A path planning method and device based on graph modeling

Abstract

The application discloses a path planning method and device based on graph modeling, and the method comprises the following steps: acquiring flight task information, three-dimensional space map information and flight cluster information set; the flight task information comprises a starting point and a terminal point of a flight task; the three-dimensional space map information comprises position coordinates, sizes and dangerous area information of all obstacles in a flight space; the flight cluster information set comprises positions, speeds and attitudes of each aircraft of a flight cluster; performing data initialization processing on the flight task information, the three-dimensional space map information and the flight cluster information set to obtain an initialization data set; and performing flight path planning processing on the initialization data set to obtain flight path information of the aircraft cluster.

Description

Technical Field

[0001] This invention relates to the fields of trajectory control, industrial data processing, and strategy optimization, specifically to a path planning method and apparatus based on graph modeling. Background Technology

[0002] With the rapid iteration and development of flight swarm technology, medium-to-large-scale swarms, with their advantages of wide coverage, high operational efficiency, and strong fault tolerance, are widely used in various complex three-dimensional airspace mission scenarios such as material delivery, emergency rescue, and formation collaborative operations. Flight path planning, as a core key technology for autonomous aircraft swarm operations, directly determines the safety, collaborative stability, and mission execution efficiency of swarm flights, and is a fundamental prerequisite for ensuring that aircraft swarms can efficiently complete complex airspace missions.

[0003] Currently, most existing aircraft swarm path planning technologies are suitable for small-scale aircraft swarm operation scenarios. Traditional path planning methods often adopt a fixed topology collaborative architecture, where the collaborative communication relationships of the aircraft swarm are fixed and cannot be dynamically adjusted according to the real-time spatial position changes of the swarm. This easily leads to problems such as local swarm communication blockage and global information transmission lag, resulting in swarm collaboration breakdown and local aircraft flight disorder, making it difficult to adapt to the dynamic flight requirements of large-scale swarms.

[0004] Meanwhile, existing environmental modeling methods are relatively crude, often employing two-dimensional planar modeling or simple three-dimensional spatial modeling modes. These methods cannot accurately adapt to the obstacle distribution characteristics of complex three-dimensional airspace and lack standardized safety redundancy protection mechanisms. They also struggle to accurately distinguish between safe and dangerous airspace areas, greatly increasing the risk of aircraft flying close to obstacles or even colliding. Furthermore, traditional planning algorithms lack a cluster state coordination and synchronization mechanism during trajectory generation, resulting in poor consistency of core flight state parameters such as aircraft speed and turn angle. The accuracy of controlling the spacing between aircraft within the cluster is insufficient, frequently leading to the risk of inter-aircraft collisions.

[0005] In addition, conventional path planning algorithms have a single node sampling and path expansion method, mostly extending the path only towards the target point. They have weak environmental exploration capabilities, are prone to generating redundant and tortuous paths, have poor path optimality, and cannot balance computational load and planning accuracy. They suffer from excessive computing power consumption and low planning efficiency, making it difficult to meet the requirements of long-term, high-precision, and high-stability autonomous flight path planning for medium and large-scale aircraft clusters. This greatly limits the operational capabilities and application scope of aircraft clusters in complex scenarios. Summary of the Invention

[0006] This invention discloses an aircraft cluster path planning technology based on dynamic topology collaboration and random sampling optimization, aiming to solve the technical problems of poor global connectivity, low collaboration efficiency, insufficient trajectory obstacle avoidance accuracy and poor real-time performance in existing aircraft cluster path planning. Through the organic integration of multiple algorithm ideas, it realizes safe, efficient and collaborative path planning for medium and large-scale aircraft clusters in complex obstacle environments.

[0007] In a first aspect, this invention discloses a path planning method based on graph modeling, comprising: S1, acquire flight mission information, 3D spatial map information, and flight cluster information set; the flight mission information includes the start and end points of each aircraft in the flight cluster required by the flight mission; the 3D spatial map information includes the flight space range, the position coordinates and dimensions of all obstacles in the flight space; the flight cluster information set includes the position, speed, turn angle, and performance limits of each aircraft in the flight cluster; the performance limits include communication radius, maximum flight speed, and maximum turn angle; S2, perform data initialization processing on the flight mission information, three-dimensional spatial map information and flight cluster information set to obtain an initialization data set; S3, perform flight path planning processing on the initial data set to obtain the flight path information of the aircraft cluster.

[0008] The process of initializing the flight mission information, 3D spatial map information, and flight cluster information sets to obtain an initialized data set includes: S21, perform information representation processing on the flight mission information, three-dimensional spatial map information and flight cluster information set to obtain a set of information variables; S22, Based on the flight cluster information set, an initial adjacency matrix is ​​constructed; S23, Based on the three-dimensional spatial map information, generate a three-dimensional raster environment model; S24. Using the set of information variables, an initialization data set is constructed.

[0009] The step of generating a three-dimensional raster environment model based on the three-dimensional spatial map information includes: Based on the flight space range in the aforementioned three-dimensional spatial map information, a three-dimensional spatial boundary range variable is constructed. , This provides information on the x-axis range. This provides information on the y-axis range. This provides information on the z-axis range. The grid discretization method is used to decompose the three-dimensional spatial boundary range variables. Divided into several uniform grids; For each grid cell, obstacle occupancy is determined, and a grid state function is defined. Its expression is: in, The center coordinates of the current grid. For the first An obstacle, For obstacle safety redundancy boundaries, , , For the first The center coordinates of the obstacles The equivalent radius of the obstacle; when an obstacle exists. If the Euclidean distance from the center of the grid to the center of the obstacle is less than or equal to the equivalent radius of the obstacle a, then the grid is a danger zone and aircraft are prohibited from entering; otherwise, it is a safe zone and aircraft are allowed to fly. Utilize all The grid is used to construct a set of dangerous areas. ; A three-dimensional grid environment model is constructed using the set of hazardous areas, the grid state function, and the three-dimensional spatial boundary range variables.

[0010] The process of constructing an initial data set using the information variable set includes: Assign a unique identifier to each aircraft in the flight cluster. , This ensures that each aircraft can be individually identified and controlled; Initialize the initial position of each aircraft. initial velocity Initial steering angle Initial path length ; The initial set of state parameters for each aircraft is constructed. , ; An initial data set is constructed using the initial set of state parameters of all aircraft, the three-dimensional grid environment model, and the initial adjacency matrix.

[0011] The step of performing flight path planning processing on the initial data set to obtain the flight path information of the aircraft cluster includes: S31, Collect the current time information t0; S32, at t0+ At any given time, the current position information of each aircraft is collected, and the adjacency matrix in the initial data set is updated to obtain the updated adjacency matrix and neighbor set of each aircraft; This is the preset first time interval; S33, at t0+ At any given time, based on the adjacency matrix and the neighbor set, the state parameter set of the aircraft is synchronized in a distributed and consistent manner to obtain the synchronized state parameter set of the aircraft and the neighbor constraint parameters. This is the preset second time interval; S34, based on the aforementioned three-dimensional grid environment model, the synchronized aircraft state parameter set, and neighbor constraint parameters, generate t0+ Path length at any given time; S35, based on t0+ The path length at each moment is used to calculate the subsequent position information of each aircraft; S36, based on the subsequent position information of all aircraft, determines whether the stopping planning conditions are met. If not, updates t0 = t0 + Execute S32; if satisfied, output the flight path information of the aircraft cluster as the set of path lengths and state parameters of all aircraft at all times.

[0012] The above at t0+ At any given time, the current position information of each aircraft is collected, and the adjacency matrix in the initial data set is updated to obtain the updated adjacency matrix and neighbor set for each aircraft, including: S321, at t0+ At any given time, set the time variable t = t0 + Collect the current location information of each aircraft ; S322, calculate the Euclidean distance between any two aircraft at the current moment. ; Filter out the reachable neighbor set for each aircraft The reachable neighbor set refers to the set of neighbors whose distance from the aircraft is within the communication radius. The collection of aircraft within; S323, for each aircraft reachable neighbor set Generate a random number between (0, 1). ; S324, if The reachable neighbor set A neighboring aircraft randomly selected from the middle Remove, randomly select an aircraft with the number from those outside the reachable neighbor set. aircraft joining To form the neighbor set at time t p is a preset scaling factor; like Set the neighbor set at time t. ; S325, iterates through all aircraft in the flight cluster, and if any isolated aircraft exist... ,Right now For aircraft i, the distance to the aircraft is forced to be... The most recent f aircraft joined ; S326, Neighbor set of all aircraft at the current moment The adjacency matrix is ​​updated to obtain the updated adjacency matrix. .

[0013] The above at t0+ At any given time, based on the adjacency matrix and neighbor set, the aircraft's state parameter set is synchronized using a distributed consensus mechanism to obtain the synchronized aircraft's state parameter set and neighbor constraint parameters, including: S331, at t0+ At any given time, set the time variable t = t0 + ; S332, based on adjacency matrix The set of state parameters of the aircraft is updated; the update expressions for speed and turn angle are: , , in, For the current moment, the first The speed of the aircraft For the previous moment The speed of the aircraft; For the current moment, the first The aircraft's turning angle For the previous moment The turning angle of the aircraft, where n is the column dimension of the adjacency matrix; Representing the adjacency matrix The element in the i-th row and j-th column; S333, for each aircraft Based on neighbor set Determine neighbor constraint parameters ,in, Let b be the minimum safe distance between aircraft i and its neighboring aircraft j, and let b be the preset minimum safe distance value. The current position of neighboring aircraft j; S334 outputs the synchronized aircraft status parameters for each aircraft. Neighbor constraint parameters for each aircraft .

[0014] A second aspect of this invention discloses a path planning device based on graph modeling, the device comprising: Memory containing executable program code; A processor coupled to the memory; The processor calls the executable program code stored in the memory to execute the graph-based path planning method.

[0015] In a third aspect, the present invention discloses a computer-readable storage medium storing computer instructions, which, when invoked by a computer, are used to execute the graph-based path planning method.

[0016] In a fourth aspect of this invention, an information data processing terminal is disclosed, which is used to implement the path planning method based on graph modeling.

[0017] The beneficial effects of this invention are as follows: This invention performs comprehensive data initialization processing on flight missions, space environment, and cluster information to achieve standardized integration of multi-source heterogeneous data. It can accurately adapt to medium-to-large-scale aircraft cluster operation scenarios. While ensuring the accuracy of cluster collaborative operations, it effectively balances the computational power consumption of path planning with the overall operational efficiency, avoiding the shortcomings of traditional technologies in adapting to small-scale clusters, computational power redundancy, or insufficient accuracy.

[0018] This invention employs a uniform three-dimensional grid discretization method to complete airspace environment modeling. Combined with obstacle equivalent treatment and safety redundancy distance settings, it accurately divides airspace into safe and dangerous areas. It can completely and accurately restore the obstacle distribution of complex three-dimensional flight airspace, effectively avoiding the risk of collision between aircraft and airspace obstacles. At the same time, by performing safety verification and fine-tuning of take-off and landing points, it ensures the safety of aircraft take-off and landing from the source, solving the problems of low accuracy and lack of safety protection in traditional environmental modeling.

[0019] This invention constructs a dynamically updatable cluster collaborative topology mechanism, which dynamically filters communication neighbors based on the real-time location of aircraft, and introduces a random reconnection mechanism to optimize the cluster topology structure. It can effectively generate cross-regional information transmission shortcuts, shorten the global state information transmission latency, and set a forced connectivity mechanism for isolated aircraft to completely avoid the problems of cluster topology splitting and local collaborative failure, and significantly improve the global connectivity and collaborative stability of medium and large-scale aircraft clusters during flight. Attached Figure Description

[0020] Figure 1This is a flowchart illustrating the implementation of the method of the present invention. Detailed Implementation

[0021] To better understand the content of this invention, an embodiment is provided here.

[0022] Figure 1 This is a flowchart illustrating the implementation of the method of the present invention.

[0023] In a first aspect, this invention discloses a path planning method based on graph modeling, comprising: S1, acquire flight mission information, 3D spatial map information, and flight cluster information set; the flight mission information includes the start and end points of each aircraft in the flight cluster; the 3D spatial map information includes the flight space range, the position coordinates and dimensions of all obstacles in the flight space; the flight cluster information set includes the position, speed, turn angle, and performance limits of each aircraft in the flight cluster; the performance limits include communication radius, maximum flight speed, and maximum turn angle; S2, perform data initialization processing on the flight mission information, three-dimensional spatial map information and flight cluster information set to obtain an initialization data set; S3, perform flight path planning processing on the initial data set to obtain the flight path information of the aircraft cluster.

[0024] The process of initializing the flight mission information, 3D spatial map information, and flight cluster information sets to obtain an initialized data set includes: S21, perform information representation processing on the flight mission information, three-dimensional spatial map information and flight cluster information set to obtain a set of information variables; The set of information variables includes: Number of aircraft clusters (positive integer, ), representing the total number of aircraft participating in path planning, with a value range suitable for medium to large-scale cluster scenarios, balancing collaborative efficiency and computing cost; Aircraft cluster starting point set ,in , , , , The first The three-dimensional coordinates (unit: m) of the starting position of each aircraft are used to ensure that the initial positions of the cluster do not overlap. Aircraft cluster target point set ,in , , , , The first The three-dimensional coordinates (unit: m) of the target position of an aircraft can be set as a unique target point for each aircraft or a unified target point for the cluster, depending on the mission requirements. , indicating the first aircraft at time Real-time location coordinates (unit: m). , The maximum planning time (unit: s) is used to dynamically update the real-time position as the aircraft flies, serving as the core input for topology update and trajectory planning. In this invention, the aircraft's flight trajectory is planned for several subsequent time points based on the real-time position of the aircraft at several sampling times. , for time Cluster collaborative adjacency matrix, , Indicates the first frame and the first Establish collaborative communication relationships between aircraft. It indicates no cooperative communication relationship, the matrix dimension matches the number of aircraft, and reflects the cluster's cooperative topology in real time; , is a collection of spatial obstacles. The number of obstacles (positive integer). , , , , For the first The center coordinates of each obstacle (unit: m) The equivalent radius of the obstacle (unit: m) is used to treat irregular obstacles as spherical obstacles, simplifying collision detection calculations. , for time No. frame and the first The Euclidean distance (in meters) between aircraft is used to determine whether aircraft are reachable and whether there is a risk of collision. It is the core calculation parameter for topology update and collision detection. , for time No. The planned path length of an aircraft (unit: m) is used to measure the effectiveness of path optimization and ensure the near-optimal nature of the path; , for time No. The flight speed of an aircraft (unit: m / s). Adhere to aircraft performance constraints and avoid loss of control due to excessive speed; , for time No. The turning angle of an aircraft (unit: , This ensures that the aircraft turns smoothly and avoids attitude imbalance caused by sharp turns.

[0025] S22, based on the flight cluster information set, constructs an initial adjacency matrix, including: Obtain the initial position of the aircraft from the flight cluster information set. Set the communication radius Initial number of neighbors (Based on extensive simulations, the initial six neighbors can balance local clustering with computational complexity.) Calculate the initial Euclidean distance between any two aircraft. This is used to determine the initial reachability between aircraft, and the calculation formula is as follows: in, For the first The initial position coordinates of the aircraft. For the first The initial position coordinates of the aircraft. This indicates that the distance between the same aircraft is not calculated; the straight-line distance between any two aircraft can be calculated using this formula, which intuitively reflects the initial positional relationship; For each aircraft Filter out (i.e., within the communication radius, where communication can be established) and the closest Let each aircraft be an initial neighbor, denoted as the initial neighbor set. and For the front The nearest neighbor}; Construct an initial adjacency matrix based on the initial neighbor set. ,in If and only if (i.e., the first) frame and the first (If an aircraft is the initial neighbor, cooperative communication can be established); otherwise... ; Output the initial adjacency matrix (Define initial cooperative relationships), the initial neighbor set for each aircraft. ( (Identify the initial collaborating partners for each aircraft).

[0026] S23, Based on the three-dimensional spatial map information, generate a three-dimensional raster environment model; S24. Using the set of information variables, an initialization data set is constructed.

[0027] The step of generating a three-dimensional raster environment model based on the three-dimensional spatial map information includes: Based on the flight space range in the aforementioned three-dimensional spatial map information, a three-dimensional spatial boundary range variable is constructed. , This provides information on the x-axis range. This provides information on the y-axis range. This provides information on the z-axis range. The grid discretization method is used to decompose the three-dimensional spatial boundary range variables. Divided into A uniform grid, with the grid size chosen to balance environmental modeling accuracy and computational efficiency; For each grid cell, obstacle occupancy is determined, and a grid state function is defined. The grid state function is used to distinguish between safe and dangerous areas and is defined as follows: in, The center coordinates of the current grid (in meters). For the first An obstacle, This represents the obstacle safety redundancy boundary (where 'a' is the safety distance used to prevent aircraft from approaching the obstacle and causing a collision). , , For the first The center coordinates of the obstacles The equivalent radius of the obstacle; when an obstacle exists. A grid is considered a danger zone when the Euclidean distance from the grid center to the obstacle center is less than or equal to the obstacle's equivalent radius *a*. Aircraft are prohibited from entering; otherwise, it is a safe area. ), allowing aircraft to fly; Utilize all The grid is used to construct a set of dangerous areas. ; A three-dimensional grid environment model is constructed by using the set of hazardous areas, the grid state function, and the three-dimensional spatial boundary range variables.

[0028] The process of constructing an initial data set using the information variable set includes: Assign a unique identifier to each aircraft in the flight cluster. , This ensures that each aircraft can be individually identified and controlled; Initialize the initial position of each aircraft. initial velocity (Starting from a standstill), initial steering angle (Initial attitude level), initial path length (The path length is 0 before flight begins); at the same time, the starting point must be verified. and target point Whether it is located in a safe area, i.e. and If the conditions are not met (i.e., the starting point or target point is located in a dangerous area), the starting point / target point is fine-tuned (fine-tuning distance ≤ 2m). The fine-tuning direction should prioritize a safe area away from obstacles to ensure that the adjusted starting point and target point are both located in a safe area. The initial set of state parameters for each aircraft is constructed. ( (to clearly define the initial state of each aircraft). An initial data set is constructed using the initial set of state parameters of all aircraft, the three-dimensional grid environment model, and the initial adjacency matrix.

[0029] The initial position, initial speed, and initial turn angle of each aircraft can also be determined based on the aircraft information in the flight cluster information set.

[0030] The step of performing flight path planning processing on the initial data set to obtain the flight path information of the aircraft cluster includes: S31, Collect the current time information t0; S32, at t0+ At any given time, the current position information of each aircraft is collected, and the adjacency matrix in the initial data set is updated to obtain the updated adjacency matrix and neighbor set of each aircraft; S33, at t0+ At any given time, based on the adjacency matrix and the neighbor set, the state parameter set of the aircraft is synchronized in a distributed and consistent manner to obtain the synchronized state parameter set of the aircraft and the neighbor constraint parameters. S34, based on the aforementioned three-dimensional grid environment model, the synchronized aircraft state parameter set, and neighbor constraint parameters, generate t0+ Path length at any given time; S35, based on t0+ The path length at each moment is used to calculate the subsequent position information of each aircraft; S36, based on the subsequent position information of all aircraft, determines whether the stopping planning conditions are met. If not, updates t0 = t0 + Execute S32; if satisfied, output the flight path information of the aircraft cluster as the set of path lengths and state parameters of all aircraft at all times.

[0031] In application, .

[0032] The stopping planning condition refers to the sum of the relative distances between the subsequent position information of all aircraft and their corresponding termination points being less than a preset threshold.

[0033] The above at t0+ At any given time, the current position information of each aircraft is collected, and the adjacency matrix in the initial data set is updated to obtain the updated adjacency matrix and neighbor set for each aircraft, including: S321, at t0+ At any given time, set the time variable t = t0 + Collect the current location information of each aircraft ; S322, calculate the Euclidean distance between any two aircraft at the current moment. ; Filter out the reachable neighbor set for each aircraft The reachable neighbor set refers to the set of neighbors whose distance from the aircraft is within the communication radius. An internal collection of aircraft capable of establishing collaborative communication; S323, for each aircraft reachable neighbor set Generate a random number between (0, 1). ; S324, if The reachable neighbor set A neighboring aircraft randomly selected from the middle Remove, in non-reachable neighbors ( One aircraft with a random number is selected. aircraft joining To form the neighbor set at time t p is a preset scaling factor; like , The original set of reachable neighbors is maintained; the purpose of dynamic reconnection is to generate cross-regional shortcuts and shorten the global information transmission path. S325, iterates through all aircraft in the flight cluster, and if any isolated aircraft exist... ,Right now For aircraft i, the distance to the aircraft is forced to be... The most recent f aircraft joined This ensures global connectivity within the cluster and prevents collaborative failures caused by cluster splits; f is 1 or 2. S326, Neighbor set of all aircraft at the current moment The adjacency matrix is ​​updated to obtain the updated adjacency matrix. This ensures that the adjacency matrix reflects the current cluster collaboration relationships in real time. Euclidean distance can be calculated using the Euclidean distance formula.

[0034] The current position information of each aircraft can be collected through the aircraft's onboard GPS and inertial measurement unit (IMU). Due to the slight error in the sensor, linear interpolation is used to correct the position deviation. The correction formula is as follows: ,in For deviation correction amount ( ), to ensure the accuracy of location data; The above at t0+ At any given time, based on the adjacency matrix and neighbor set, the aircraft's state parameter set is synchronized using a distributed consensus mechanism to obtain the synchronized aircraft's state parameter set and neighbor constraint parameters, including: S331, at t0+ At any given time, set the time variable t = t0 + ; S332, based on adjacency matrix The system updates the set of aircraft state parameters to achieve coordinated consistency in the flight states of aircraft within the cluster, avoiding collisions caused by excessive differences in local speed and turn angles. The update formulas for speed and turn angle are as follows: , , in, For the current moment, the first The speed of the aircraft For the previous moment The speed of the aircraft; the core logic of this formula is: the first... The current speed of an aircraft is composed of its own speed at the previous moment, plus the weighted sum of the differences between the speeds of all cooperating neighbors and its own speed at the previous moment (the weights are elements of the adjacency matrix), thus achieving coordinated speed consistency. For the current moment, the first The aircraft's turning angle For the previous moment The turning angle of the aircraft is logically synchronized with the speed to achieve coordinated consistency of turning angles and avoid cluster chaos caused by some aircraft turning too fast and others turning too slow. n is the column dimension of the adjacency matrix. Representing the adjacency matrix The element in the i-th row and j-th column; S333, for each aircraft Based on neighbor set Determine neighbor constraint parameters ,in The minimum safe distance between aircraft i and its neighboring aircraft j. The current position of the neighboring aircraft; this constraint parameter is used for collision detection in subsequent trajectory generation to ensure that the trajectory of the current aircraft does not overlap with the position and trajectory of the neighboring aircraft; S334, output synchronized aircraft status parameters ( (used for subsequent trajectory generation) and neighbor constraint parameters for each aircraft. (For subsequent collision detection). Multiply by the speed of the previous period The product of the two, and the path length at the previous time step. Summation yields, This is obtained by collecting the aircraft's location.

[0035] This invention uses a distributed consistency and collaborative synchronization strategy to synchronize the flight speed and turn angle of all aircraft in the cluster in real time, thereby achieving dynamic balance and uniformity of the flight status of the entire aircraft cluster. This effectively avoids problems such as cluster flight disorder and inter-aircraft collisions caused by excessive differences in the flight status of individual aircraft, and significantly improves the regularity and stability of cluster formation flight.

[0036] This invention employs a hybrid sampling mechanism that combines target point priority sampling with random sampling of the entire safe area. This mechanism can accelerate the convergence of aircraft flight paths to target points, improve mission execution efficiency, and fully explore the airspace environment, avoiding the shortcomings of traditional path planning such as fixed paths and poor environmental adaptability. At the same time, through fixed step size node adjustment and multi-layer safety constraint verification, it ensures that each generated flight path meets obstacle avoidance safety requirements.

[0037] Based on the three-dimensional grid environment model, the synchronized aircraft state parameter set, and neighbor constraint parameters, t0+ is generated. The path length at any given time includes: S341, for each aircraft, with probability Directly sample the corresponding target points (Accelerate path convergence to the target point), with probability In the three-dimensional safe area ( That is, all Random sampling is performed within a region (composed of a grid) to obtain a set of candidate nodes; during sampling, the grid state function is used. Real-time determination of whether the sampling point is located in a dangerous area; if the sampling point satisfy (Dangerous areas) are immediately resampled until safe candidate nodes are obtained. The candidate node set includes several candidate nodes; S342, for each candidate node in the candidate node set, calculate the distance relative to the current aircraft position. distance Its calculation expression is: , in,[ [ represents the position coordinates of the i-th aircraft,] , , [ ] represents the position coordinates of candidate node q; S343, Determine whether the distance is less than a preset distance value. If the value is less than 0, the corresponding candidate node is determined to be a valid candidate node, and S346 is executed; if the value is greater than 0, S344 is executed. S344: Resample along the directed line segment pointing from the current aircraft position to the candidate node position to obtain new candidate nodes. , making (Ensure the sampling step size meets the setting to avoid coarse paths caused by excessively large distances between candidate nodes), calculate The position of each neighboring aircraft of the i-th aircraft distance ; S345, if any distance exists If the neighbor safety distance constraint is violated, then execute S344; otherwise, determine... If a node is a valid candidate, execute S346; S346: Using all valid candidate nodes, construct a set of valid candidate nodes.

[0038] S347 builds an independent path tree for each aircraft. The root node of the path tree is the initial position. The leaf nodes of the path tree are the valid candidate nodes for the aircraft; traverse the path tree to find the nearest generated valid candidate node. Recent For each of the k nodes (to reduce computation while ensuring effective path optimization), calculate the path length (the path length is the Euclidean distance between nodes) from each of these k nodes to the most recently generated valid candidate node. Select the node with the shortest path length as the candidate node. Update the path tree to the parent node. This ensures the path is near optimal; Update the path length of the current aircraft based on the parent node. The calculation formula is as follows: in, This represents the path length of the aircraft at the previous moment. Valid candidate nodes coordinates The coordinates of the current aircraft position are given. This formula calculates the total path length at the current moment by summing the distances from the current position to the new candidate node, which is then used for subsequent path optimization evaluation. Determine the path length of the current aircraft , for t0+ The path length at any given time.

[0039] The neighboring aircraft is an aircraft that is part of the neighbor set.

[0040] The basis of t0+ The subsequent position information of each aircraft is calculated based on the path length at a given time. Combining the aircraft's previous position and turning angle, the path length is extended in the direction of the turning angle. This allows us to obtain the aircraft's subsequent location information.

[0041] A leaf node is the terminal node in the path tree that has no lower-level extended child nodes. It is the latest flight point expanded from the path tree and represents the latest spatial location that can be reached in the current plan. Subsequent new nodes are all extended based on the existing leaf nodes.

[0042] The root node is the unique starting reference node of a single aircraft's path tree. It corresponds to the aircraft's initial position coordinates and is the origin of the entire path tree. All subsequent flight nodes extend from this node, making it the starting point for path planning.

[0043] The parent node is the superior node in the path tree that is directly connected to the new node and is located at the top level. Its function is to determine the attachment position of the new node in the path tree, thereby connecting the previous and subsequent flight points to form a continuous feasible flight path.

[0044] The flight path information of the aircraft can be constructed using the aircraft's path tree, or using the set of path lengths and state parameters of the aircraft at all times.

[0045] The method of using the path length and status parameters of all aircraft at all times as the flight path information output for the aircraft cluster is to use the path length, position information and path tree of all aircraft at all times as the flight path information output for the aircraft cluster.

[0046] This invention constructs an independent aircraft path tree structure, selects the optimal parent node to complete the iterative expansion of the path, and continuously optimizes the length of the aircraft flight path. It can effectively avoid redundant and tortuous paths, ensure the near-optimality of the flight path of a single aircraft, and greatly improve the overall flight operation efficiency of the cluster.

[0047] In all embodiments of the present invention, the variables involved in all computational expressions or mathematical functions have been dimensionlessized before computation.

[0048] In all embodiments of the present invention, the values ​​of the independent variables in the input of all computational expressions or mathematical functions meet the reasonable requirements of the input range of the computational expressions or mathematical functions, and can ensure that the computational expressions or mathematical functions can be calculated smoothly without violating physical laws or mathematical rules.

[0049] A second aspect of this invention discloses a path planning device based on graph modeling, the device comprising: Memory containing executable program code; A processor coupled to the memory; The processor calls the executable program code stored in the memory to execute the graph-based path planning method.

[0050] In a third aspect, the present invention discloses a computer-readable storage medium storing computer instructions, which, when invoked by a computer, are used to execute the graph-based path planning method.

[0051] In a fourth aspect of this invention, an information data processing terminal is disclosed, which is used to implement the path planning method based on graph modeling.

[0052] The above description is merely an embodiment of the present invention and is not intended to limit the invention. Various modifications and variations can be made to the present invention by those skilled in the art. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principle of the present invention should be included within the scope of the claims of the present invention.

Claims

1. A path planning method based on graph modeling, characterized in that, include: S1, acquire flight mission information, 3D spatial map information, and flight cluster information set; the flight mission information includes the start and end points of each aircraft in the flight cluster required by the flight mission; the 3D spatial map information includes the flight space range, the position coordinates and dimensions of all obstacles in the flight space; the flight cluster information set includes the position, speed, turn angle, and performance limits of each aircraft in the flight cluster; the performance limits include communication radius, maximum flight speed, and maximum turn angle; S2, perform data initialization processing on the flight mission information, three-dimensional spatial map information and flight cluster information set to obtain an initialization data set; S3, perform flight path planning processing on the initial data set to obtain the flight path information of the aircraft cluster.

2. The path planning method based on graph modeling as described in claim 1, characterized in that, The process of initializing the flight mission information, 3D spatial map information, and flight cluster information sets to obtain an initialized data set includes: S21, perform information representation processing on the flight mission information, three-dimensional spatial map information and flight cluster information set to obtain a set of information variables; S22, Based on the flight cluster information set, an initial adjacency matrix is ​​constructed; S23, Based on the three-dimensional spatial map information, generate a three-dimensional raster environment model; S24. Using the set of information variables, an initialization data set is constructed.

3. The path planning method based on graph modeling as described in claim 2, characterized in that, The step of generating a three-dimensional raster environment model based on the three-dimensional spatial map information includes: Based on the flight space range in the aforementioned three-dimensional spatial map information, a three-dimensional spatial boundary range variable is constructed. , This provides information on the x-axis range. This provides information on the y-axis range. This provides information on the z-axis range. The grid discretization method is used to decompose the three-dimensional spatial boundary range variables. Divided into several uniform grids; For each grid cell, obstacle occupancy is determined, and a grid state function is defined. Its expression is: in, The center coordinates of the current grid. For the first An obstacle, For obstacle safety redundancy boundaries, , , For the first The center coordinates of the obstacles The equivalent radius of the obstacle; when an obstacle exists. If the Euclidean distance from the center of the grid to the center of the obstacle is less than or equal to the equivalent radius of the obstacle a, then the grid is a danger zone and aircraft are prohibited from entering; otherwise, it is a safe zone and aircraft are allowed to fly. Utilize all The grid is used to construct a set of dangerous areas. ; A three-dimensional grid environment model is constructed using the set of hazardous areas, the grid state function, and the three-dimensional spatial boundary range variables.

4. The path planning method based on graph modeling as described in claim 3, characterized in that, The process of constructing an initial data set using the information variable set includes: Assign a unique identifier to each aircraft in the flight cluster. , This ensures that each aircraft can be individually identified and controlled; Initialize the initial position of each aircraft. initial velocity Initial steering angle Initial path length ; The initial set of state parameters for each aircraft is constructed. , ; An initial data set is constructed using the initial set of state parameters of all aircraft, the three-dimensional grid environment model, and the initial adjacency matrix.

5. The path planning method based on graph modeling as described in claim 4, characterized in that, The step of performing flight path planning processing on the initial data set to obtain the flight path information of the aircraft cluster includes: S31, Collect the current time information t0; S32, at t0+ At any given time, the current position information of each aircraft is collected, and the adjacency matrix in the initial data set is updated to obtain the updated adjacency matrix and neighbor set of each aircraft; This is the preset first time interval; S33, at t0+ At any given time, based on the adjacency matrix and the neighbor set, the state parameter set of the aircraft is synchronized in a distributed and consistent manner to obtain the synchronized state parameter set of the aircraft and the neighbor constraint parameters. This is the preset second time interval; S34, based on the aforementioned three-dimensional grid environment model, the synchronized aircraft state parameter set, and neighbor constraint parameters, generate t0+ Path length at any given time; S35, based on t0+ The path length at each moment is used to calculate the subsequent position information of each aircraft; S36, based on the subsequent position information of all aircraft, determines whether the stopping planning conditions are met. If not, updates t0 = t0 + Execute S32; if satisfied, output the flight path information of the aircraft cluster as the set of path lengths and state parameters of all aircraft at all times.

6. The path planning method based on graph modeling as described in claim 5, characterized in that, The above at t0+ At any given time, the current position information of each aircraft is collected, and the adjacency matrix in the initial data set is updated to obtain the updated adjacency matrix and neighbor set for each aircraft, including: S321, at t0+ At any given time, set the time variable t = t0 + Collect the current location information of each aircraft ; S322, calculate the Euclidean distance between any two aircraft at the current moment. ; Filter out the reachable neighbor set for each aircraft The reachable neighbor set refers to the set of neighbors whose distance from the aircraft is within the communication radius. The collection of aircraft within; S323, for each aircraft reachable neighbor set Generate a random number between (0, 1). ; S324, if The reachable neighbor set A neighboring aircraft randomly selected from the middle Remove, randomly select an aircraft with the number from those outside the reachable neighbor set. aircraft joining To form the neighbor set at time t p is a preset scaling factor; like Set the neighbor set at time t. ; S325, iterates through all aircraft in the flight cluster, and if any isolated aircraft exist... ,Right now For aircraft i, the distance to the aircraft is forced to be... The most recent f aircraft joined ; S326, Neighbor set of all aircraft at the current moment The adjacency matrix is ​​updated to obtain the updated adjacency matrix. .

7. The path planning method based on graph modeling as described in claim 5, characterized in that, The above at t0+ At any given time, based on the adjacency matrix and neighbor set, the aircraft's state parameter set is synchronized using a distributed consensus mechanism to obtain the synchronized aircraft's state parameter set and neighbor constraint parameters, including: S331, at t0+ At any given time, set the time variable t = t0 + ; S332, based on adjacency matrix The set of state parameters of the aircraft is updated; the update expressions for speed and turn angle are: , ， in, For the current moment, the first The speed of the aircraft For the previous moment The speed of the aircraft; For the current moment, the first The aircraft's turning angle For the previous moment The turning angle of the aircraft, where n is the column dimension of the adjacency matrix; Representing the adjacency matrix The element in the i-th row and j-th column; S333, for each aircraft Based on neighbor set Determine neighbor constraint parameters ,in, Let b be the minimum safe distance between aircraft i and its neighboring aircraft j, and let b be the preset minimum safe distance value. The current position of neighboring aircraft j; S334 outputs the synchronized aircraft status parameters for each aircraft. Neighbor constraint parameters for each aircraft .

8. A path planning device based on graph modeling, characterized in that, The device includes: Memory containing executable program code; A processor coupled to the memory; The processor calls the executable program code stored in the memory to execute the path planning method based on graph modeling as described in any one of claims 1 to 7.

9. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores computer instructions, which, when invoked by a computer, are used to execute the path planning method based on graph modeling as described in any one of claims 1 to 7.

10. An information data processing terminal, characterized in that, The information data processing terminal is used to implement the path planning method based on graph modeling as described in any one of claims 1 to 7.

Images