Inter-satellite and ground-based joint topology planning method for satellite networks oriented to integrated data interaction

Abstract

A satellite network inter-satellite and satellite-ground joint topology planning method for integrated data interaction is disclosed. The satellite network, i.e., the satellite system, comprises three network elements: domestic satellites, overseas satellites, and ground stations. In a satellite network integrated with communication, navigation, and remote sensing, the satellite system adopts narrow-beam directional inter-satellite links. In a satellite network scenario integrated with communication, navigation, and remote sensing, both inter-satellite and satellite-ground links adopt a time-division polling system. Under the premise of meeting system measurement requirements, the delay of satellite-ground data interaction is optimized. The target problem of the inter-satellite and satellite-ground integrated topology planning is modeled as a traffic-agnostic integer linear programming problem. The steps are as follows: 1) deploying ground stations supporting narrow-beam satellite-ground links so that the satellite-ground links adopt the same time-division polling system as the inter-satellite links; 2) performing traffic-agnostic delay modeling. The satellite system can use inter-satellite links for high-precision orbit determination and space-based end-to-end telemetry, remote control, and communication.

Description

Technical Field

[0001] The present invention relates to the technical field of satellite network topology planning, and in particular to a satellite network inter-satellite and satellite-ground joint topology planning method oriented to integrated data interaction. Background Art

[0002] Satellite network topology planning is a crucial issue in satellite system design, as the quality of the planned topology directly impacts the system's measurement and communication performance. Limited by the number of narrow-beam antenna terminals onboard, the number of intersatellite links a satellite can establish simultaneously is far fewer than the number of satellites within its visible range. To meet the measurement and communication performance requirements for normal satellite system operation, intersatellite links can utilize a time-division polling system, slotted communication, and a store-and-carry-forward approach for intersatellite data transmission. However, traditional data transmission between ground station nodes and satellites utilizes wide-beam satellite-to-ground links with a non-time-division fixed link establishment system. In this system, the ground station node and satellite maintain a link within the visible arc. The non-time-division fixed link establishment system of traditional satellite-to-ground links differs significantly from the time-division polling system of intersatellite links, resulting in inefficient data exchange. Integrated communication, navigation, and remote sensing is a future trend in space satellite networks. This integration refers to the integration of high-resolution multi-mode remote sensing, two-way IoT communications, and satellite-based navigation enhancements within a single satellite. Currently, satellites operate independently in these three areas, lacking a holistic approach, resulting in low efficiency in information acquisition and utilization. Among them, integrated communication, navigation, and telemetry is an important component and key direction for the current development of communication, navigation, and telemetry systems. In integrated communication, navigation, and telemetry systems, navigation and telemetry messages are transmitted back to ground stations via communication links. However, since ground stations in my country can generally only be deployed within the country, if satellites in the system are divided into domestic and overseas satellites based on whether they can connect to any domestic ground station, overseas satellites can exchange information with ground stations through domestic satellites via inter-satellite links. Therefore, if inter-satellite and satellite-to-ground link systems can be unified, and the network topology and link establishment between nodes can be comprehensively planned, the efficiency of data exchange across the entire network will be greatly improved.

[0003] If the traditional satellite-to-ground link system actively adjusts its data distribution method to adapt to the inter-satellite time-division polling system, the ground station needs to divide the data into time slots and interact with each link-building satellite node in a specific time slot. This method requires a cumbersome technical upgrade of the ground data system. If the current data distribution method remains unchanged, the injected data of all time slots cannot be sent to the optimal access node in each time slot. The data will be stranded at the access node and thus affect the transmission performance of the inter-satellite network. In the satellite network, the inter-satellite link and the satellite-to-ground link both adopt the time-division polling system for joint topology planning, which can effectively improve the efficiency of data interaction. Among them, the traditional delay optimal modeling method is to use the idea of ​​network flow optimization to plan the optimal path for each data flow, and model the inter-satellite-to-ground joint topology planning problem as a traffic-aware integer linear programming model (Traffic-Aware ILP, TAILP). However, this model has a very large number of variables and is difficult to apply to the integrated communication, navigation and remote sensing satellite network scenario. Summary of the Invention

[0004] The present invention aims to provide a satellite-to-ground joint topology planning method for integrated data interaction in a satellite network integrating communication, navigation, and remote sensing. The topology planning scheme obtained by the method can overcome the low efficiency of satellite-to-ground data interaction under the traditional fixed-link satellite-to-ground link system and solve the problem that the integer linear programming model of traffic perception has an extremely large number of variables. The designed algorithm has low computational complexity and can approach the delay performance of the delay optimal algorithm.

[0005] The technical solution of the present invention is: a satellite network inter-satellite and satellite-ground joint topology planning method for integrated data interaction. In the integrated communication, navigation, and remote sensing satellite network scenario, both inter-satellite and satellite-ground links adopt a time-division polling system. Under the premise of meeting system measurement requirements, the delay of satellite-ground data interaction is optimized. The inter-satellite and satellite-ground integrated topology planning target problem is modeled as a traffic-insensitive integer linear programming problem. The designed topology planning algorithm has low computational complexity and is feasible and suboptimal in the integrated communication, navigation, and remote sensing satellite network scenario. The steps of the inter-satellite and satellite-ground joint topology planning method are as follows:

[0006] Step 1: Deploy ground stations supporting narrow-beam satellite-to-ground links, enabling them to adopt the same time-division polling system as inter-satellite links, paving the way for integrated planning of inter-satellite and satellite-to-ground links. In each topology planning state under the time-division polling system, calculate the visibility matrix based on the ephemeris, including inter-satellite and satellite-to-ground visibility conditions, as input for topology planning. Basic and measurement constraints are then established.

[0007] Step 2: Since the traffic-aware integer programming model for planning the optimal path for each data stream has a very large number of variables and is difficult to apply to the integrated satellite network scenario of communication, navigation and remote sensing, a delay modeling without traffic awareness is performed. The delay from the overseas satellite to the domestic satellite is described as the number of waiting time slots between the current time slot and the next earliest time slot for linking with the domestic satellite; the delay from the domestic satellite to the ground station is described as the number of waiting time slots between the current time slot and the next earliest time slot for linking with the ground station. The detection matrix P is constructed. (t) Used to detect the total delay.

[0008] Step 3: The inter-satellite and satellite-ground joint topology planning problem is modeled as a traffic-oblivious integer programming (TOILP) model. This model aims to minimize the weighted sum of the delays from an overseas satellite to a domestic satellite and from a domestic satellite to a ground station. This traffic-oblivious integer programming model significantly reduces the number of decision variables and can solve the inter-satellite and satellite-ground joint topology planning problem in integrated communication, navigation, and telemetry satellite networks.

[0009] Furthermore, the satellite-to-ground link described in step 1 adopts the same time-division polling system as the inter-satellite link; that is, a ground station that supports narrow-beam satellite-to-ground links is deployed to realize integrated satellite-to-ground data interaction. In this way, the ground station and the satellite node communicate with each other in time slots as the basic unit. The ground station can switch the link-building satellite node object in different time slots to realize fast data interaction. The inter-satellite link adopts a time-division polling system, and the time-slot communication uses a store-carry-forward method to complete inter-satellite data transmission; the satellite system is as follows Figure 1 As shown, there are three network elements: domestic satellites, overseas satellites and ground stations. a To represent domestic satellites, assuming there are N a domestic stars; using the set V n To represent overseas satellites, assuming there are N n foreign stars; use the collection V g To represent the ground station, and use the antenna beam as the basic unit to distinguish, assuming there are N g Ground station nodes in the country. s =V a ∪V n Represents all N s The set of satellite nodes, then the set of all N nodes in this network is expressed as V = V s ∪V g ={v i |1≤i≤N}, and N=N s +N g =N a +N n +N g .

[0010] In integrated communication, navigation, and remote sensing satellite networks, narrow-beam, directional intersatellite links are used to increase information rates and enhance anti-interference capabilities. However, due to the limited number of narrow-beam antenna terminals onboard, the number of intersatellite links a satellite can establish simultaneously is far fewer than the number of satellites within its visible range. This approach is therefore adopted to ensure the measurement and communication performance required for normal satellite system operation. Traditionally, data transmission between ground station nodes and satellites uses wide-beam satellite-to-ground links with a non-time-division fixed link establishment system. In this system, the ground station node and satellite maintain a link within the visible arc.

[0011] The time-division polling system includes a three-layer topology processing structure. First, based on a finite state automaton modeling method, the constellation regression cycle to be planned is divided into K consecutive topology planning states of equal length, denoted as K = {1, 2, ..., k, ..., K}. Second, each topology planning state is divided into multiple identical polling cycles, and the chain establishment plan for each topology planning state is the same for U polling cycles. Finally, each polling cycle is divided into T shorter equal-length time slots, represented by the set T = {1, 2, ..., t, ..., T}. Under this structure, topology planning is performed once for each topology planning state, planning the topology within T time slots, and the planning results are reused U times within U polling cycles.

[0012] This assumes a simple one-hop routing strategy, meaning that data from an outbound satellite can only be transmitted to a domestic satellite in one hop, and vice versa. There will be no data transmission from an outbound satellite to an outbound satellite, or vice versa. Within each time slot, a domestic satellite can choose between two link establishment options: establishing an inter-satellite link with an outbound satellite to relay data from the outbound satellite, or establishing a satellite-to-ground link with a ground station node to transmit its own data accumulated over multiple time slots. Furthermore, topology planning must consider inter-satellite measurement requirements.

[0013] Furthermore, under the time-division polling system, the basic constraints and measurement constraints that need to be met by the entire system are established as follows; the basic constraints and measurement constraints are based on the three network elements of the satellite system: domestic satellites, overseas satellites, and ground stations;

[0014] (1) Basic constraints

[0015] The three-dimensional symmetric decision variable matrix X represents the result of the topological planning of the k-th state, the two-dimensional symmetric matrix Y represents the visibility relationship of the state, and a represents the number of narrow beam antennas carried by each satellite. The satellite-to-ground link must also meet the link establishment quantity constraint. The link establishment quantity constraint, link bidirectionality constraint, and visibility constraint are:

[0016]

[0017] (2) Measurement constraints

[0018] Satellites achieve precise orbit determination through inter-satellite ranging to meet the normal operation requirements of the constellation. Ignoring the geometric distribution factor, the performance of inter-satellite ranging is characterized by the number of ranging links. The number of ranging links is the number of inter-satellite links established with different satellites in one polling cycle. In order to meet the required precise orbit determination performance, the minimum number of ranging links that each satellite needs to establish in one polling cycle is L. min ; with l i,j Indicates satellite v i and satellite v j Whether an intersatellite link has been established within the T time slots of a polling cycle, and the BigM method is used to transform the constraint into a linear expression, and the measurement constraint of the network is expressed as:

[0019]

[0020] Among them, M is a parameter that needs to be set. The value of M should be large enough to make the inequality hold. Here, the value of M should be greater than the total number of time slots T.

[0021] The BigM method, also known as the Big M method, is a method for finding an initial basis feasible solution to a linear programming problem using artificial variables when the constraints are equal (=) and greater than (=). The law defines that after adding artificial variables to the constraints of the linear programming problem, a term with a coefficient of M or -M is added to the objective function accordingly.

[0022] Furthermore, the delay modeling under traffic-unawareness described in step 2: The optimal delay modeling method under traffic-awareness adopts the idea of ​​network flow optimization to plan the optimal path for each data flow. However, the traffic-aware integer programming model has a very large number of variables and is difficult to apply to the integrated satellite network scenario of communication, navigation and remote sensing. Therefore, delay modeling under traffic-unawareness is performed.

[0023] The number of waiting time slots for data from an overseas satellite to be transmitted to a domestic satellite is given by the following modeling method:

[0024]

[0025] Among them, the matrix Ψ is N n ×T two-dimensional matrix, the element ψ in Ψ i,t If it is equal to 1, it means that the i-th foreign satellite has established an inter-satellite link with a domestic satellite in the t-th time slot, and if it is equal to 0, no inter-satellite link has been established. i The maximum delay of data accumulated in multiple time slots is equal to the number of consecutive time slots without establishing an overseas-domestic ISL, which is represented by the number of consecutive 0s in the corresponding time slot column in the i-th row of the matrix Ψ. Construct the matrix P (t)Used to detect the number of consecutive t zeros in the matrix Ψ, and obtain the matrix Using the BigM method, through the matrix Δ (t) Will Elements greater than 1 are converted to 1, where is a parameter that needs to be set. The number of all 0s in the matrix is ​​the total number of waiting time slots for all foreign satellites to establish links with domestic satellites in T time slots, that is, the total delay. If the matrix Δ is maximized (1) to Δ (T) The total number of 1 elements in the , that is, the total number of 0 elements is minimized, that is, the total delay from the overseas satellite to the domestic satellite is optimized; Γ n→a The foreign satellite traffic f i is the number of non-zero elements of the weight; here, the traffic generated by all satellites in different time slots is a constant, f i,t is the traffic generated by the i-th satellite in the t-th time slot, that is, the traffic generated by the i-th satellite in any time slot f i =f i,t , you can also use f i,1 To express;

[0026] The number of waiting time slots for data transmission from domestic satellites to ground stations is given by the same modeling method as above:

[0027]

[0028] Among them, the element φ in the matrix Φ i,t If it is equal to 1, it means that the domestic star v i At time slot t, a satellite-to-ground link is established with a ground station node. If it is equal to 0, no satellite-to-ground link is established. Similarly, the matrix P (t) Used to detect the number of consecutive t zeros in the matrix Φ, is a parameter that needs to be set, and is related to the matrix Λ (t) Together the matrix Numbers greater than 1 are converted to 1, Γ a→g The domestic satellite traffic f i is the number of non-zero elements of the weight.

[0029] Furthermore, the traffic-aware integer programming model described in step 3. Based on the traffic-aware delay modeling in step 2, the inter-satellite-ground joint topology planning problem is modeled as a traffic-aware integer programming model for solution. The inter-satellite-ground joint topology planning in topology planning state k is expressed as follows:

[0030]

[0031] st basic constraints (1) to (4),

[0032] Measurement constraints (5) to (7),

[0033] The delay modeling constraints (8) to (15) and (16) are used. The traffic-aware integer programming model is dedicated to maximizing the matrix Δ (t) With the matrix Λ (t) The weighted sum of the number of non-zero elements in , that is, minimizing the weighted sum of the time delay from the overseas satellite to the domestic satellite and from the domestic satellite to the ground station, where γ is the weight factor.

[0034] Transformation of the target problem during the solution process: The target problem of inter-satellite and satellite-ground joint topology planning with the satellite-ground data interaction delay as the optimization target is a traffic-aware integer programming problem. It is transformed into a traffic-aware integer programming model to reduce the computational complexity.

[0035] To solve the target problem: First, establish basic constraints and measurement constraints, then perform delay modeling without traffic perception, that is, calculate the delay by the number of inter-satellite links and satellite-to-ground links and the time slot interval, and construct the detection matrix P (t) It is used to detect the total delay, and finally the inter-satellite and satellite-ground joint topology planning problem is modeled as a traffic-aware integer programming model for solution. The model is committed to minimizing the weighted sum of the delay from the overseas satellite to the domestic satellite and the domestic satellite to the ground station, with the overseas satellite traffic and domestic satellite traffic as weights respectively.

[0036] The present invention provides a satellite network inter-satellite and ground joint topology planning method for integrated data interaction. The satellite system adopts inter-satellite links (ISL) for high-precision orbit determination and space-based end-to-end telemetry, remote control and communication. Since the number of ISL terminals that can be carried on each satellite is limited, a time-division polling system is usually proposed for space link layer networking. By extending the polling mechanism to ground-satellite links (GSL), a unified management system for space segments and ground segments is realized. The present invention models the topology planning problem as a traffic-insensitive integer linear programming problem, and proposes a novel time delay modeling method to minimize the average delay of data interaction between satellites and ground stations while meeting the ranging requirements required for orbit determination. In addition, the algorithm has low computational complexity and is feasible in integrated communication, navigation and remote control satellite networks including GSL.

[0037] Beneficial Effects: The present invention proposes for the first time a satellite network inter-satellite, satellite-ground joint topology planning method for integrated data interaction in a communication, navigation, and remote sensing integrated satellite network, where both inter-satellite and satellite-to-ground links adopt a time-division polling system for link establishment, and the constraint of a limited number of link terminals applies to both inter-satellite and satellite-to-ground links. The satellite system adopts an inter-satellite, satellite-to-ground integrated data interaction system, which can unify the satellite-to-ground and inter-satellite link data distribution and processing mechanisms, as well as unified scheduling and management of the space segment and ground segment. Compared with satellite-to-ground data interaction under a fixed satellite-to-ground link system, inter-satellite, satellite-to-ground integrated topology planning, under the premise that the satellite-to-ground link adopts a time-division polling system, can significantly reduce data interaction latency. The designed topology planning algorithm can achieve near-optimal satellite-to-ground data interaction delay under the premise of meeting system measurement constraints with low computational complexity. The traffic-aware integer programming model established by the present invention is very close to the optimal delay value obtained by TAILP compared with the traffic-aware integer programming model with optimal delay performance. However, the traffic-aware model greatly reduces the number of decision variables, thereby reducing complexity and greatly reducing running time. It is feasible and suboptimal in the integrated satellite network scenario of communication, navigation and remote sensing. BRIEF DESCRIPTION OF THE DRAWINGS

[0038] Figure 1 It is a satellite system for integrated data interaction between satellites and the ground;

[0039] Figure 2 This is a schematic diagram of the time delay modeling from an overseas satellite to a domestic satellite (with the detection matrix P (3) for example);

[0040] Figure 3 is the running time of TAILP and TOILP in the specific embodiment scenario;

[0041] Figure 4 It is the average delay of TAILP and TOILP planning topologies in the specific embodiment scenario. DETAILED DESCRIPTION

[0042] To better illustrate the technical content of the present invention, the following, in conjunction with the accompanying drawings, uses a global navigation satellite system (GNSS) in which each satellite carries only one phased array antenna as a specific embodiment of an integrated communication and navigation satellite network. The proposed method for satellite network inter-satellite and satellite-ground joint topology planning for integrated data interaction is described in detail. A satellite node can only establish a link and communicate with one other satellite node or ground station at a time. The specific embodiments described herein are intended only to illustrate the present invention, and the scope of protection of the present invention is not limited to the following embodiments.

[0043] The present invention provides a satellite network inter-satellite and satellite-ground joint topology planning method for integrated data interaction in a satellite network integrating communication, navigation and remote sensing, and the steps are as follows:

[0044] Step 1: Deploy ground stations supporting narrow-beam satellite-to-ground links, enabling them to adopt the same time-division polling system as inter-satellite links, paving the way for integrated planning of inter-satellite and satellite-to-ground links. In each topology planning state under the time-division polling system, a visibility matrix, including inter-satellite and satellite-to-ground visibility, is calculated based on the ephemeris. This matrix serves as the input for topology planning, and basic and measurement constraints are established.

[0045] Step 2: Since the traffic-aware integer programming model for planning the optimal path for each data stream has a very large number of variables and is difficult to apply to the integrated satellite network scenario of communication, navigation and remote sensing, a delay modeling without traffic awareness is performed. The delay from the overseas satellite to the domestic satellite is described as the number of waiting time slots between the current time slot and the next earliest time slot for linking with the domestic satellite; the delay from the domestic satellite to the ground station is described as the number of waiting time slots between the current time slot and the next earliest time slot for linking with the ground station. The detection matrix P is constructed. (t) Used to detect the total delay.

[0046] Step 3: Model the inter-satellite-ground joint topology planning problem as a traffic-agnostic integer programming model. This model aims to minimize the weighted sum of delays from an overseas satellite to a domestic satellite and from a domestic satellite to a ground station. This traffic-agnostic integer programming model significantly reduces the number of decision variables and can solve the inter-satellite-ground joint topology planning problem in integrated communication, navigation, and telemetry satellite networks.

[0047] The scenario in the specific embodiment includes 4 overseas satellite nodes, 3 domestic satellite nodes and 1 ground station node. The inter-satellite visibility relationship and the satellite-ground visibility relationship are as follows: Figure 2 The parameter settings of the scene are listed in Table 1.

[0048] Table 1 Scene parameter settings

[0049]

[0050] The satellite-to-ground link described in step 1 uses the same time-division polling system as the inter-satellite link. In this way, the ground station and the satellite node communicate with each other in time slots. The ground station can switch the satellite node object in different time slots to achieve fast data interaction. Figure 1 As shown, there are three network elements: domestic satellites, overseas satellites and ground stations. a To represent domestic satellites, assuming there are N a = 3 domestic stars; use collection V n To represent overseas satellites, assuming there are N n= 4 foreign stars; use the set V g To represent the ground station, and use the antenna beam as the basic unit to distinguish, assuming there are N g = 1 domestic ground station node. s =V a ∪V n Represents all N s =7 satellite nodes, then the set of all N=8 nodes in this network is expressed as V=V s ∪V g ={v i |1≤i≤N}, and N=N s +N g =N a +N n +N g .

[0051] The three-layer topology processing structure based on a time-division polling system first divides the constellation regression cycle to be planned into K consecutive topology planning states of equal length, denoted as K = {1, 2, …, k, …, K}, using a finite state automaton modeling method. Next, each topology planning state is divided into multiple identical polling cycles, and the link establishment plan for each topology planning state is identical for U polling cycles. Finally, each polling cycle is divided into T = 6 shorter time slots of equal length, represented by the set T = {1, 2, …, t, …, T}. In this structure, topology planning is performed once for each topology planning state, planning the topology within T = 6 time slots. The planning results are reused U times within U polling cycles.

[0052] Taking topology planning state k as an example, we illustrate how to perform inter-satellite and satellite-ground joint topology planning. Within each time slot, a domestic satellite can either establish an inter-satellite link with an overseas satellite to relay data received from the overseas satellite, or establish a satellite-ground link with a ground station node to transmit its own data accumulated over multiple time slots. Furthermore, topology planning must consider inter-satellite measurement requirements.

[0053] Under the time-division polling system, the basic constraints and measurement constraints that need to be met in establishing the entire system are as follows.

[0054] (1) Basic constraints

[0055] The result of the topology planning for the kth state is represented by the three-dimensional symmetric decision variable matrix X, and the visibility relationship of the state is represented by the two-dimensional symmetric matrix Y. The satellite-to-ground link must also meet the constraint on the number of established links. The constraint on the number of established links, the constraint on the bidirectionality of the link, and the constraint on the visibility can be expressed as follows:

[0056]

[0057] (2) Measurement constraints

[0058] In the case that ground stations cannot be deployed globally, satellites achieve precise orbit determination through inter-satellite ranging to meet the normal operation requirements of the constellation. In this invention, the geometric distribution factor is ignored, and the performance of inter-satellite ranging is characterized by the number of ranging links. Since repeated ranging in a short period of time cannot improve the orbit determination performance, the ranging link of a satellite is defined as the inter-satellite link established with different satellites within a polling cycle. The number of ranging links is the number of inter-satellite links established with different satellites within a polling cycle. In order to meet the required precise orbit determination performance, the minimum number of ranging links that each satellite needs to establish within a polling cycle is given as L min Here, inter-satellite measurement is the main constraint, that is, to ensure that the minimum number of measurements between satellites reaches L min =2. i,j Indicates satellite v i and satellite v j Whether an intersatellite link has been established within a polling period of T = 6 time slots can be expressed as the measurement constraint of the network:

[0059]

[0060] It can be seen that even if satellite i and satellite j establish an intersatellite link within multiple time slots of a polling cycle, l i,j The value of is still 1, which is consistent with the definition of the ranging link.

[0061] However, it should be noted that constraint (21) is not a linear expression, so the BigM method is used to transform constraint (21) into the following linear expression:

[0062]

[0063] Among them, M is a parameter that needs to be set. The value of M should be large enough to make the inequality hold. Here, the value of M is 25. When ∑ t∈T x i,j,t = 0, that is, when satellite i and satellite j do not establish an intersatellite link in T = 6 time slots of a polling cycle, in order to satisfy the inequality on the left side of formula (24), l i,j is constrained to be 0; conversely, when ∑ t∈T x i,j,t > 0, that is, when satellite i and satellite j have established more than or equal to one intersatellite link in one polling cycle, in order to satisfy the inequality on the right side of formula (24), l i,j It is constrained to take the value of 1, and a sufficiently large value of M ensures that the inequality on the right side holds true in this case regardless of how many intersatellite links satellite i and satellite j have established.

[0064] Delay modeling without traffic awareness as described in step 2. The traffic-aware integer linear programming algorithm uses the concept of network flow optimization to plan the optimal path for each data flow. This model can achieve optimal delay modeling, but due to the extremely large number of variables, it is difficult to apply to the integrated communication, navigation and remote sensing satellite network scenario. Therefore, delay modeling without traffic awareness is performed.

[0065] Here, a relatively simple one-hop routing strategy is adopted, meaning that data from an outbound satellite can only be transmitted to a domestic satellite in one hop, and data from a domestic satellite can only be transmitted to a ground station in one hop. Regardless of the data volume and distribution of the traffic, it is assumed that all data accumulated by a node over multiple time slots can be transmitted to the link-establishing node within a single time slot. In this case, the latency from an outbound satellite to a domestic satellite is described as the number of waiting time slots between the current time slot and the earliest time slot for link establishment with the domestic satellite; the latency from a domestic satellite to a ground station is described as the number of waiting time slots between the current time slot and the earliest time slot for link establishment with the ground station. To simplify the aforementioned TAILP model, only the decision variables of topology planning are used to model latency. This latency is calculated using the number of inter-satellite links and satellite-to-ground links established and the time slot interval. This latency is traffic-agnostic and, while it cannot accurately represent the latency under known traffic conditions, it best reflects the communication performance of the planned topology without adding decision variables.

[0066] In order to model the number of waiting time slots for data generated by the foreign satellite to be transmitted to the domestic satellite, a size of N is defined. n ×T two-dimensional matrix Ψ, the elements in the matrix ψ i,t If it is equal to 1, it means that the i-th foreign satellite has established an inter-satellite link with a domestic satellite in the t-th time slot, which can be expressed as:

[0067]

[0068] like Figure 2 As shown, taking the network of the embodiment consisting of four overseas satellite nodes, three domestic satellite nodes, and one ground station node as an example, and considering a total of six time slots (T=6), overseas satellite v1 establishes an intersatellite link with the domestic satellite in the second and sixth time slots. Therefore, the first row of the matrix Ψ is [0 1 0 0 0 1]. In one-hop routing, the overseas satellite can only send data when it establishes an intersatellite link with the domestic satellite. Based on the above link establishment results, the data delays generated by overseas satellite v1 in all six time slots are [1 0 3 2 1 0], represented by squares in the figure. It can be seen that the maximum data delay accumulated by overseas satellite v1 in multiple time slots is equal to the number of consecutive time slots without an overseas-domestic ISL, represented by the number of consecutive zeros in the corresponding time slot column in the first row of matrix Ψ. Assuming the maximum data delay accumulated in multiple time slots is d, the total accumulated data delay is expressed as d + (d-1) + ... + 1.

[0069] In order to accurately express the delay of each overseas satellite node to the domestic satellite node in all time slots, T detection matrices are constructed, represented as P (1) To P (T) . Among them, the t-th detection matrix P (t) The definition is as follows:

[0070] There are T rows (T-t+1) and columns, column i It consists of three parts and can be expressed as:

[0071]

[0072] where p 1,i to p i-1,i is the first part, the value is 0; p i,i to p i+t-1,i is the second part, the value is 1; p i+t,i to p T,i is the third part, and its value is 0. (t) The first column, that is, There is no first part; for the detection matrix P (t) The last column, that is, There is no third part. In addition, when t=T, the detection matrix P (T) There is only one column, and all elements are 1.

[0073] Probe matrix P (t) The function of is to detect how many consecutive t elements are 0 in each row of the matrix Ψ, that is, to detect how many times the foreign satellite has not established a link with the domestic satellite in t consecutive time slots. Specifically, first detect the matrix P (t) The following matrix is ​​obtained:

[0074]

[0075] matrix The 0 element in the matrix Ψ indicates that there are t consecutive 0s in the corresponding row. (3) For example, see Figure 2 , P (3) After multiplication with Ψ, the resulting matrix The elements can be seen and elements It is equal to 0, which means that there are three consecutive zeros (t=3) in the first row of the matrix Ψ and two consecutive three zeros in the second row. It should be noted that the second row of the matrix Ψ is actually a continuous four zeros, that is, [0 0 0 0], and the detection matrix P is used. (3)During detection, four consecutive zeros can be regarded as two consecutive three zeros, that is, two [0 0 0].

[0076] Although the original intention of constructing the detection matrix is ​​to find the largest number of consecutive zeros in the overseas-domestic ISL link matrix Ψ, and then calculate the total delay d+(d-1)+...+1 of multiple accumulated data time slots based on the maximum delay d, it is difficult to detect all the consecutive zeros in the number matrix Ψ with a single detection matrix. (1) To P (T) By multiplying it with the matrix Ψ, we can get the situation where all rows of the matrix Ψ corresponding to all outbound stars have continuous 0s. Figure 2 For example, if the overseas satellite v1 has no domestic satellite link in three consecutive time slots t=3, t=4, and t=5, the third, fourth, and fifth columns of the first row of the matrix Ψ are 0. After multiplying Ψ by the detection matrix, we get:

[0077] ·With P (1) After multiplication, the resulting matrix The corresponding position in is 0, which means that there are 3 [0]s at the corresponding position of Ψ. Its physical meaning is that the earliest time slot without link establishment with the domestic satellite needs to wait for 3 time slots (d=3), that is, node v1 needs to wait for 3 time slots at time slot t=3;

[0078] ·With P (2) After multiplication, the resulting matrix The corresponding position in is 0, which means that there are 2 [0 0] at the corresponding position of Ψ. Its physical meaning is that node v1 needs to wait for 2 time slots at time slot t = 4;

[0079] ·With P (3) After multiplication, the resulting matrix The corresponding position in is 0, which means that there is 1 [00 0] at the corresponding position of Ψ. Its physical meaning is that node v1 needs to wait for 1 time slot at time slot t = 5;

[0080] ·With P (4) 、P (5) 、P (6) After multiplication, since there are no more than 3 consecutive 0s in the first row of the matrix Ψ, the resulting matrix The matrix is ​​full of 1s and has no 0 elements.

[0081] And d+(d-1)+...+1 is exactly the total delay of all data accumulated by the overseas satellite in d consecutive time slots without establishing a link with the domestic satellite. Therefore, we have the following conclusion: After multiplying Ψ with all detection matrices, the resulting matrix is to The number of all 0s in is the total number of waiting time slots for all overseas satellites to establish links with domestic satellites in T time slots, that is, the total delay.

[0082] Given the number of time slots T = 6 and the number of satellites N n =4, then the matrix to The total number of elements in is a fixed value. If the matrix is ​​maximized to The total number of non-zero elements in the matrix is ​​minimized, that is, the total delay from the foreign satellite to the domestic satellite is optimized. To achieve this goal, the matrix Δ (t) is used to Elements greater than 1 are converted to 1:

[0083]

[0084] Δ (t) All elements in are 0-1 Boolean variables. The number of non-zero elements in is converted to Δ (t) The number of elements 1 in Figure 2 by As an example, It is a parameter that needs to be set. The value of should be large enough to make the inequality hold. The elements greater than 1 in are converted to 1. Set to 25. Therefore, as long as all matrices Δ are maximized (t) The sum of the elements in , that is, maximizing the number of elements 1, can minimize the total delay from the overseas satellite to the domestic satellite. The optimization goal to be maximized can be expressed as follows:

[0085]

[0086] in, It corresponds to overseas star v i The number of non-zero elements, f i is the corresponding traffic weight factor. Here, the traffic generated by all satellites in different time slots is a constant, that is, f i =f i,t , you can also use f i,1 To express.

[0087] The number of waiting time slots for data transmission from domestic satellites to ground stations is given by the same modeling method as above:

[0088]

[0089] Among them, the element φ in the matrix Φ i,tIf it is equal to 1, it means that the domestic star v i At time slot t, a satellite-to-ground link is established with a ground station node. If it is equal to 0, no satellite-to-ground link is established. Similarly, the matrix P (t) Used to detect the number of consecutive t zeros in the matrix Φ, is a parameter that needs to be set, and is related to the matrix Λ (t) Together the matrix In the specific embodiment, the number greater than 1 is converted to 1. Set to 50. a→g It is the number of non-zero elements with domestic satellite traffic as weight.

[0090] The traffic-aware integer programming model described in step 3. Based on the traffic-aware delay modeling in step 2, the inter-satellite and ground-to-space joint topology planning problem is modeled as a traffic-aware integer linear programming model for solution.

[0091] The inter-satellite and ground-to-space joint topology planning in topology planning state k is expressed as follows:

[0092]

[0093] st basic constraints (17) to (20),

[0094] Measurement constraints (21) to (24),

[0095] Delay modeling constraints (25) to (34). (35),

[0096] The traffic-aware integer programming model aims to maximize the matrix Δ (t) With the matrix Λ (t) The weighted sum of the number of non-zero elements in [ ] is used to minimize the weighted sum of the delays from the overseas satellite to the domestic satellite and from the domestic satellite to the ground station, where γ is a weighting factor, set to 0.5 in this embodiment. Compared to TAILP, which uses optimal delay modeling, TOILP does not plan the optimal path for each data stream. This significantly reduces the number of decision variables and can solve the inter-satellite, satellite-ground joint topology planning problem in integrated communication, navigation, and remote sensing satellite networks.

[0097] Figure 3 The following demonstrates how the runtime of the TAILP and TOILP algorithms changes with the number of time slots in a specific embodiment scenario. As expected, the TOILP algorithm proposed in this invention significantly reduces the runtime by reducing the integer variables compared to the TAILP algorithm, and the runtime of TOILP only increases slightly with the number of time slots. Under the topologies planned by the two algorithms, the average delay from satellite to ground station is as follows: Figure 4As shown in Figure 2, the average delay is the average of the delays between data frames generated by all satellite nodes in all time slots and sent to any ground station node. As can be seen, the delay of TOILP is very close to the optimal delay value obtained by TAILP, which also proves the suboptimal nature of TOILP.

[0098] The above descriptions are merely embodiments of the present invention and are 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 in the scope of protection of the present invention.

Claims

1. A satellite network inter-satellite and ground joint topology planning method for integrated data interaction, characterized in that: Satellite network is a satellite system with three network elements: domestic satellites, overseas satellites and ground stations. a To represent domestic satellites, assuming there are N a domestic stars; using the set V n To represent overseas satellites, assuming there are N n foreign stars; use the collection V g To represent the ground station, and use the antenna beam as the basic unit to distinguish, assuming there are N g Domestic ground station nodes; V s =V a ∪V n Represents all N s The set of satellite nodes, then the set of all N nodes in this network is expressed as V = V s ∪V g ={v i |1≤i≤N}, and N=N s +N g =N a +N n +N g In an integrated satellite network for communication, navigation, and remote sensing, the satellite system uses narrow-beam directional inter-satellite links. In this scenario, both inter-satellite and satellite-to-ground links adopt a time-division polling system. While meeting system measurement requirements, the latency of satellite-to-ground data interaction is optimized. The inter-satellite-to-ground integrated topology planning objective problem is modeled as a traffic-agnostic integer linear programming problem. The steps are as follows: Step 1: Deploy ground stations supporting narrow-beam satellite-to-ground links, enabling them to adopt the same time-division polling system as inter-satellite links, paving the way for integrated planning of inter-satellite and satellite-to-ground links. In each topology planning state under the time-division polling system, calculate the visibility matrix based on the ephemeris, including inter-satellite and satellite-to-ground visibility conditions, as input for topology planning. Basic and measurement constraints are then established. Step 2: Delay modeling is performed without traffic perception. The delay from the overseas satellite to the domestic satellite is described as the number of waiting time slots between the current time slot and the next earliest time slot for establishing a link with the domestic satellite; the delay from the domestic satellite to the ground station is described as the number of waiting time slots between the current time slot and the next earliest time slot for establishing a link with the ground station. The detection matrix P is constructed. (t) Used to detect the total delay; Step 3: Model the inter-satellite, satellite-ground joint topology planning problem as a traffic-agnostic integer programming model and solve it. The model aims to minimize the weighted sum of the delays from overseas satellites to domestic satellites and from domestic satellites to ground stations. This traffic-agnostic integer programming model can significantly reduce the number of decision variables and solve the inter-satellite, satellite-ground joint topology planning problem in integrated communication, navigation, and remote sensing satellite networks. The satellite-to-ground link described in step 1 adopts the same time division polling system as the inter-satellite link; That is, deploying ground stations that support narrow-beam satellite-to-ground links to achieve integrated data interaction between satellites and the ground. In this way, the ground station and satellite nodes communicate using time slots as the basic unit, and the ground station can switch the satellite node objects for link establishment in different time slots to achieve rapid data interaction. In this way, the ground station and satellite nodes communicate using time slots as the basic unit, and the ground station can switch the satellite node objects for link establishment in different time slots to achieve rapid data interaction. The inter-satellite link adopts a time-division polling system, and the time-division communication adopts a store-carry-forward method to complete inter-satellite data transmission. The time-division polling system includes a three-layer topology processing structure. First, based on a finite state automaton modeling method, the constellation regression cycle to be planned is divided into K consecutive topology planning states of equal length, denoted as K = {1, 2, ..., k, ..., K}. Second, each topology planning state is divided into multiple identical polling cycles, and the chain establishment plan for each topology planning state is the same for U polling cycles. Finally, each polling cycle is divided into T shorter equal-length time slots, represented by the set T = {1, 2, ..., t, ..., T}. Under this structure, topology planning is performed once for each topology planning state, planning the topology within T time slots, and the planning results are reused U times within U polling cycles. A simple one-hop routing strategy is set up, that is, the overseas satellite can only transmit to the domestic satellite in one hop, and the domestic satellite can only transmit to the ground station in one hop, and there will be no situation where the overseas satellite transmits data to the overseas satellite or the domestic satellite transmits data to the domestic satellite; in each time slot, there are two types of link establishment options for the domestic satellite. The first is to choose to establish an inter-satellite link with the overseas satellite as a transit to receive overseas satellite data, and the second is to choose to establish a satellite-to-ground link with the ground station node to transmit its own accumulated data in multiple time slots; in addition, the need for inter-satellite measurement must also be considered when conducting topology planning.

2. The satellite network inter-satellite and satellite-ground joint topology planning method for integrated data interaction according to claim 1 is characterized in that: Under the time-division polling system, the basic constraints and measurement constraints that need to be met in establishing the entire system are as follows; (1) Basic constraints The result of the k-th state topology planning is represented by a three-dimensional symmetric decision variable matrix X, where a represents the number of narrow-beam antennas carried by each satellite. The satellite-to-ground link must also meet the link establishment quantity constraint. The link establishment quantity constraint, link bidirectionality constraint, and visibility constraint are: (2) Measurement constraints Satellites achieve precise orbit determination through inter-satellite ranging to meet the normal operation requirements of the constellation. Ignoring the geometric distribution factor, the performance of inter-satellite ranging is characterized by the number of ranging links. The number of ranging links is the number of inter-satellite links established with different satellites in one polling cycle. In order to meet the required precise orbit determination performance, the minimum number of ranging links that each satellite needs to establish in one polling cycle is L. min ; with l i,j Indicates satellite v i and satellite v j Whether an intersatellite link has been established within the T time slots of a polling cycle, and the BigM method is used to transform the constraint into a linear expression, and the measurement constraint of the network is expressed as: Among them, M is a parameter that needs to be set. The value of M should be large enough to make the inequality hold. Here, the value of M should be greater than the total number of time slots T.

3. The satellite network inter-satellite and satellite-ground joint topology planning method for integrated data interaction according to claim 2 is characterized in that: Delay modeling without traffic awareness as described in step 2: The optimal delay modeling method with traffic awareness uses the concept of network flow optimization to plan the optimal path for each data flow. However, the traffic-aware integer programming model has a large number of variables and is difficult to apply to integrated satellite networks for communication, navigation, and remote sensing. Therefore, delay modeling without traffic awareness is performed. The number of waiting time slots for data from an overseas satellite to be transmitted to a domestic satellite is given by the following modeling method: Among them, the matrix Ψ is N n ×T two-dimensional matrix, the element ψ in Ψ i,t =1 means that the i-th foreign satellite has established an inter-satellite link with a domestic satellite in the t-th time slot, and =0 means that no inter-satellite link has been established; foreign satellite v i The maximum delay of the data accumulated in multiple time slots is equal to the number of consecutive time slots without establishing an overseas-domestic ISL, which is represented by the number of consecutive 0s in the corresponding time slot column in the i-th row of the matrix Ψ; construct the matrix P (t) Used to detect the number of consecutive t zeros in the matrix Ψ, and obtain the matrix Using the BigM method, through the matrix Δ (t) Will Elements greater than 1 are converted to 1, where is a parameter that needs to be set. The number of all 0s in the matrix is ​​the total number of waiting time slots for all foreign satellites to establish links with domestic satellites in T time slots, that is, the total delay. If the matrix Δ is maximized (1) to Δ (T) The total number of 1 elements in the , that is, the total number of 0 elements is minimized, that is, the total delay from the overseas satellite to the domestic satellite is optimized; Γ n→a is the number of non-zero elements with the outbound satellite traffic as the weight; here, the traffic generated by all satellites in different time slots is a constant, f i,t is the traffic generated by the i-th satellite in the t-th time slot, that is, the traffic generated by the i-th satellite in any time slot f i =f i,t ; The number of waiting time slots for data transmission from domestic satellites to ground stations is given by the same modeling method as above: Among them, the element φ in the matrix Φ i,t If it is equal to 1, it means that the domestic star v i At time slot t, a satellite-to-ground link is established with a ground station node. If it is equal to 0, no satellite-to-ground link is established. Similarly, the matrix P (t) Used to detect the number of consecutive t zeros in the matrix Φ, the matrix Λ (t) The matrix The numbers greater than 1 are converted to 1, Γ a→g It is the number of non-zero elements with domestic satellite traffic as weight.

4. The satellite network inter-satellite and satellite-ground joint topology planning method for integrated data interaction according to claim 3 is characterized in that: Furthermore, the traffic-aware integer programming model described in step 3 is based on the traffic-aware delay modeling in step 2, and the inter-satellite and ground joint topology planning problem is modeled as a traffic-aware integer programming model for solution. The inter-satellite and ground joint topology planning in the topology planning state k is expressed as follows: st basic constraints (1) to (4), Measurement constraints (5) to (7), Delay modeling constraints (8) to (15). (16), The traffic-aware integer programming model aims to maximize the matrix Δ (t) With the matrix Λ (t) The weighted sum of the number of non-zero elements in , that is, minimizing the weighted sum of the time delay from the overseas satellite to the domestic satellite and from the domestic satellite to the ground station, where γ is the weight factor.

Images