A navigation system data injection route planning method and system

Abstract

The present invention provides a navigation system data upload route planning method and system. The method obtains the set of node satellites and destination satellites participating in data upload within a planning period, as well as the node satellite configuration parameter requirements; then, based on the node and destination satellite sets, calculates the shortest path sequence from any node satellite to the destination satellite; constructs an integer programming model for the ground data upload route optimization problem based on the node and destination satellite sets and the shortest path sequence from any node satellite to the destination satellite; solves the integer programming model, and obtains a routing plan. This method can quickly generate an upload route planning plan from a domestic node satellite to an overseas destination satellite. Experimental results demonstrate the effectiveness of the present invention in solving the upload route planning problem.

Description

Technical Field

[0001] The present invention belongs to the field of satellite mission planning, and in particular relates to a navigation system data injection route planning method and system. Background Art

[0002] To ensure the stable operation of the navigation constellation and provide high-precision positioning services, the ground control system must upload various service data to designated destination satellites. This data includes navigation messages, telecommands, time slot tables, and routing tables. Due to the geographical constraints of BeiDou ground stations, the ground data uploading process categorizes navigation satellites into domestic and overseas satellites. Satellites that are continuously visible to the ground station during the uploading period are designated domestic satellites, while all other satellites are designated overseas satellites. If the destination satellite for data uploading is a domestic satellite, uploading can be accomplished directly using satellite-to-ground links, resulting in short transmission latency and high stability. However, if the destination satellite is an overseas satellite, the data must first be uploaded to the domestic satellite via the satellite-to-ground link. The data is then transferred to the overseas destination satellite via inter-satellite links, using the domestic satellite as a transit node. This results in longer transmission times and higher risks. To achieve rapid ground data uploading and minimize the risk of packet loss and bit errors during data transmission, an effective data uploading routing planning method is needed to minimize the transmission latency of each data element while meeting constraints.

[0003] The above-noted latency is divided into satellite-to-ground transmission latency and inter-satellite transmission latency. Satellite-to-ground transmission latency (milliseconds) is much smaller than inter-satellite transmission latency (seconds). Therefore, when calculating data for the above-noted latency, only the inter-satellite transit transmission latency from domestic satellites to overseas satellites is considered.

[0004] Due to factors such as satellite hardware capabilities and system stability, the ground system limits the number of available node satellites and the relay capacity of each node satellite during the upload process. Relay capacity refers to the number of destination satellites a node satellite can simultaneously relay data from. Therefore, when planning data upload routes, it is important to consider the proper allocation of node satellite resources, ensuring that overseas data is uploaded to a node satellite closer to the destination satellite, enabling rapid data injection. Upload transmission latency is measured in time slots, denoted as slots.

[0005] Existing research on navigation system data routing can be divided into two categories: one focuses solely on the scheduling of satellite-to-ground links, without considering the inter-satellite transit transmission process; the other, while planning inter-satellite data transmission paths, fails to consider the number and capacity constraints of node satellites. Therefore, existing research on routing planning for data routing cannot meet the routing constraints and requirements of BeiDou global navigation system engineering applications, necessitating the development of new data routing planning methods for routing. Summary of the Invention

[0006] The technical problem to be solved by the present invention is how to quickly provide a routing plan from a domestic node satellite to an overseas destination satellite when the number and capacity of node satellites are limited. A navigation system data routing planning method and system are proposed.

[0007] In order to solve the above technical problems, the technical solution adopted by the present invention is:

[0008] A method for planning a route for uploading navigation system data includes the following steps:

[0009] Step 1: Obtain the node satellite set and destination satellite set participating in data uploading within the planning period, as well as the configuration parameter requirements of the node satellites. The destination satellite refers to the satellite to which the data is ultimately uploaded, and the node satellite refers to the domestic satellite that starts uploading data from the ground station.

[0010] Step 2: Calculate the shortest path sequence from any node star to the destination star based on the node star set and the destination star set;

[0011] Step 3: Construct an integer programming model for the ground data routing optimization problem based on the node star set, the destination star set, and the shortest path sequence from any node star to the destination star;

[0012] Step 4: Solve the integer programming model;

[0013] Step 5: Output the routing plan obtained.

[0014] Furthermore, the integer programming model is:

[0015] The objective function is:

[0016]

[0017] The constraints are:

[0018]

[0019] x ij ≤y j , i=1,...,m; j=1,...,n (3)

[0020]

[0021]

[0022] y j ∈{0, 1}, j=1,...,n (6)

[0023] x ij ∈{0, 1}, i=1,...,m; j=1,...,n (7)

[0024] Among them, the objective function is to minimize the average transmission path length from each destination satellite to the node satellite; x ij 、y j are two decision variables, x ij Indicates whether the destination star i data chooses node star j to forward. The value 1 means selection, and the value 0 means no selection. j Indicates whether node star j is selected. The value is 1 if it is selected, otherwise it is 0. Both are 0-1 variables. i represents the target star number, j represents the node star number, m represents the number of target stars, and n represents the number of node stars. The shortest path length l ij represents the shortest transmission delay between destination satellite i and node satellite j;

[0025] Formula 2 indicates that each target star must and can only select one node star;

[0026] Formula 3 indicates that only the selected node satellite can be selected by the destination satellite for intersatellite transfer;

[0027] Formula 4 limits the transfer capacity of each selected node star; c j is the transit capacity of node star j;

[0028] Formula 5 represents the number constraint of node stars; P is the upper limit of the number of node stars that can be selected;

[0029] Furthermore, the method for solving the integer programming model in step 4 is a hybrid genetic programming algorithm based on gene expression.

[0030] Furthermore, the hybrid genetic programming algorithm based on gene expression is specifically as follows:

[0031] Step 4.1: Randomly generate N chromosomes according to the gene expression code to form an initial population. The chromosomes are composed of a single or multiple genes. A gene is a set of linear character expressions of fixed length. The gene is divided into a head segment and a tail segment. The head segment consists of function characters and parameter characters, and the tail segment consists of parameter characters. The parameter characters are characteristic parameters of the path in the shortest path sequence.

[0032] Step 4.2: Calculate the fitness of each individual in the population;

[0033] Step 4.3: Perform selection, crossover, and mutation operations within the population;

[0034] Step 4.4: Generate a new generation of population;

[0035] Step 4.5: Determine whether the termination condition is met. If so, output the optimal result; if not, return to step 4.2.

[0036] Furthermore, the method for calculating individual fitness in step 4.2 is:

[0037] Step 4.2.1: Decoding the gene expression into a mathematical equation;

[0038] Step 4.2.2: Obtain the characteristic parameter values ​​of each path in the shortest path sequence, substitute them into the mathematical equation, and use the calculation result of the equation as the score of each path;

[0039] Step 4.2.3: Form a new path sequence according to the order of path scores;

[0040] Step 4.2.4: Prioritize the paths with high scores for the destination star path allocation, perform constraint checks on the allocated paths, delete infeasible paths, and obtain a feasible routing solution;

[0041] Step 4.2.5: Substitute the feasible routing scheme into the objective function to obtain the sum of the transmission path lengths of all destination stars, which is used as the individual fitness value.

[0042] Furthermore, the function characters in step 4.1 include +, -, *, / , max, min, and the characteristic parameters include:

[0043] WN: path length;

[0044] WAVE: the mean path length of the destination star to all node stars related to the path;

[0045] WVAR: path length variance of the path-related destination star to all node stars;

[0046] WMAX: The maximum path length of the path-related destination star to all node stars;

[0047] WMIN: The minimum path length of the destination star to all node stars related to the path;

[0048] NAVE: the mean path length from the node star to all destination stars related to the path;

[0049] NVAR: Path length variance of path-related node stars to all destination stars;

[0050] NMAX: The maximum path length of the path-related node star to all destination stars;

[0051] NMIN: The minimum path length of the node star related to the path to all destination stars;

[0052] NCAP: the remaining transit capacity of the node star associated with the path;

[0053] NRLY: Whether the node star associated with the path is available.

[0054] Furthermore, in the gene expression coding of step 4.1, if the length of the head segment is h, the length t of the tail segment must satisfy: t≥h(n-1)+1; where n is the maximum number of input parameters of all functions in the head segment.

[0055] Furthermore, the method for decoding gene expression in step 4.2.1 is:

[0056] Step 4.2.1.1: Decode the gene into a syntax tree. Start with the first character of the gene as the root node of the syntax tree. Then, fill in the child nodes of the function node in the syntax tree from left to right and from top to bottom. When all the leaf nodes in the tree are parameter nodes, the decoding is complete.

[0057] Step 4.2.1.2, decode the syntax tree into a mathematical equation. The overall decoding order is from leaf nodes to the root node. Starting from the lower left corner of the syntax tree, use the parent node's operator to connect the child node's equations according to the operation rules to form a new equation, and then traverse all nodes in order from right to top.

[0058] Furthermore, the method for performing constraint checking on the assigned paths in step 4.2.4 and removing infeasible paths to obtain a feasible routing solution set is:

[0059] Step 4.2.1.1: Select the first path in the path sequence and check the number of node stars and transit capacity constraints;

[0060] Step 4.2.1.2: If any constraint is violated, delete this path and return to step 4.2.1.1; otherwise, assign this path to the destination star corresponding to the path and delete the paths in the path sequence that have the same overseas destination star as this path;

[0061] Step 4.2.1.3: If the path sequence is not empty, update the characteristic parameters of each path, re-complete the path scoring and return to step 4.2.1.1; otherwise, the path allocation is completed and the feasible routing solution for the problem is output.

[0062] Furthermore, in step 2, the method for calculating the shortest path sequence from any node star to the destination star according to the node star set and the destination star set is a breadth-first search algorithm.

[0063] The present invention also provides a navigation system data injection route planning system, which is characterized by comprising the following modules:

[0064] Information acquisition module: used to obtain the node satellite set and destination satellite set participating in data uploading within the planning period, as well as the configuration parameter requirements of the node satellite. The destination satellite refers to the satellite to which the data is finally uploaded, and the node satellite refers to the domestic satellite that starts uploading data from the ground station.

[0065] The shortest path sequence calculation module is used to calculate the shortest path sequence from any node star to the destination star based on the node star set and the destination star set obtained by the information acquisition module;

[0066] Integer programming model construction module: used to construct an integer programming model for the ground data routing optimization problem based on the node star set and the destination star set as well as the shortest path sequence from any node star to the destination star;

[0067] Solving module: used for solving the integer programming model;

[0068] Output module: used to output the routing plan obtained by the solution module.

[0069] By adopting the above technical solution, the present invention has the following beneficial effects:

[0070] This invention provides a method and system for planning navigation system data routing. With the goal of minimizing the transmission delay of each overseas data point, a 0-1 integer programming model is established for this problem. This model is then solved using a hybrid genetic programming based on gene expression (HGPGE) algorithm. This method can quickly generate routing plans from domestic node satellites to overseas destination satellites. Experimental results demonstrate the effectiveness of this method in solving routing planning problems. BRIEF DESCRIPTION OF THE DRAWINGS

[0071] Figure 1 This is a flow chart of the system of the present invention;

[0072] Figure 2 This is the framework diagram of the genetic programming algorithm;

[0073] Figure 3 Schematic diagram of gene expression;

[0074] Figure 4 This is a schematic diagram of the gene syntax tree expression;

[0075] Figure 5 Schematic diagram for decoding gene syntax tree into mathematical equation;

[0076] Figure 6 Schematic diagram of the construction algorithm based on path scoring;

[0077] Figure 7 It is a schematic diagram of a single point crossover;

[0078] Figure 8 Schematic diagram of fragment crossover;

[0079] Figure 9 Schematic diagram of single-point mutation operator;

[0080] Figure 10 Schematic diagram of the fragment flip operator;

[0081] Figure 11 Schematic diagram of the fragment transfer operator. DETAILED DESCRIPTION

[0082] The technical solution of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the embodiments described are only some embodiments of the present invention, not all embodiments. All other embodiments obtained by ordinary technicians in this field based on the embodiments of the present invention without making any creative efforts shall fall within the scope of protection of the present invention.

[0083] The transmission requirements and constraints for ground data uploading differ from those for onboard data downlinking. Compared to onboard data, ground data is generated less frequently and requires less data transmission, so the uploading and transit process does not require consideration of inter-satellite link bandwidth capacity. Domestic data can be uploaded directly via the satellite-ground link, while overseas data uploading requires the selection of a specific node satellite and the planning of an inter-satellite transmission route from the node satellite to the destination satellite. Due to the complex nature of ground data, the generation time and uploading period of some data types are not fixed. To facilitate routing management, the system stipulates that after arriving at a domestic node satellite, ground data must wait for the start of the next superframe before inter-satellite transmission can begin. Therefore, the minimum planning period for uploading routing design is the superframe. Inter-satellite transit routes for overseas data all start at the first transient of the superframe. Taking into account factors such as satellite hardware capabilities and system operational stability, the navigation system's ground control system limits the number of node satellites available for data uploading and the transit capacity of each node satellite. Transmission capacity refers to the number of destination satellites a node satellite can simultaneously transmit data from. Therefore, when designing the upload route, it is necessary to consider the reasonable allocation of node satellite resources, so that overseas data can be uploaded to a node satellite closer to the destination satellite, thereby completing the rapid injection of data.

[0084] Figures 1 to 11 A specific embodiment of a navigation system data injection route planning method is shown. Figure 1 As shown, the following steps are included:

[0085] Step 1: Obtain the set of node satellites and destination satellites participating in data upload within the planning period, as well as the transfer capabilities and constraints of the node satellites. The destination satellite refers to the satellite to which data is ultimately uploaded, and the node satellite refers to the domestic satellite that initiates data upload from the ground station. Using the BeiDou navigation system as an application scenario, an experimental case set consisting of 30 satellites was designed.

[0086] In this embodiment, the minimum planning period for topological routing optimization is a superframe. The topological structure of each superframe within a topological period is consistent. Therefore, one topological routing optimization case can be generated for each topological period. This embodiment randomly generates 120 topological period cases, selects 100 of these cases as the training set for the HGPGE algorithm, and the remaining 20 cases as the algorithm test set. The upper limit of the available number of node stars is set to 3, and the node star transit capacity is set to 7.

[0087] Step 2: Based on the set of node stars and the set of destination stars, a sequence of shortest paths from any node star to the destination star is calculated. In this embodiment, the shortest path sequence refers to a sequence of various optional paths from the node star to the destination star, obtained using a shortest path algorithm. The shortest path algorithm in this embodiment is a commonly used breadth-first search algorithm, which treats the inter-satellite network as a general directed graph and uses this algorithm to obtain the shortest path from each node star to the destination star.

[0088] Step 3: Construct an integer programming model for the ground data routing optimization problem based on the node star set, the destination star set, and the shortest path sequence from any node star to the destination star;

[0089] In this embodiment, in order to establish the planning model, the following assumptions are made: 1) The satellite-to-ground link is continuously connected throughout the entire topological cycle, and the ground station can achieve full-segment tracking and data transmission of the satellite within the territory; 2) The inter-satellite network topology structure is determined within each superframe, and link failures are not considered; 3) The transmission time of data from the ground station to the node satellite is 0; 4) Only data that arrives at the node satellite before the start of the superframe can be transferred between satellites; 5) The bandwidth capacity of the inter-satellite link is large enough, and there is no need to consider bandwidth limitations during the inter-satellite transfer process.

[0090] The integer programming model is:

[0091] The objective function is:

[0092]

[0093] The constraints are:

[0094]

[0095] x ij ≤y j , i=1,...,m; j=1,...,n (3)

[0096]

[0097]

[0098] y j∈{0, 1}, j=1,...,n (6)

[0099] x ij ∈{0, 1}, i=1,...,m; j=1,...,n (7)

[0100] Among them, the objective function is to minimize the average transmission path length from each destination satellite to the node satellite; x ij 、y j are two decision variables, x ij Indicates whether the destination star i data chooses node star j to forward. The value 1 means selection, and the value 0 means no selection. j Indicates whether node star j is selected. The value is 1 if it is selected, otherwise it is 0. Both are 0-1 variables. i represents the destination star, j represents the node star, m represents the number of destination stars, and j represents the number of node stars. The shortest path length l ij represents the shortest transmission delay between destination satellite i and node satellite j;

[0101] Formula 2 indicates that each target star must and can only select one node star;

[0102] Formula 3 indicates that only the selected node satellite can be selected by the destination satellite for intersatellite transfer;

[0103] Formula 4 limits the transfer capacity of each selected node star; c j is the transit capacity of node star j;

[0104] Formula 5 represents the selection quantity constraint of node stars; P is the upper limit of the number of optional node stars.

[0105] Step 4: Solve the integer programming model. In this embodiment, the integer programming model is solved by using a hybrid genetic programming algorithm based on gene expression (HGPGE). In the genetic programming method used in the present invention, the encoding object of the chromosome usually includes a function set and a parameter set. When decoding, a mathematical equation about the characteristic parameters is first generated, and a feasible solution is generated by the mathematical equation, and then substituted into the programming model to obtain the objective function value. Figure 2 As shown in Figure 2, the hybrid genetic programming algorithm based on gene expression is specifically:

[0106] Step 4.1: Randomly generate N chromosomes according to the gene expression code to form an initial population. The chromosome means that the chromosome is composed of a single or multiple genes. A gene refers to a set of linear character expressions of fixed length. The character types include function characters and parameter characters. The gene is divided into a header segment and a tail segment. The header segment consists of function characters and parameter characters, and the tail segment consists of parameter characters. The parameter characters are characteristic parameters of the path in the shortest path sequence. In this embodiment, the function characters include +, -, *, / , max, min, where max and min represent the maximum and minimum values ​​of the two parameters, respectively. The number of input parameters for each function is 2, so the maximum number of input parameters for all functions is n=2. The parameter set is usually composed of various characteristic parameters, and the extraction of characteristic parameters needs to be closely combined with the characteristics of the specific problem. The present invention takes each shortest path as the object, combines the optimization goal of the problem and the number of node stars and the transit capacity constraint properties, and extracts 11 characteristic parameters related thereto, specifically:

[0107] WN: path length;

[0108] WAVE: the mean path length of the destination star to all node stars related to the path;

[0109] WVAR: path length variance of the path-related destination star to all node stars;

[0110] WMAX: The maximum path length of the path-related destination star to all node stars;

[0111] WMIN: The minimum path length of the destination star to all node stars related to the path;

[0112] NAVE: the mean path length from the node star to all destination stars related to the path;

[0113] NVAR: Path length variance of path-related node stars to all destination stars;

[0114] NMAX: The maximum path length of the path-related node star to all destination stars;

[0115] NMIN: The minimum path length of the node star related to the path to all destination stars;

[0116] NCAP: the remaining transit capacity of the node star associated with the path;

[0117] NRLY: Whether the node star associated with the path is available.

[0118] Among them, WN reflects the objective attributes of the path itself, WAVE, WVAR, WMAX, WMIN, NAVE, NVAR, NMAX, and NMIN reflect the transmission capabilities of two types of resources related to the path from eight statistical perspectives, and NCAP and NRLY reflect the constraint attributes of the number of node stars and the transfer capacity. Figure 3 Schematic diagram of the gene expression coding shown,

[0119] Step 4.2: Calculate the fitness of each individual in the population in turn; Figure 5 As shown, the fitness is calculated as:

[0120] Step 4.2.1: Decoding the gene expression into a mathematical equation. In this embodiment, the method for decoding the gene expression is:

[0121] Step 4.2.1.1: Decode the gene into a syntax tree. Start with the first character of the gene as the root node of the syntax tree. Then, fill in the child nodes of the function node in the syntax tree from left to right and from top to bottom. When all the leaf nodes in the tree are parameter nodes, the decoding is complete.

[0122] Step 4.2.1.2, decode the syntax tree into a mathematical equation. The overall decoding order is from leaf nodes to the root node. Starting from the lower left corner of the syntax tree, use the parent node's operator to connect the child node's equations according to the operation rules to form a new equation, and then traverse all nodes in order from right to top.

[0123] like Figure 4 As shown, starting from the first character + of the gene, it is used as the root node of the syntax tree. Since the operator + requires two input parameters, the second and third characters are used as the left and right child nodes of the root node respectively. Then, from left to right and from top to bottom, the child nodes of the function node in the syntax tree are filled in. When all the leaf nodes in the tree are parameter nodes, the decoding is completed. Figure 4 In the example, the terminal characters {a, b, c} in the gene tail segment are not represented in the syntax tree. This article refers to them as the implicit expression region, while the remaining character regions are the explicit expression region. The lengths of the explicit and implicit expression regions are not fixed. This flexible setting allows genes of fixed length to remain viable after crossover and mutation operations. To ensure the integrity of the explicit expression region, the composition of the gene expression must meet the following two constraints:

[0124] 1) The tail segment can only contain parameter characters, and the head segment can contain function and parameter characters;

[0125] 2) If the length of the header segment is h, the length of the tail segment t must satisfy:

[0126] t≥h(n-1)+1

[0127] Where n is the maximum number of input parameters of all functions in the header segment.

[0128] by Figure 3 As shown, the function set of the gene expression is {+,*, / }, and the parameter set is {x,y,z,a,b,c}.

[0129] The header segment contains function characters {+, *, / } and parameter characters {x}, while the tail segment contains only parameter characters. The header segment length h = 4, the maximum number of input parameters for all functions n = 2, and the tail segment length t = 6 > h(n-1) + 1 = 5, satisfying the above two constraints. Because the extra tail segment length is unnecessary, in gene expression coding applications, the tail segment length t is usually directly set to h(n-1) + 1.

[0130] like Figure 6 As shown, the syntax tree nodes are divided into three layers. First, the child nodes of the third layer are connected and combined with the operators of their parent nodes to form two equations (x*x) and (y / z) respectively. Then, the tree is traversed upwards and the root node is connected with the two child node equations to form (x*x)+(y / z). The decoding is completed.

[0131] Step 4.2.2: Obtain the characteristic parameter values ​​of each path in the shortest path sequence, substitute them into the mathematical equation, and use the calculation result of the equation as the score of each path.

[0132] Step 4.2.3: Form a new path sequence according to the order of path scores from high to low;

[0133] Step 4.2.4: Prioritize the destination star path allocation based on the path with the highest score, and perform constraint checks on the allocated paths to remove infeasible paths to obtain a feasible routing solution. In this embodiment, the path with the highest score is allocated first to meet the upload requirements of various overseas data.

[0134] Step 4.2.4.1: Select the first path in the path sequence and check the constraints such as the number of node stars and transfer capacity;

[0135] Step 4.2.4.2: If any constraint is violated, delete this path and return to step 4.2.4.1; otherwise, assign this path to the destination star corresponding to the path and delete the paths in the path sequence that have the same overseas destination star as this path;

[0136] Step 4.2.4.3: If the path sequence is not empty, update the characteristic parameters of each path, re-complete the path scoring and return to step 4.2.4.1; otherwise, the path allocation is completed and the feasible routing solution for the problem is output.

[0137] Step 4.2.5: Substitute all paths in the feasible routing solution into the objective function to obtain the sum of the transmission path lengths of all destination stars, which is used as the individual fitness value.

[0138] Step 4.3: Perform selection, crossover, and mutation operations within the population.

[0139] In the genetic programming algorithm of this embodiment, the selection operator used is tournament selection, which means randomly selecting N from the parent population each time. ts Individuals enter the competition pool to compete, and the best individuals are selected to enter the evolution pool. Then, the individuals in the competition pool are put back into the parent population, and selection and competition continue until the number of individuals in the evolution pool meets the population size. In order to ensure the convergence efficiency of the algorithm, an elite retention strategy is added to the tournament selection, which selects the N individuals with the best fitness in the parent population. el Individuals skip the crossover and mutation stage and directly enter the next generation population.

[0140] This embodiment provides two crossover operators: a single-point crossover operator and a segment crossover operator.

[0141] (1) Single-point crossover operator:

[0142] Randomly select a character position in the individual gene expression and swap the characters of the two parent individuals at that position to generate two offspring individuals. The operation process is as follows: Figure 7 shown.

[0143] (2) Fragment crossover operator:

[0144] Randomly select two different breakpoints in the individual gene expression, and exchange the gene fragments between the two parent individuals' breakpoints to obtain two offspring individuals. The specific operation process is as follows: Figure 8 shown.

[0145] This embodiment provides three mutation operators, namely a single-point mutation operator, a fragment flip operator, and a fragment transfer operator.

[0146] (1) Single-point mutation operator

[0147] Select a position in the gene expression of the parent individual and define the corresponding single-point replacement rule according to different position characteristics as follows:

[0148] 1) If the position is the beginning of the header segment, a function character is randomly selected from the function set to replace the original character;

[0149] 2) If the character at that position is a constant parameter, a random number is generated to replace the original character;

[0150] 3) If the position is any other position in the header segment, randomly select a character from the function set or parameter set to replace the original character;

[0151] 4) If the position is in the tail segment, a characteristic parameter character is randomly selected from the parameter set to replace the original character.

[0152] The specific implementation of the single-point mutation operator is as follows Figure 9 shown.

[0153] (2) Fragment flip operator

[0154] Select a gene segment from the head segment of the parent individual gene (excluding the character at the start position), and flip the gene segment to generate a new offspring individual. The specific operation is as follows Figure 10 shown.

[0155] (3) Fragment transfer operator

[0156] Select a gene fragment from the gene expression of the parent individual. If the starting position of the gene fragment is a function character, randomly select a position in the gene expression and move the gene fragment to that position; if the starting position of the fragment is not a function character, randomly select a position other than the starting position of the head and move the gene fragment to that position to generate a new individual. The specific operation method is as follows Figure 11 shown.

[0157] Step 4.4: Generate a new generation of population;

[0158] Step 4.5: Determine whether the termination condition is met. If so, output the optimal result. If not, return to step 4.2. In this embodiment, the optimal result is the path sequence with the highest fitness.

[0159] Step 5: Output the path planning solution obtained. Finally, output the path sequence as the optimal path planning solution.

[0160] A navigation system data injection route planning system, characterized by comprising the following modules:

[0161] Information acquisition module: used to obtain the node satellite set and destination satellite set participating in data uploading within the planning period, as well as the configuration parameter requirements of the node satellite. The destination satellite refers to the satellite to which the data is finally uploaded, and the node satellite refers to the domestic satellite that starts uploading data from the ground station.

[0162] The shortest path sequence calculation module is used to calculate the shortest path sequence from any node star to the destination star based on the node star set and the destination star set obtained by the information acquisition module;

[0163] Integer programming model construction module: used to construct an integer programming model for the ground data routing optimization problem based on the node star set and the destination star set as well as the shortest path sequence from any node star to the destination star;

[0164] Solving module: used for solving the integer programming model;

[0165] Output module: used to output the routing plan obtained by the solution module.

[0166] Experimental comparison

[0167] The HGPGE algorithm of the present invention is an unsupervised learning algorithm that can achieve rapid problem solving. To evaluate the fast-solving performance of the HGPGE algorithm, this paper used four fast-solving algorithms as comparison algorithms: three empirical knowledge-based construction heuristic algorithms: Path Length First (PLF), which prioritizes paths with the shortest path length; Destination Satellite Time Delay First (DTF), which prioritizes paths with the shortest communication delay; and Node Satellite Time Delay First (NTF), which prioritizes paths starting from nodes with the shortest communication delay. Furthermore, a neural network algorithm, the Self-Organizing Feature Map (SOM) neural network algorithm, was used. The experimental environment was designed with an Intel Core i5 1.8 GHz processor, 8 GB of memory, and the Windows 10 operating system. The implementation language of HGPGE, PLF, DTF and NTF algorithms is Python 3.7, and the SOM algorithm is implemented based on the neural network toolbox of Matlab2017a.

[0168] (1) Algorithm parameters

[0169] Based on the scale of the navigation system scenario, the parameters of the HGPGE algorithm are configured as shown in Table 1. The gene header segment length h = 10, and the maximum number of input parameters for each operator in the function set n = 2, resulting in a single gene expression length of h(n-1) + 1 = 11. The crossover and mutation operators have equal probabilities. The number of training generations for the SOM algorithm is set to 2000, and the competition layer structure is a one-dimensional structure.

[0170] Table 1 Navigation system scenario algorithm parameter configuration

[0171] Parameter name Parameter value Number of genes 5 Gene header length h 10 Population size 50 Maximum evolutionary generation 50 <![CDATA[Number of elites N el > 1 Crossover probability 0.8 Mutation probability 0.1

[0172] (2) Analysis of algorithm comparison results

[0173] Based on 100 test set examples, the HGPGE algorithm was called to train the decision rules to obtain the gene expression of the optimal individual. The mathematical equation was decoded to form the decision rule for path scoring. The test set examples were then solved using a feasible solution construction method based on path scoring, and compared with other fast solution algorithms.

[0174] Table 2 shows the solution results of the HGPGE algorithm and four comparison algorithms for 20 test set examples, and the time unit in the table is seconds (s). As can be seen from the results in Table 2, except that the solution quality of the HGPGE algorithm in Example 15 is slightly worse than that of the SOM algorithm, the solution quality of the HGPGE algorithm of the present invention in all other examples is better than or equal to that of the four comparison algorithms. Among the four comparison algorithms, the NTF algorithm performs the worst, and the other three algorithms have their own advantages and disadvantages. In terms of calculation time, the SOM algorithm is <2s, the HGPGE algorithm is <1s, and the calculation time of the other three greedy algorithms is about 0s. It can be seen that the five algorithms can all meet the rapid solution of the problem. Using the trained decision rules, the solution quality obtained by the HGPGE algorithm is better than that of the four comparison algorithms. The comparison results show the effectiveness and superiority of the HGPGE algorithm.

[0175] Table 2 Comparison results of navigation system scenario algorithms

[0176]

[0177]

[0178] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit it. Although the present invention has been described in detail with reference to the above embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the above embodiments, or replace some or all of the technical features therein with equivalents. However, these modifications or replacements do not cause the essence of the corresponding technical solutions to deviate from the scope of the technical solutions of the embodiments of the present invention.

Claims

1. A method for routing navigation system data, characterized in that: The following steps are involved: Step 1: Obtain the node satellite set and destination satellite set participating in data uploading within the planning period, as well as the configuration parameter requirements of the node satellites. The destination satellite refers to the satellite to which the data is ultimately uploaded, and the node satellite refers to the domestic satellite that starts uploading data from the ground station. Step 2: Calculate the shortest path sequence from any node star to the destination star based on the node star set and the destination star set; Step 3: Construct an integer programming model for the ground data routing optimization problem based on the node star set, the destination star set, and the shortest path sequence from any node star to the destination star; The integer programming model is: The objective function is: （1） The constraints are: （2） （3） （4） （5） （6） （7） Among them, the objective function is to minimize the average transmission path length from each destination satellite to the node satellite; 、 are two decision variables, x ij Indicates the destination star i Whether the data selects the node star j To forward, the value is 1 for selection, and the value is 0 for non-selection. y j Indicates node star j Whether it is selected, the value is 1 to indicate selection, otherwise it is 0, both are 0-1 variables, i Indicates the destination star number, j represents the node star number, m represents the number of destination stars, n represents the number of node stars; the shortest path length l ij Indicates the destination star i With node star j The shortest transmission delay between Formula 2 indicates that each target star must and can only select one node star; Formula 3 indicates that only the selected node satellite can be selected by the destination satellite for intersatellite transfer; Formula 4 limits the transfer capacity of each selected node star; c j is the transit capacity of node star j; Formula 5 represents the number constraint of node stars; P is the upper limit of the number of node stars that can be selected; Step 4: Solve the integer programming model; Step 5: Output the routing plan obtained.

2. The routing planning method according to claim 1, wherein: The method for solving the integer programming model in step 4 is a hybrid genetic programming algorithm based on gene expression.

3. The routing planning method according to claim 2, wherein: The hybrid genetic programming algorithm based on gene expression is specifically: Step 4.1: Randomly generate according to gene expression code A chromosome is formed to form an initial population. The chromosome is composed of a single or multiple genes. A gene refers to a set of linear character expressions of fixed length. The gene is divided into a head segment and a tail segment. The head segment is composed of function characters and parameter characters, and the tail segment is composed of parameter characters. The parameter characters are characteristic parameters of the path in the shortest path sequence. Step 4.2: Calculate the fitness of each individual in the population in turn; Step 4.3: Perform selection, crossover, and mutation operations within the population; Step 4.4: Generate a new generation of population; Step 4.5: Determine whether the termination condition is met. If so, output the optimal result; if not, return to step 4.

2.

4. The routing planning method according to claim 3, wherein: The method for calculating individual fitness in step 4.2 is: Step 4.2.1: Decoding the gene expression into a mathematical equation; Step 4.2.2: Obtain the characteristic parameter values ​​of each path in the shortest path sequence, substitute them into the mathematical equation, and use the calculation result of the equation as the score of each path; Step 4.2.3: Form a new path sequence according to the order of path scores; Step 4.2.4: Prioritize the paths with high scores for the destination star path allocation, perform constraint checks on the allocated paths, delete infeasible paths, and obtain a feasible routing solution; Step 4.2.5: Substitute the feasible routing scheme into the objective function to obtain the sum of the transmission path lengths of all destination stars, which is used as the individual fitness value.

5. The routing planning method according to claim 3, wherein: The function characters in step 4.1 include , the characteristic parameters include: WN: path length; WAVE: the mean path length of the destination star to all node stars related to the path; WVAR: path length variance of the path-related destination star to all node stars; WMAX: The maximum path length of the path-related destination star to all node stars; WMIN: The minimum path length of the destination star to all node stars related to the path; NAVE: the mean path length from the node star to all destination stars related to the path; NVAR: Path length variance of path-related node stars to all destination stars; NMAX: The maximum path length of the path-related node star to all destination stars; NMIN: The minimum path length of the path-related node star to all destination stars; NCAP: the remaining transit capacity of the node star associated with the path; NRLY: Whether the node star associated with the path is available.

6. The routing planning method according to claim 3, wherein: In the gene expression coding of step 4.1, if the length of the head segment is h, the length of the tail segment t must satisfy: t ≥ h(n − 1) + 1 ; where n is the maximum number of input parameters of all functions in the header segment.

7. The routing planning method according to claim 4, wherein: The method for decoding gene expression in step 4.2.1 is: Step 4.2.1.1: Decode the gene into a syntax tree. Start with the first character of the gene as the root node of the syntax tree. Then, fill in the child nodes of the function node in the syntax tree from left to right and from top to bottom. When all the leaf nodes in the tree are parameter nodes, the decoding is complete. Step 4.2.1.2, decode the syntax tree into a mathematical equation. The overall decoding order is from leaf nodes to the root node. Starting from the lower left corner of the syntax tree, use the parent node's operator to connect the child node's equations according to the operation rules to form a new equation, and then traverse all nodes in order from right to top.

8. The routing planning method according to claim 4, wherein: The method for performing constraint checking on the assigned paths in step 4.2.4 and removing infeasible paths to obtain a feasible routing solution set is: Step 4.2.1.1: Select the first path in the path sequence and check the number of node stars and transit capacity constraints; Step 4.2.1.2: If any constraint is violated, delete this path and return to step 4.2.1.1; otherwise, assign this path to the destination star corresponding to the path and delete the paths in the path sequence that have the same overseas destination star as this path; Step 4.2.1.3: If the path sequence is not empty, update the characteristic parameters of each path, re-complete the path scoring and return to step 4.2.1.1; otherwise, the path allocation is completed and the feasible routing solution for the problem is output.

9. The routing planning method according to any one of claims 1 to 8, characterized in that: In step 2, the method for calculating the shortest path sequence from any node star to the destination star based on the node star set and the destination star set is a breadth-first search algorithm.

10. A navigation system data annotation route planning system, characterized in that: Includes the following modules: Information acquisition module: used to obtain the node satellite set and destination satellite set participating in data uploading within the planning period, as well as the configuration parameter requirements of the node satellite. The destination satellite refers to the satellite to which the data is finally uploaded, and the node satellite refers to the domestic satellite that starts uploading data from the ground station. The shortest path sequence calculation module is used to calculate the shortest path sequence from any node star to the destination star based on the node star set and the destination star set obtained by the information acquisition module; Integer programming model construction module: used to construct an integer programming model for the ground data injection routing optimization problem based on the node star set and the destination star set and the shortest path sequence from any node star to the destination star; the integer programming model is: The objective function is: （1） The constraints are: （2） （3） （4） （5） （6） （7） Among them, the objective function is to minimize the average transmission path length from each destination satellite to the node satellite; 、 are two decision variables, x ij Indicates the destination star i Whether the data selects the node star j To forward, the value is 1 for selection, and the value is 0 for non-selection. y j Indicates node star j Whether it is selected, the value is 1 to indicate selection, otherwise it is 0, both are 0-1 variables, i Indicates the destination star number, j represents the node star number, m represents the number of destination stars, n represents the number of node stars; the shortest path length l ij Indicates the destination star i With node star j The shortest transmission delay between Formula 2 indicates that each target star must and can only select one node star; Formula 3 indicates that only the selected node satellite can be selected by the destination satellite for intersatellite transfer; Formula 4 limits the transfer capacity of each selected node star; c j is the transit capacity of node star j; Formula 5 represents the number constraint of node stars; P is the upper limit of the number of node stars that can be selected; Solving module: used for solving the integer programming model; Output module: used to output the routing plan obtained by the solution module.

Images