A collaborative planning method for UAV swarm ground attack missions

Abstract

This application discloses a collaborative planning method for UAV swarm ground attack missions, comprising: S1 randomly distributing the UAV swarm over a designated area and assessing the relative states of the UAVs and ground targets; S2 assigning targets to the swarm of m UAVs and n ground targets based on a mission benefit calculation method; S3 planning the flight path of each UAV to the target using the Dubins path planning method based on the situation assessment results and the target assignment scheme; and S4 replanning threatened paths, considering obstacle threats. This invention addresses the problems of assessing the situation, assigning corresponding tasks, and planning the shortest path when multiple UAVs are performing ground attack missions.

Description

Technical Field

[0001] This invention pertains to multi-UAV cooperative control technology, specifically a method for collaborative planning of UAV swarm ground missions. Background Technology

[0002] Leveraging the advantages of high speed, stealth, and low cost of unmanned aerial vehicles (UAVs), the ability to quickly and accurately execute ground missions in complex environments has become crucial for the collaborative control of multiple UAVs. The collaborative planning problem for UAV swarm ground missions refers to the process where, after receiving ground missions, UAVs at different locations in the air assign numerous ground tasks to each UAV, enabling them to avoid obstacles, plan the shortest paths to the target, and achieve optimal mission results.

[0003] Due to the limited capabilities of a single drone, multi-drone collaborative control will become the main form of drone applications in the future. These applications include military missions such as surveillance, reconnaissance, jamming, and strike; and civilian applications such as flight performances, light shows, collaborative drone operations for Earth observation, drone logistics and delivery, geological surveys, and forest fire prevention. Collaborative planning for drone swarm ground attack missions is a crucial aspect of drone swarm systems executing ground attack tasks. Efficient and safe execution of ground attack missions is paramount, and how multiple drones can perform ground attack missions remains a technological bottleneck in this field. Summary of the Invention

[0004] This patent addresses the problem of collaborative planning for UAV swarm ground missions by proposing a method that combines situational awareness, target allocation, trajectory planning, and obstacle avoidance technologies.

[0005] This method first performs a situation assessment based on the state information of the UAV swarm and the ground mission information. Based on the assessment results and the value estimate of the ground mission, it assigns mission objectives to the UAV swarm. Finally, it divides the trajectory planning process into an obstacle avoidance process and a mission execution process, and uses Dubins curves to plan the shortest safe flight trajectory for each process.

[0006] The present invention has the following three main technical features.

[0007] (1) Situation assessment algorithm for ground missions. The situational advantage of the UAV swarm over ground targets is assessed using the angle advantage index, speed advantage index, and distance advantage index. The comprehensive advantage of the UAV swarm over ground missions is obtained through weighted calculation, forming an advantage function matrix.

[0008] (2) Target allocation algorithm based on matrix method. Based on the selection rule of maximum benefit, the element that best meets the conditions is found from the advantage function matrix, and the UAV swarm corresponding to that element is paired with the ground task. Considering the principle of cooperative control, a cooperative priority ranking is performed, and the target allocation scheme that is most beneficial to the cooperative task of the UAV swarm is selected.

[0009] (3) Dubins curve-based trajectory planning technology. Based on the location and heading angle of the UAV swarm, the location and heading angle of the ground target, and the obstacle threat area, a safe flight trajectory from the starting point to the target point is found that satisfies the maneuverability constraints, the direction of travel, and the direction of velocity.

[0010] (4) Track replanning technology. When the original track passes through an obstacle area, it may threaten the safety of the UAV flight. It is necessary to replan the track with collision risk to guide the UAV to avoid the obstacle threat and ultimately safely perform the collaborative ground mission.

[0011] Invention Effects

[0012] The beneficial effects of this invention are as follows:

[0013] (1) During the situation assessment phase, the search radius and attack radius of the UAVs were fully considered. Combined with the advantages of angle, speed and distance, the overall situation of each UAV in the cluster against ground targets under the ground mission was assessed, providing a basis for target allocation.

[0014] (2) In the target allocation phase, based on the situation assessment results, the traditional matrix method is used to assign a corresponding individual task target to each aircraft in the cluster. The target allocation result can avoid redundancy in the allocation of task targets and also prevent the omission of targets. This algorithm has a fast iteration speed, accurate target allocation, and high advantage in assessment and sorting speed. It not only ensures the flexibility of the system but also ensures the stability of the system in the face of changing environments, thereby improving the efficiency of the system in processing tasks.

[0015] (3) In the trajectory planning stage, based on the Dubins curve method, obstacle avoidance is further considered to plan a safe trajectory for each UAV to avoid obstacles and achieve the shortest flight distance. In actual flight, this method can increase flight safety and reduce fuel consumption. Attached Figure Description

[0016] Figure 1 This is a diagram of the situational advantage model;

[0017] Figure 2 Schematic diagrams of three Dubins path shapes;

[0018] Figure 3 This is a schematic diagram of the Dubins route solution in the embodiment;

[0019] Figure 4 This diagram illustrates four feasible paths based on the Dubins algorithm.

[0020] Figure 5A trajectory map for collaborative planning of UAV swarm ground attack missions. Detailed Implementation

[0021] The UAV swarm ground attack mission collaborative planning method proposed in this invention includes four stages: situation assessment, target allocation, trajectory planning, and replanning. In the trajectory planning stage, the shortest path from the starting point to the mission endpoint is planned based on the UAV's situation assessment and target allocation results. The replanning stage is divided into obstacle avoidance and mission execution. In the obstacle avoidance stage, the UAV needs to avoid obstacle-threatening areas to ensure flight safety; in the mission execution stage, the UAV needs to fly to the location of the ground target to complete the ground attack mission.

[0022] 1. The situation assessment phase includes the following steps:

[0023] In scenarios where multiple drones collaborate to perform ground attack missions, the drone swarm is distributed over the base. Assuming that the drone command and control system knows the number and location of ground targets, a situation assessment is conducted before target allocation.

[0024] 1) Establish a situational advantage model for UAVs over ground targets, such as... Figure 1 As shown. Where D ij α ij This represents the distance and azimuth of the UAV to the ground target. The azimuth is the distance between the target line and the UAV's velocity vector v. i The angle between the two lines is positive when the line deviates to the right relative to the target line, and satisfies α. ij ∈[-π, π].

[0025] 2) The situational advantage of UAVs over ground targets can be assessed using the angle advantage index, velocity advantage index, and range advantage index, denoted as Sa respectively. ij ,Sv ij and Sr ij .

[0026] The azimuth advantage index primarily considers the azimuth angle of the drone and is designed as follows:

[0027]

[0028] As can be seen from equation (1), the smaller the angle between the UAV's velocity direction and the dual-target line, the greater the angular advantage. That is, the smaller the azimuth angle of the UAV to the target, the faster it can adjust its direction of motion to fly toward the target.

[0029] The speed advantage index primarily considers the speed of the drone relative to the ground target, and is designed as follows:

[0030]

[0031] In the formula, v i and vj Let be the speeds of the drone and the ground target, respectively. As can be seen from equation (2), the greater the speed of the drone relative to the ground target, the greater its speed advantage.

[0032] The range advantage index is related to the distance between the UAV and the target, the maximum range of the missiles carried by the UAV, and the maximum detection range of the airborne radar. It is designed as follows:

[0033]

[0034] In the formula, D ij r is the distance between the UAV and the ground target. a r is the maximum range of the missiles carried by the drone. s This represents the maximum detection range of the UAV's onboard radar. As can be seen from equation (3), the closer the UAV is to the ground target, the greater its range advantage.

[0035] 3) Obtain the situational advantage model. The total situational advantage S of the i-th UAV in the cluster over the j-th ground target is... ij Linear weighting of various advantage indices based on their importance:

[0036] S ij =ε1·Sa ij +ε2·Sv ij +ε3·Sr ij (4)

[0037] In the formula, Let ε1, ε2, and ε3 be the advantage evaluation weight vector, where ε1, ε2, and ε3 are the advantage index weights for angle, velocity, and distance, respectively, and satisfy the following condition:

[0038] 2. The target allocation phase includes the following steps:

[0039] Target allocation is performed for a scenario involving a cluster of m UAVs and n (n < m) ground targets.

[0040] 1) Calculate the autonomy priority and determine the advantage assessment matrix S for both sides. Through advantage assessment, the m×n (m>n) dimensional advantage assessment matrix of the UAV swarm against the ground target is obtained as follows:

[0041]

[0042] In the formula, S ji The element in row j and column i represents the size of the drone R. i ground target B j The comprehensive advantage index value. Where 1≤i≤m, 1≤j≤n.

[0043] 2) Taking into account the survival probability and value of ground targets, design the UAVs R in the cluster for the mission benefit matrix C. i ground target B j The method for calculating task rewards is as follows:

[0044]

[0045] In the formula, and Ground target B j The survival probability and value; α is the weighting coefficient. The larger α is, the more the drone tends to preserve its greatest advantage, while the smaller α is, the more the drone tends to maximize the benefits of mission completion.

[0046] 3) Calculate the benefit assessment index for each drone in the cluster. Benefit assessment represents the risk level of the mission execution strategy adopted by each drone. The benefit assessment yields the drone's R... i The return assessment index f i The definition is as follows:

[0047]

[0048] In the formula, Indicates drone R i ground target B j The maximum profit value also means that the drone swarm has adopted the most profitable task execution strategy, that is, the most conservative and safest task execution plan.

[0049] 4) Evaluation index f of the revenue of each drone in the cluster i Sort the vectors (1≤i≤m) in descending order to obtain an ordered vector f. i ';

[0050] 5) Target allocation. Based on sorted f i ', record f in sequence i 'middle element f i '(1) Corresponding UAV R k The index k is selected from the k-th column of matrix C as the closest to f. i '(1) element C jk Record line number j, and set target B. j Assigned to drone R k ;

[0051] 6) Update. After allocation, update the allocated f. i From f i Remove from '(1), thus f iThe second element in the matrix will be used in the next allocation. Furthermore, the C matrix needs to be updated. Moving the j-th row and i-th column of the C matrix out of the matrix yields the following new C matrix:

[0052]

[0053] 7) When all ground targets have been assigned to the corresponding UAV targets, the assignment is complete; otherwise, repeat steps 3 to 6.

[0054] 3. The trajectory planning phase includes the following steps:

[0055] 1) Without considering obstacles, use the Dubins path planning method to plan the flight path of the UAV to the target. Dubins paths have three types: "CSC", "CCS", or "CCC", such as... Figure 2 As shown, it consists of arcs and straight lines with fixed curvature, where C represents an arc segment and S represents a straight line segment. Further, it can be subdivided into six types: "LSL", "RSR", "LSR", "RSL", "RLR", and "LRL", where L represents a left-turning (counter-clockwise) arc, S represents a straight line, and R represents a right-turning (clockwise) arc. Figure 3 For example, the steps for solving the Dubins path are given:

[0056] a) Determine the starting and ending circles: based on the position and velocity information P of the starting and ending points. S P F v s v f Determine the starting circle C s and the endpoint circle C f The coordinates of the center of the circle O s (x cs ,y cs ), O f (x cf ,y cf );

[0057] b) Find the ingress point P of the Dubins path. N and the tangent point P X : with O f Center R f -R s Draw a circle C with radius C. t From O s Draw a tangent from point C to circle C t Let the point of tangency be T, and connect O. f And T, line O f T and circle C f The intersection point is P. N , past O s Make a line with Of A straight line parallel to T, and circle C s The intersection point is P. X ;

[0058] c) Obtain the total path and its length: Connect P X P N Obtain the straight segments of the Dubins path, and select two minor arcs to connect P. S P X and P N P F Two circular arc segments are obtained. The lengths of the three segments are added together to obtain the total path and its length.

[0059] 2) Considering obstacle threats, the threatened paths are replanned. Replanning divides the entire phase into two processes: a safe obstacle avoidance process and a task execution process. Given the known starting and ending points and their velocity directions, there are four Dubins paths, such as... Figure 4 As shown.

[0060] a) Locate the original path threatened by the obstacle, obtain the two intersection points of the original path and the threat circle, and the heading angle at this point. Connect the two intersection points with a straight line. Draw a perpendicular line through the center of the obstacle threat. The two intersection points of this perpendicular line and the threat circle are the candidate intermediate waypoints, while keeping the heading angle unchanged. Determine a new intermediate waypoint based on the principle of shortest path length.

[0061] b) For the obstacle avoidance process, the starting point is selected as the UAV's departure position, and the ending point is selected as the new intermediate waypoint. The Dubins path planning method is also used to replan this path so that the UAV can bypass the obstacle.

[0062] c) For the mission execution process, the starting point is selected as a new intermediate waypoint and the ending point is selected as the ground target location. The Dubins path planning method is also used to replan this path to complete the ground mission.

[0063] d) Replace the original path that did not take obstacles into account with the newly planned obstacle avoidance path.

[0064] To verify the effectiveness of this invention in collaborative planning for UAV swarm ground attack missions, the UAV swarm ground attack mission collaborative planning method was simulated in the Matlab environment to verify the algorithm. Experimental settings: There were 15 UAVs and 8 ground targets. The initial heading angles of the UAVs were random, and their speed was 80 m / s. The ground targets were stationary with random heading angles. The UAVs were distributed in the region with x-coordinates of [0, 15000] and y-coordinates of [0, 5000]. The targets were distributed in the region with both x-coordinates of [15000, 20000]. One static obstacle was centered at (10000, 10000), with a threat radius of 3000 m. The simulation results of the UAV swarm ground attack mission collaborative planning are as follows: Figure 5 As shown in the diagram, the thicker lines represent the replanned flight paths of the UAVs, while the thinner lines represent the initial routes planned using the Dubins method that will not cross obstacle threat areas. Experiments demonstrate that the UAV swarm ground mission collaborative planning method described in this invention can effectively achieve multi-UAV ground mission planning under obstacle threat conditions.

[0065] This invention addresses the challenges of assessing situational awareness, assigning tasks, and planning the shortest path when multiple UAVs are performing ground attack missions. The invention establishes an advantage assessment index, employs a matrix method to assign tasks starting with the UAV with the greatest advantage, generates an assignment matrix, then uses Dubins curves for trajectory planning, and finally utilizes the Dubins path planning method for obstacle avoidance based on obstacle information.

[0066] The above description is merely a specific embodiment of the present invention, but the scope of protection of the present invention is not limited thereto. Those skilled in the art can make various modifications and adjustments within the technical scope disclosed in the present invention. Therefore, the scope of protection of the present invention should be determined by the scope of the claims.

Claims

1. A collaborative planning method for UAV swarm ground attack missions, characterized in that, The collaborative planning method includes: S1 randomly distributes a swarm of drones over a designated area to assess the relative status of the drones and ground targets. S2 allocates targets to a scenario consisting of a cluster of m UAVs and n ground targets based on the task benefit calculation method; Based on the situation assessment results and target allocation scheme, S3 uses the Dubins path planning method to plan the flight path of each UAV to the target; S3 specifically refers to: Without considering obstacles, S31 divides the path into arcs with fixed curvature and straight lines, with C representing arc segments and S representing straight line segments; S32 determines the starting and ending circles: based on the position and velocity information of the starting and ending points. , , , Determine the starting circle and the final circle The coordinates of the center of the circle , ; S33 determines the entry point for the Dubins path. and cut-out point ; S34 obtains the total path and its length: connect Obtain the straight segments of the Dubins path, and connect two minor arcs. and Two circular arc segments are obtained. The lengths of the three segments are added together to obtain the total path and its length. S4 considers obstacle threats and replans threatened paths.

2. The method for collaborative planning of UAV swarm ground missions according to claim 1, characterized in that, Specifically, S1 is: S11 randomly distributes the drone swarm over a designated area, and sets the number and location information of known ground targets for the drone command and control system. S12 derives a comprehensive advantage index based on the angle advantage index, velocity advantage index, and distance advantage index, and establishes a situational advantage model of the UAV over ground targets. The [number]th [unit] in the cluster... The first drone to the ground Situational advantage model for individual task objectives Linear weighting of various advantage indices based on their importance: (1) In the formula, Let be the advantage evaluation weight vector, where , ,and These are the weights of the efficiency advantage index for angle, speed, and distance, respectively, S S S These are the performance advantage indices for angle, speed, and distance, respectively, and they satisfy... ; S13 uses the angle advantage index, speed advantage index, and distance advantage index to evaluate the situational advantage of UAVs over ground targets. The angle advantage index is: (2) In the formula, The azimuth angle of the UAV relative to the ground target is such that the smaller the angle between the UAV's velocity direction and the target line of the two aircraft, the greater the angular advantage. The speed advantage index is: (3) In the formula, and These are the speeds of the drone and the ground target, respectively. The greater the speed of the drone relative to the ground target, the greater its speed advantage. The range advantage index is related to the distance between the UAV and the target, the maximum range of the missiles carried by the UAV, and the maximum detection range of the airborne radar. It is designed as follows: (4) In the formula, The distance between the drone and the ground target. This refers to the maximum range of the missiles carried by the drone. This represents the maximum detection range of the drone's onboard radar. The closer the drone is to the ground target, the greater its range advantage.

3. The method for collaborative planning of UAV swarm ground missions according to claim 1, characterized in that, Specifically, S2 is: S21 targets A swarm of drones and Target allocation is performed in scenarios involving multiple ground targets. Calculate autonomous priorities and determine the advantage assessment matrix between the UAV and the target. Through advantage assessment, the effectiveness of the drone swarm against ground targets is obtained. The dimensional advantage assessment matrix is ​​as follows: (5) In the formula, For the first Okay, number The elements in the column, whose size represents the drone. ground targets The comprehensive advantages; S22 is the task reward matrix. Design cluster of drones ground targets The method for calculating task rewards is as follows: (6) In the formula, and Ground targets Survival probability and value; These are the weighting coefficients. The larger the value, the more likely the drone is to preserve its greatest advantage. The smaller the value, the more the drone is inclined to maximize the efficiency of mission completion; Benefit evaluation index of each UAV in the S23 computing cluster, UAV Benefit evaluation index The definition is as follows: (7) In the formula, Indicates drone Maximum benefit value against ground targets; S24 is an evaluation index of the effectiveness of each drone in the cluster. Sort in descending order. , thus obtaining ordered vectors ; S25 is based on sorting. Record in sequence medium elements Corresponding drones Serial number ,exist Matrix number Select the closest one from the list elements Record line number , target Assigned to drones ; After S26 is allocated, the allocated... from Remove from the middle, thus The second element in the allocation is used in the next allocation, and will The first in the matrix Okay, number Columns are shifted out of the matrix to obtain a new one. The matrix is ​​as follows: (8) S27 When all ground targets have been assigned to the corresponding UAV targets, the assignment is complete; otherwise, repeat steps S23 to S26.

4. The method for collaborative planning of UAV swarm ground missions according to claim 1, characterized in that, The path includes six structures, divided into six types: LSL, RSR, LSR, RSL, RLR, and LRL. L represents a left-turning arc, S represents a straight line, and R represents a right-turning arc.

5. The method for collaborative planning of UAV swarm ground missions according to claim 1, characterized in that, The replanning divides the entire phase into two processes: a safe obstacle avoidance process and a task execution process, specifically: S41 finds the original path threatened by the obstacle, obtains the two intersection points of the original path and the threat circle and the heading angle at this time, connects the two intersection points with a straight line, draws a perpendicular line through the center of the obstacle threat, and the two intersection points of this perpendicular line and the threat circle are the candidate intermediate waypoints, while keeping the heading angle unchanged, and determines a new intermediate waypoint according to the principle of the shortest path length. For the obstacle avoidance process, S42 selects the starting point as the drone's departure position and the ending point as a new intermediate waypoint, and uses the Dubins path planning method to replan this path so that the drone can bypass obstacles. For the mission execution process, S43 selects a new intermediate waypoint as the starting point and the ground target location as the ending point, and uses the Dubins path planning method to replan this path in order to complete the ground mission.

Images