellites as constituent elements.

[0063] Step S2, predicting the service data in future time slots by using the historical service data in the service area of each ground station, and determining the function selection of the integrated satellites in future time slots based on the prediction of the service data;

[0064] The content of step S2 comprises:

[0065] Step S21, collecting the historical service data in the service area of each ground station, including the remote sensing service demand and the communication service demand, and performing data preprocessing;

[0066] Step S22, predicting the future service data by using an LSTM neural network prediction algorithm;

[0067] Step S23, the prediction results are evaluated by using two indicators of mean absolute error and root mean square error;

[0068] Step S24, according to the proportion of service demand in the region, the satellite function within the visible range of the ground station and outside the visible range is set.

[0069] The content of step S24 is as follows:

[0070] After the service prediction based on the LSTM neural network prediction algorithm, the communication service demand and the remote sensing service demand of each ground station have been predicted; the time average of each satellite in the corresponding ground station service area is taken as a time slot; for each ground station, the predicted communication service demand and the predicted remote sensing service demand are compared, and the service with larger predicted value is selected as the mainstream service in this time slot, and the satellite within the visible range of the ground station is set as the corresponding mainstream service function, and the satellites not within the visible range of any ground station are uniformly set as communication mode and remote sensing mode.

[0071] Step S3, a communication and remote sensing integrated satellite constellation clustering model based on Stackelberg game is constructed to maximize the system utility;

[0072] The content of step S3 is as follows:

[0073] Step S31, the domestic remote sensing ground station is projected to the global by using virtual space-time projection method;

[0074] Step S32, a communication and remote sensing integrated satellite constellation clustering model is established;

[0075] Step S33, the leader and the follower are determined and the strategy selection modeling is performed;

[0076] Step S34, the reward function of the Stackelberg game model is designed.

[0077] The principle of the virtual space-time projection method in step S31 is as follows:

[0078] Based on the orbit characteristics of Walker constellation, the satellite closest to the subsatellite point and each ground station is found, the virtual image of the ground station is translated and copied to other geographical positions along the orbit of the corresponding satellite, and the projection times can be reasonably set according to the constellation management requirements; the ground station includes a projected ground station virtual image.

[0079] The content of the communication and remote sensing integrated satellite constellation clustering model in step S32 is as follows:

[0080] In an integrated satellite constellation, each integrated satellite is divided into several non-overlapping clusters. Within each cluster, a node is selected as the cluster head, and the other nodes become cluster members. The cluster head is responsible for cluster management, namely routing scheduling and resource allocation. Indicates the assembly at the ground station. ,in For the first A ground station, a cluster of satellites consisting of express, ,in For the first A cluster head satellite, and a collection of cluster member satellites consisting of... It means that, among them, , ,in For the first Clusters managed by a single satellite, For the first in the cluster Cluster member satellites, Indicates the maximum number of cluster members. For the first in the cluster Each cluster consists of a member satellite; a ground station corresponds one-to-one with the cluster head. The ground station and the satellites with the cluster head as the main satellites have the capability to transmit back to Earth, that is, they transmit between satellites within the ground station's line of sight. The cluster head selects member satellites from other satellites according to the corresponding conditions, and the member satellites select Earth-reaching satellites to transmit back to the ground station.

[0081] The deterministic modeling and strategy selection modeling in step S33 include:

[0082] With the ground station as the leader and the cluster head satellite as the follower, within the ground station's line of sight, the ground station selects a satellite as the cluster head. After the cluster head is determined, the cluster head searches for nearby satellites in ascending order of geometric distance to serve as cluster members in order to work together and complete the return of on-board data.

[0083] The design method for the reward function in step S34 is as follows:

[0084] Prioritizing the satellite-to-ground backhaul performance after cluster head selection, and comprehensively considering the impact of inter-satellite stability and intra-cluster routing distance on clustering utility, a system utility function is defined and used as the payoff function in the Stackelberg game. The expression is as follows:

[0085] ;

[0086] In the formula, Indicates system utility. This indicates the transmission rate between the ground station and the cluster satellite. This represents the transmission rate between the ground station and member satellites, with the transmission rate between the station and non-visible member satellites being zero. is an inter-satellite stability factor, is an intra-cluster distance factor, is the sum of remote sensing service demand and communication service demand of the ground station;

[0087] wherein the transmission rate expression between the ground station and the cluster head satellite is:

[0088] ;

[0089] wherein, denotes the bandwidth of the link between the ground station and the cluster head satellite, denotes the signal-to-noise ratio of the signal transmitted in this link;

[0090] the transmission rate between the ground station and the cluster member satellite is:

[0091] ;

[0092] wherein, denotes the bandwidth of the link between the ground station and the cluster member satellite, denotes the signal-to-noise ratio of the signal transmitted in this link;

[0093] the expression of the inter-satellite stability factor is:

[0094] ;

[0095] wherein, is the received power between the cluster head and each cluster member, determined by the distribution of the deviation angle between the cluster head and the cluster member, is the maximum received power of the receiver for establishing the inter-satellite link;

[0096] the expression of the intra-cluster distance factor is:

[0097] , ;

[0098] wherein, is the routing distance between the cluster member and the satellite for backhaul, obtained by the shortest path algorithm, denotes the satellite for backhaul.

[0099] Step S4, introducing a random best response strategy, solving the Stackelberg game model, and completing the satellite clustering of the communication and remote sensing integrated satellite constellation.

[0100] The content of the step S4 includes:

[0101] Introducing a random best response strategy, first calculating the strategy update probability, the expression is:

[0102] ;

[0103] wherein, is a policy update probability, is a system utility of the previous time slot, is a normalization operation, indicating that the clustering reward function of the previous time slot acts on the policy update probability of the current time slot;

[0104] Under the premise of knowing the star-ground visibility and obtaining the ground station task demand through business prediction, if there is visibility between the ground station and some satellites with the ability to communicate with the ground in the region, the ground station selects the satellite with the maximum estimated transmission rate as the cluster head; after preliminarily deciding on some satellites as cluster heads, the cluster head satellite executes the cluster member selection strategy and obtains the inter-satellite stability factor and the routing distance factor, and calculates the clustering reward function; in the process of policy updating, only the cluster head selection of one ground station is changed in each loop, the satellite with the suboptimal channel state is selected as the cluster head, the cluster member selection strategy is executed, and the clustering reward function after the change is estimated; if the new function value is greater than the previous optimal value, a random number between 0 and 1 is generated and compared with the policy update probability; if the random number is less than the policy update probability, the cluster head selection strategy is updated, and after the update, each cluster head re-executes the cluster member selection strategy; if the new function value is less than or equal to the previous optimal value or the random number is greater than or equal to the policy update probability, the cluster head selection strategy is not updated.

[0105] wherein, the specific content of the cluster member selection strategy is:

[0106] After determining the cluster head, the cluster members are selected in descending order of the received power after the cluster head establishes a link with other satellites, until the received power between the cluster head and each cluster member is lower than a threshold; within each cluster, the satellite covering the ground station by the beam has the backhaul capability, and the minimum routing distance between each cluster member satellite and the satellite in the cluster that communicates with the ground is calculated.

[0107] Embodiment 2 provides a satellite clustering method based on user demand, which can predict the demand of users in the service area of different ground stations on the earth for remote sensing business communication services, such as time resources required by remote sensing business, scheduling resources, data volume required by voice and video, data transmission and the like. According to the user demand, the integrated satellite function is determined, and according to the satellite characteristics in each function state, the satellite clustering is performed to reduce the network overhead and adapt to heterogeneous business. The specific implementation process can refer to the implementation process of the method steps in Embodiment 1, which will not be repeated here.

[0108] Embodiment 3, the embodiment provides a communication and remote sensing resource coordination and satellite clustering method based on emergency demand, which can be based on business prediction to perceive emergency communication and emergency remote sensing demand in the service range of each ground station, such as natural disasters like tsunami, earthquake and other malignant events, and needs to urgently switch integrated satellite functions to serve emergency business when the events occur, to ensure that the emergency business is completed in priority. On this basis, according to the characteristics of the integrated satellite in each function state, satellite clustering is performed to realize fast and low-cost management under emergency business. The specific implementation process can refer to the implementation process of the method steps in embodiment 1, which will not be described here.

[0109] The above is only the preferred embodiment of the present application, and does not limit the patent scope of the present application, and any equivalent structure or equivalent process transformation made by using the specification of the present application, or directly or indirectly applied in other related technical fields, are also included in the patent protection scope of the present application.

Claims

1. A satellite constellation clustering method based on service forecasting for integrated communication and remote sensing, characterized in that, The method includes: Step S1: Construct an integrated communication and remote sensing satellite model that includes a generalized satellite platform, communication payload, and remote sensing payload, as well as an integrated communication and remote sensing satellite constellation model composed of the integrated communication and remote sensing satellites; Step S2: Use historical service data within the service areas of various ground stations to predict service data in several future time slots. Based on the prediction of service data, determine the function selection of the integrated satellite in each future time slot. Step S3: With the goal of maximizing system utility, construct a clustering model for an integrated communication and remote sensing satellite constellation based on Stackelberg game theory; Step S4: Introduce a stochastic optimal response strategy, solve the Stackelberg game model, and complete the satellite clustering of the integrated communication and remote sensing satellite constellation.

2. The method for integrated communication and remote sensing satellite constellation clustering based on service forecasting according to claim 1, characterized in that, In step S1, the construction methods for the integrated communication and remote sensing satellite model and the integrated communication and remote sensing satellite constellation model are as follows: First, a general-purpose satellite platform is constructed, changing the electrical, software, and network interface forms, and adopting an open and standardized architecture at the interface, instead of designing a dedicated platform for communication or remote sensing functions; the payload to be carried on the general-purpose satellite platform is designed; next, based on the Walker constellation orbit, an integrated communication and remote sensing satellite constellation is constructed with integrated communication and remote sensing satellites as components.

3. The method for integrated communication and remote sensing satellite constellation clustering based on service forecasting according to claim 1, characterized in that, The contents of step S2 include: Step S21: Collect historical business data within the service area of ​​each ground station, including remote sensing business demand and communication business demand, and perform data preprocessing; Step S22: Use the LSTM neural network prediction algorithm to predict future business data; Step S23: Evaluate the prediction results using two indicators: mean absolute error and root mean square error. Step S24: Based on the proportion of business demand in this area, set the satellite functions within and outside the line of sight of the ground station.

4. The method for integrated communication and remote sensing satellite constellation clustering based on service forecasting according to claim 3, characterized in that, The content of step S24 is as follows: Based on the LSTM neural network prediction algorithm, the communication service demand and remote sensing service demand of each ground station have been predicted. The average time that each satellite spends within the service area of ​​the corresponding ground station is taken as a time slot. For each ground station, the predicted communication service demand and the predicted remote sensing service demand are compared, and the service with the larger predicted value is selected as the mainstream service in this time slot. Satellites within the line of sight of the ground station are set as the corresponding mainstream service functions. For satellites not within the line of sight of any ground station, they are uniformly set to communication mode and remote sensing mode.

5. The method for integrated communication and remote sensing satellite constellation clustering based on service forecasting according to claim 1, characterized in that, The content of step S3 is as follows: Step S31: Using the virtual spatiotemporal projection method, project domestic remote sensing ground stations onto the globe; Step S32: Establish a cluster model for an integrated communication and remote sensing satellite constellation; Step S33: Using the ground station as the leader and the cluster satellite as the follower, perform deterministic modeling and strategy selection modeling for the leader and the follower. Step S34: Design the reward function for the Stackelberg game model.

6. The method for integrated communication and remote sensing satellite constellation clustering based on service forecasting according to claim 5, characterized in that, The principle of the virtual spatiotemporal projection method in step S31 is as follows: Based on the orbital characteristics of the Walker constellation, the satellite whose nadir point is closest to each ground station is found, and the virtual image of the ground station is translated and copied along the orbit of the corresponding satellite to other geographical locations. The number of projections is set according to the constellation management requirements; the ground station includes the projected virtual image of the ground station.

7. A satellite constellation clustering method based on service forecasting for integrated communication and remote sensing, as described in claim 5, is characterized in that... The content of the integrated communication and remote sensing satellite constellation clustering model in step S32 is as follows: In an integrated satellite constellation, each integrated satellite is divided into several non-overlapping clusters. Within each cluster, a node is selected as the cluster head, and the other nodes become cluster members. The cluster head is responsible for cluster management, namely routing scheduling and resource allocation. Indicates the assembly at the ground station. ,in For the first A ground station, a cluster of satellites consisting of express, ,in For the first A cluster head satellite, and a collection of cluster member satellites consisting of... It means that, among them, , ,in For the first Clusters managed by a single satellite, For the first in the cluster Cluster member satellites, Indicates the maximum number of cluster members. For the first in the cluster Each cluster consists of a member satellite; a ground station corresponds one-to-one with the cluster head. The ground station and the satellites with the cluster head as the main satellites have the capability to transmit back to Earth, that is, they transmit between satellites within the ground station's line of sight. The cluster head selects member satellites from other satellites according to the corresponding conditions, and the member satellites select Earth-reaching satellites to transmit back to the ground station.

8. A satellite constellation clustering method based on service forecasting for integrated communication and remote sensing, as described in claim 5, is characterized in that... The deterministic modeling and strategy selection modeling in step S33 include: With the ground station as the leader and the cluster head satellite as the follower, within the ground station's line of sight, the ground station selects a satellite as the cluster head. After the cluster head is determined, the cluster head searches for nearby satellites in ascending order of geometric distance to serve as cluster members in order to work together and complete the return of on-board data.

9. A satellite constellation clustering method based on service forecasting for integrated communication and remote sensing, as described in claim 5, is characterized in that... The design method for the reward function in step S34 is as follows: Define the system utility function and use it as the payoff function in the Stackelberg game. The expression is as follows: ； In the formula, Indicates system utility. This indicates the transmission rate between the ground station and the cluster satellite. This represents the transmission rate between the ground station and member satellites, with the transmission rate between the station and non-visible member satellites being zero. As an interstellar stability factor, It is the intra-cluster distance factor. This is the sum of the remote sensing service demand and communication service demand of the ground station. The transmission rate expression between the ground station and the cluster satellite is as follows: ； In the formula, This indicates the bandwidth of the link between the ground station and the cluster satellite. This indicates the signal-to-noise ratio of the signal transmitted in this link; The transmission rate between the ground station and the cluster satellites is: ； In the formula, This indicates the bandwidth of the link between the ground station and the cluster satellites. This indicates the signal-to-noise ratio of the signal transmitted in this link; The expression for the interstellar stability factor is: ； in, The received power between the cluster head and each cluster member is determined by the angular distribution of the deviation between the cluster head and cluster members. The maximum receiving power of the receiver for establishing an inter-satellite link; The expression for the intra-cluster distance factor is: ， ； in, The route distance between cluster members and Earth-based satellites is obtained using a shortest path algorithm. This refers to the satellite responsible for transmitting data back to its origin.

10. A satellite constellation clustering method based on service forecasting for integrated communication and remote sensing, as described in claim 1, is characterized in that... The content of step S4 includes: Introducing a stochastic optimal response policy, we first calculate the policy update probability, expressed as: ； in, Update the probability for the strategy. For the system utility of the previous time slot, For normalization operations, it means that the clustering reward function of the previous time slot will affect the policy update probability of the current time slot; Given known satellite-to-ground visibility and ground station mission requirements obtained through service forecasting, if there is visibility between the ground station and several satellites with ground-to-ground capabilities within the area, the ground station selects the satellite with the highest estimated transmission rate as the cluster head. After initially deciding on several satellites as cluster heads, the cluster head satellite executes the cluster member selection strategy, obtains the inter-satellite stability factor and routing distance factor, and calculates the clustering reward function. During the strategy update process, each loop only changes the cluster head selection of one ground station, selecting the satellite with the second-best channel state as the cluster head, executing the cluster member selection strategy, and estimating the changed clustering reward function. If the new function value is greater than the previous best, a random number between 0 and 1 is generated and compared with the strategy update probability. If the random number is less than the strategy update probability, the cluster head selection strategy is updated, and each cluster head re-executes the cluster member selection strategy. If the new function value is less than or equal to the previous best or the random number is greater than or equal to the strategy update probability, the cluster head selection strategy is not updated. The cluster member selection strategy includes the following: After determining the cluster head, cluster members are selected in descending order of received power after the cluster head establishes links with other satellites, until the received power between the cluster head and each cluster member is lower than the threshold. Within each cluster, satellites covering the ground station have backhaul capability, and the minimum routing distance between each cluster member satellite and the ground satellite within the cluster is calculated.

Images