Multi-UAV Fast Scheduling Method, Device, Equipment and Readable Storage Medium

Abstract

The present invention provides a method, apparatus, device and readable storage medium for rapid scheduling of multiple unmanned aerial vehicles. The method includes: Step S11, determining the number of unmanned aerial vehicles in the unmanned aerial vehicle group; Step S12, determining an optimal pairing scheme according to the real-time position information, real-time speed information of each unmanned aerial vehicle and the position information of each target point; Step S13, based on the optimal pairing scheme, generating a preliminary path for each unmanned aerial vehicle through the A* algorithm according to the real-time modeling information of the local environment and the position information of the target point, fitting the preliminary path into a B-spline curve, performing a trajectory optimization operation on the B-spline curve, and issuing the optimized trajectory to each unmanned aerial vehicle; Step S14, if there is no parked unmanned aerial vehicle at any target point, repeating Step S12 and Step S13 at a preset time interval. Through the present invention, on the basis of ensuring no collision and dynamic feasibility during the scheduling process of multiple unmanned aerial vehicles, the time required for the unmanned aerial vehicle group to execute the scheduling task is further shortened, and the work efficiency is improved.

Description

Technical Field

[0001] The present invention relates to the technical field of multi-UAV motion planning, and particularly to a multi-UAV fast scheduling method, device, equipment and readable storage medium. Background Technique

[0002] In recent years, quadrotor UAVs have received increasing attention due to their flexibility, functionality and scalability. With the development of computing, sensing and communication technologies, UAVs have shown outstanding performance in various application fields such as search, rescue and auxiliary communication. Autonomous navigation UAVs are applied to various tasks and challenges on the earth, helping human operators to achieve various difficult or risky tasks.

[0003] Limited by the payload capacity and endurance of UAVs, the types of tasks that a single UAV can complete are very limited. Multi-UAVs are more applicable and reliable in real tasks. In these tasks, in order to complete the planning or scheduling tasks, UAVs often need to fly in complex and dense obstacle groups. However, when the current multi-UAV scheduling method is used, the UAV swarm spends a long time performing the scheduling task and has low work efficiency. Summary of the Invention

[0004] The main object of the present invention is to provide a multi-UAV fast scheduling method, device, equipment and readable storage medium, aiming to solve the technical problem that the multi-UAV scheduling method in the prior art spends a long time performing the scheduling task and has low work efficiency when used.

[0005] In a first aspect, the present invention provides a multi-UAV fast scheduling method, and the multi-UAV fast scheduling method includes:

[0006] Step S11, determining the number of UAVs in the UAV swarm, where the number of UAVs is equal to the number of target points in the scheduling task;

[0007] Step S12, determining an optimal pairing scheme according to the real-time position information, real-time speed information of each UAV and the position information of each target point, where the first reference duration of the optimal pairing scheme is the smallest among all pairing schemes, and the first reference duration is the largest predicted scheduling duration in a pairing scheme, and the predicted scheduling duration is calculated according to the real-time position information, real-time speed information of the UAV and the position information of the corresponding target point, and is used to estimate the time taken for the UAV to reach the corresponding target point;

[0008] Step S13, based on the optimal pairing scheme, generating a preliminary path for each UAV through the A* algorithm according to the local environment real-time modeling information and the position information of the target point, fitting the preliminary path into a B-spline curve, performing a trajectory optimization operation on the B-spline curve, and publishing the optimized trajectory to each UAV;

[0009] In step S14, if there is no parked drone at any target point, steps S12 and S13 are repeatedly executed at a preset time interval.

[0010] Optionally, step S12 specifically includes:

[0011] Set the first variable T opt , which is used to record the first reference duration of the optimal pairing scheme;

[0012] Set the second variable T max , which is used to record the first reference duration of the current pairing scheme;

[0013] Set an array X, which is used to record the target point allocation sequence corresponding to the optimal pairing scheme;

[0014] Select a drone r from the unassigned drone swarm by drone number, where 1 ≤ i ≤ n; i

[0015] Randomly select a target point g from the unassigned target points , where p ≤ j ≤ q, and the number of unassigned target points is equal to the number of unassigned drones; j

[0016] Calculate the predicted scheduling duration t based on the real-time position information, real-time speed information of the drone r i and the position information of the target point g j ; ij

[0017] If the predicted scheduling duration t ij is greater than the second variable T max , then overwrite the original value of the second variable T with the value of the predicted scheduling duration t ij ; max

[0018] If not all drones and all target points are paired, modify the unassigned drone swarm to Modify the unassigned target points Perform the next-level recursion to assign a target point to the drone r i+1 ;

[0019] If all drones and all target points are paired, and the second variable T max is less than the first variable T opt , then overwrite the original value of the first variable T with the value of the second variable T max and overwrite the original allocation sequence of the array X with the target point allocation sequence of the current pairing scheme; opt

[0020] If the analysis of all possible pairing schemes is completed, it is determined that the pairing scheme corresponding to the array X is the optimal pairing scheme.

[0021] Optionally, before the step of covering the original value of the second variable T with the value of the predicted scheduling duration t ij is greater than the second variable T max , the following steps are further included: ij covering the original value of the second variable T with the value of the predicted scheduling duration t max :

[0022] If the predicted scheduling duration t ij is greater than the first variable T opt , the pairing possibility of the drone r i and the target point g j is abandoned: If then randomly select a target point g from the unassigned target points m to pair with the drone r i and calculate the predicted scheduling duration t im , where p ≤ m ≤ q and m ≠ j; if then return to the previous layer of recursion, modify the unassigned drone swarm to reassign a target point for the drone r i-1 .

[0023] Optionally, after the step of setting the second variable T max for recording the first reference duration of the current pairing scheme, the following steps are further included:

[0024] Set a third variable T opt_total for recording the second reference duration of the optimal pairing scheme, where the second reference duration is the sum of the predicted scheduling durations of all drones in a pairing scheme;

[0025] Set a fourth variable T total for recording the second reference duration of the current pairing scheme;

[0026] In the step of, if all drones and all target points are not paired, modify the unassigned drone swarm to modify the unassigned target points perform the next layer of recursion, and assign a target point for the drone r i+1 , the following steps are further included:

[0027] If all drones and all target points are paired, the second variable T max is equal to the first variable T opt , the fourth variable T total is less than the third variable T opt_total , then use the fourth variable Ttotal The value of opt_total covers the third variable T, and the original allocation sequence of the array X is covered by the target point allocation sequence of the current pairing scheme with the original value of

[0028] Optionally, based on the real-time position information, real-time speed information of the drone r i and the position information of the target point g j the predicted scheduling duration t ij is calculated, and the steps include:

[0029] Based on the heuristic algorithm formula, the predicted scheduling duration t ij is calculated, and the heuristic algorithm formula is:

[0030]

[0031] where L ij is the Euclidean space distance between the position of the drone r i and the target point g j , v m is the maximum drone speed, a m is the maximum drone acceleration, v ij is the velocity component of the drone r i in the direction of the target point g j , and v ij_ver is the velocity component of the drone r i perpendicular to the direction of the target point g j .

[0032] Optionally, the B-spline curve is described by p degrees, N + 1 control points {Q0,..., Q N}, M + 1 knots {u0,..., u M}, and M = N + p + 1. There is the same time interval Δt = u i+1 - u i between adjacent knots;

[0033] The representation form of the B-spline curve is:

[0034]

[0035] where C(u) is the trajectory position represented at time u, Q i is the three-dimensional coordinate where the i-th control point is located, u i is the i-th knot, and N i,p (u) is the p-th order B-spline basis function.

[0036] Optionally, the trajectory optimization operation is based on the soft constraint optimization method, and each requirement constraint of the drone trajectory is expressed as a soft constraint cost:

[0037]

[0038] Among them, λ s 、λ c 、λ f 、λ sw 、λ v are the weight parameters corresponding to the costs of each part, and J s is the trajectory smoothness cost calculated by the jerk of the UAV; J c is the collision cost calculated from the distance between the control points and the obstacles; J f is the cost of the dynamic feasibility of the trajectory; J sw is the cluster collision avoidance cost calculated by the distance between UAVs, and J v is the velocity optimization cost guided by local path points.

[0039] In a second aspect, the present invention also provides a multi-UAV fast scheduling device, and the multi-UAV fast scheduling device includes:

[0040] A determination module, which is used for step S11 to determine the number of UAVs in the UAV group, where the number of UAVs is equal to the number of target points in the scheduling task;

[0041] A pairing module, which is used for step S12 to determine the best pairing scheme according to the real-time position information, real-time speed information of each UAV and the position information of each target point, where the first reference duration of the best pairing scheme is the smallest among all pairing schemes, and the first reference duration is the maximum predicted scheduling duration in a pairing scheme, and the predicted scheduling duration is calculated according to the real-time position information, real-time speed information of the UAV and the position information of the corresponding target point, and is used to estimate the time taken for the UAV to reach the corresponding target point;

[0042] A motion planning module, which is used for step S13 to generate the preliminary paths of each UAV based on the best pairing scheme, according to the real-time modeling information of the local environment and the position information of the target points, fit the preliminary paths into B-spline curves, perform trajectory optimization operations on the B-spline curves, and issue the optimized trajectories to each UAV;

[0043] A callback module, which is used for step S14 to repeat step S12 and step S13 at a preset time interval if there is no UAV docked at any target point.

[0044] In a third aspect, the present invention further provides a multi-UAV rapid scheduling device, which includes a processor, a memory, and a multi-UAV rapid scheduling program stored on the memory and executable by the processor. When the multi-UAV rapid scheduling program is executed by the processor, the steps of the above multi-UAV rapid scheduling method are implemented.

[0045] In a fourth aspect, the present invention further provides a readable storage medium, on which a multi-UAV rapid scheduling program is stored. When the multi-UAV rapid scheduling program is executed by a processor, the steps of the above multi-UAV rapid scheduling method are implemented.

[0046] The present invention proposes a real-time target reallocation idea, adjusts the target allocation result according to the real-time state of the UAVs, and determines the optimal pairing scheme. By continuously feeding back the real-time position and speed information of the UAV swarm and making target allocation adjustments accordingly, the entire scheduling system can be kept in an optimal time state. Through the present invention, on the basis of ensuring collision-free and dynamically feasible in the multi-UAV scheduling process, the time required for the UAV swarm to execute the scheduling task is further shortened, and the work efficiency is improved. BRIEF DESCRIPTION OF THE DRAWINGS

[0047] Figure 1 It is a flowchart of the multi-UAV rapid scheduling method in an embodiment of the present invention;

[0048] Figure 2 It is a scenario diagram of the multi-UAV rapid scheduling method in an embodiment of the present invention;

[0049] Figure 3 It is a working principle diagram of the pairing module and the motion planning module in an embodiment of the present invention;

[0050] Figure 4 It is a flowchart of step S12 in an embodiment of the present invention;

[0051] Figure 5 It is a hardware structure diagram of the multi-UAV rapid scheduling device in an embodiment of the present invention.

[0052] The realization, functional features and advantages of the object of the present invention will be further described with reference to the embodiments and the accompanying drawings. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0053] It should be understood that the specific embodiments described herein are only used to explain the present invention and are not used to limit the present invention.

[0054] In a first aspect, an embodiment of the present invention provides a multi-UAV rapid scheduling method.

[0055] Figure 1The flowchart of the multi-UAV rapid scheduling method in an embodiment of the present invention is shown.

[0056] Referring to Figure 1 , in one embodiment, the multi-UAV rapid scheduling method includes the following steps:

[0057] S11. Determine the number of UAVs in the UAV group, where the number of UAVs is equal to the number of target points in the scheduling task;

[0058] In this embodiment, the scale of the UAV group is determined according to the scheduling task to be executed. The UAV group consists of multiple quadrotor UAVs. The number of UAVs is equal to the number of target points in the scheduling task, and the starting position of the UAV group is deployed. The goal of the scheduling task in this embodiment is to ensure that the UAV group is scheduled to the target points without collision and dynamically feasible, and to minimize the time consumption as much as possible.

[0059] Figure 2 The scene diagram of the multi-UAV rapid scheduling method in an embodiment of the present invention is shown.

[0060] Referring to Figure 2 , there are unknown complex obstacles in the scene. The obstacles, the UAV group, and the target points are not all shown in Figure 2 . The UAV group needs to cross the obstacle scene from the starting position to the target points. The UAV group is represented as R = {r1,..., r n}; the starting position of the UAV group is represented as S = {s1,…, s n}; the target points are represented as G = {g1,…, g n}. Among them, n is the number of UAVs. The UAV r i ∈ R corresponds to the starting position s i ∈ S, but does not correspond to the target point g i ∈ G. Since there is no constraint that the UAV group performs a specific task at a specific target point, any UAV in the UAV group can select any target point as its own planned end point, and can change the target point during the execution of the scheduling task, but it is required that when the scheduling task is completed, there is one UAV parked at each target point. Figure 2 A situation where r1, r2, and r3 are paired with g1, g2, and g3 respectively is shown.

[0061] S12. According to the real-time position information, real-time speed information of each UAV and the position information of each target point, determine the best pairing scheme, where the first reference duration of the best pairing scheme is the smallest among all pairing schemes. The first reference duration is the maximum predicted scheduling duration in a pairing scheme, and the predicted scheduling duration is calculated according to the real-time position information, real-time speed information of the UAV and the position information of the corresponding target point, and is used to estimate the time taken for the UAV to reach the corresponding target point;

[0062] In this embodiment, the position information of each target point remains unchanged during the entire execution process of the scheduling task. After obtaining the real-time position information and real-time speed information of each UAV at a certain moment, the predicted scheduling duration for any UAV to reach any target point can be obtained through a heuristic algorithm. In the scheduling task, all UAVs start at the same time, and the scheduling task is completed when the last UAV reaches the corresponding target point. The time required for the UAV swarm to execute the scheduling task is the flight time of the UAV that takes the longest time. In this step, the time for the UAV to reach the corresponding target point is estimated through the predicted scheduling duration. Define the maximum predicted scheduling duration in a pairing scheme as the first reference duration. By exhaustively listing all possible pairing schemes, calculating the first reference duration of each pairing scheme and comparing them, the best pairing scheme with the smallest first reference duration can be obtained.

[0063] S13. Based on the best pairing scheme, according to the real-time modeling information of the local environment and the position information of the target point, generate the preliminary paths of each UAV through the A* algorithm, fit the preliminary paths into B-spline curves, perform trajectory optimization operations on the B-spline curves, and publish the optimized trajectories to each UAV;

[0064] In this embodiment, after determining the best pairing scheme, each UAV corresponds to a specific target point. According to the real-time modeling information of the local environment and the position information of the target point, a preliminary path that satisfies the start-end constraints and has no collisions can be generated through the A* algorithm. After fitting the preliminary path into a simple B-spline curve, perform trajectory optimization operations on the B-spline curve to achieve optimization in multiple aspects such as smoothness, collision avoidance, clustering, and the speed of the local path points. Publish the optimized trajectories to each UAV so that each UAV can fly according to the received trajectories.

[0065] S14. If there is no UAV parked at any target point, repeat steps S12 and S13 at a preset time interval.

[0066] When performing scheduling work in a complex obstacle environment, in order to avoid collisions between UAVs and surrounding obstacles as well as between UAVs, it is necessary to continuously adjust the trajectories of the UAV swarm during the planning process, so that the UAV swarm cannot fly according to the original plan. In this embodiment, whether all target points are parked with a UAV is used as the basis for judging whether the scheduling task is completed. Before the scheduling task is completed, repeat steps S12 and S13 at a preset time interval, determine the best pairing scheme at the corresponding moment according to the real-time position information and real-time speed information of each UAV obtained in real time, and judge whether it is necessary to modify the pairing scheme to shorten the time required for the UAV swarm to execute the scheduling task.

[0067] Specifically, in this embodiment, while steps S12 and S13 are running, the actual position and speed of the UAV are changing, and the real-time position information and real-time speed information of each UAV used in the calculations in steps S12 and S13 are fixed values obtained at a certain moment and remain unchanged during the operation. Therefore, the running duration of steps S12 and S13 should be less than the acquisition time interval of the real-time position information and real-time speed information of each UAV (i.e., the aforementioned preset time interval), so as to ensure that before obtaining the real-time position information and real-time speed information of the next moment, the flight trajectory obtained based on the real-time position information and real-time speed information of the previous moment is sent to each UAV.

[0068] Figure 3 FIG. shows the working principle diagram of the pairing module and the motion planning module in an embodiment of the present invention.

[0069] Referring to Figure 3 , in one embodiment, the pairing module is used to implement step S12, and the motion planning module is used to implement step S13. After the pairing module obtains the real-time pose information (i.e., real-time position information and real-time speed information) of each UAV, it performs the operations of step S12 to determine the optimal pairing scheme and outputs the target point allocation sequence to the motion planning module. It can be understood that when the UAV swarm is arranged in ascending order of serial numbers, the pairing scheme can be determined according to the target point allocation sequence. For example, if the UAV swarm is arranged in the order of r1, r2, r3, and the target point allocation sequence output by the pairing module is g2, g1, g3, it means that the pairing scheme is r1 corresponding to g2, r2 corresponding to g1, and r3 corresponding to g3. After receiving the target point allocation sequence, the motion planning module first performs a replanning judgment. If the current target point allocation sequence is the same as the previous one, there is no need to modify the corresponding target points in the subsequent path planning. Otherwise, it needs to be modified according to the new target point allocation sequence. Then, A* pathfinding, B-spline fitting, and trajectory optimization operations are performed in sequence to obtain the optimized trajectory and send it to the corresponding UAV.

[0070] Thus, this embodiment proposes the idea of real-time target reallocation, adjusts the target allocation result according to the real-time state of the UAV, and determines the optimal pairing scheme. By continuously feeding back the real-time position and speed information of the UAV swarm and making target allocation adjustments accordingly, the entire scheduling system can be kept in the optimal time state. Through this embodiment, on the basis of ensuring no collision and dynamic feasibility during the multi-UAV scheduling process, the time required for the UAV swarm to execute the scheduling task is further shortened, and the work efficiency is improved.

[0071] In one embodiment, step S12 specifically includes:

[0072] Set the first variable T opt , which is used to record the first reference duration of the optimal pairing scheme;

[0073] Set the second variable T max , which is used to record the first reference duration of the current pairing scheme;

[0074] Set the array X, which is used to record the target point allocation sequence corresponding to the optimal pairing scheme;

[0075] Select the drone r from the unassigned drone swarm by drone serial number, where 1 ≤ i ≤ n; i Select the target point g randomly from the unassigned target points

[0076] , where p ≤ j ≤ q, and the number of unassigned target points is equal to the number of unassigned drones; j

[0077]

[0078] Calculate the predicted scheduling duration t based on the real-time position information, real-time speed information of the drone r i and the position information of the target point g j ; ij

[0078] If the predicted scheduling duration t ij is greater than the second variable T max , then overwrite the original value of the second variable T with the value of the predicted scheduling duration t ij ; max

[0079] If not all drones are paired with all target points, then modify the unassigned drone swarm to Modify the unassigned target points Perform the next-level recursion to allocate a target point for the drone r i+1 ;

[0080] If all drones and all target points are paired, and the second variable T max is less than the first variable T opt , then overwrite the original value of the first variable T with the value of the second variable T max , and overwrite the original allocation sequence of the array X with the target point allocation sequence of the current pairing scheme; opt

[0081] If all possible pairing schemes have been analyzed, then determine the pairing scheme corresponding to the array X as the optimal pairing scheme.

[0082] In this embodiment, the determination of each pairing scheme and the recording of relevant data are manifested as a recursive process. Target points are randomly allocated to each drone in ascending order of drone serial number. One drone is allocated one target point, and the number of unassigned target points is always equal to the number of unassigned drones. Calculate the drone r in the current recursive level i ​​​With the target point g j The predicted scheduling duration t ij , and compare the predicted scheduling duration t ij with the second variable T max , determine whether numerical coverage is required to ensure that the second variable T max records the maximum predicted scheduling duration in the current pairing scheme (i.e., the first reference duration). If the current recursion level is not the last level, after completing the current recursion level, enter the next recursion level. After the last recursion level is completed, all drones and all target points are paired, and a pairing scheme is determined. Compare the second variable T max with the first variable T opt , determine whether numerical coverage is required and modify the target point allocation sequence in the array X to ensure that the first variable T opt records the minimum first reference duration among all the pairing schemes determined so far, and the array X records the target point allocation sequence corresponding to the minimum first reference duration. After exhausting all possible pairing schemes, the pairing scheme corresponding to the array X is the optimal pairing scheme.

[0083] Further, in one embodiment, before the step of covering the original value of the second variable T ij with the value of the predicted scheduling duration t max if the predicted scheduling duration t ij is greater than the second variable T max , further includes:

[0084] If the predicted scheduling duration t ij is greater than the first variable T opt , then abandon the pairing possibility of the drone r i and the target point g j : If then randomly select a target point g from the unassigned target points m to pair with the drone r i and calculate the predicted scheduling duration t im , where p ≤ m ≤ q and m ≠ j; if then return to the previous recursion level, modify the unassigned drone swarm to re-allocate a target point for the drone r i-1 .

[0085] In this embodiment, compare the predicted scheduling duration t ij in the current recursion level with the first variable T opt . If the predicted scheduling duration t ij is greater than the first variable T opt , it can be understood that it includes the drone r i and the target point gj The pairing scheme of the pairing cannot be the optimal pairing scheme. For the determination process of the current pairing scheme, when the above situation occurs at the current recursive level, if there are other unassigned target points, then for the drone r i Allocate other target points. If there are no other unassigned target points, return to the previous level of recursion and re-allocate target points for the drone r i-1 The first reference duration of the pairing scheme obtained after the recursion of this embodiment is necessarily less than or equal to the first variable T opt , and there is a possibility of becoming the optimal pairing scheme. In this way, some unnecessary operations are skipped during the process of exhausting all possible pairing schemes, which helps to reduce the amount of calculation and shorten the calculation time.

[0086] Further, in one embodiment, after the step of setting the second variable T max , which is used to record the first reference duration of the current pairing scheme, the following steps are further included:

[0087] Set the third variable T opt_total , which is used to record the second reference duration of the optimal pairing scheme, where the second reference duration is the sum of the predicted scheduling durations of all drones in a pairing scheme;

[0088] Set the fourth variable T total , which is used to record the second reference duration of the current pairing scheme;

[0089] If all drones and all target points are not paired, then modify the unassigned drone group to Modify the unassigned target points Perform the next level of recursion and allocate target points for the drone r i+1 The following steps are further included after the step of:

[0090] If all drones and all target points are paired, the second variable T max is equal to the first variable T opt , the fourth variable T total is less than the third variable T opt_total , then overwrite the original value of the third variable T total with the value of the fourth variable T opt_total , and overwrite the original allocation sequence of the array X with the target point allocation sequence of the current pairing scheme.

[0091] The foregoing embodiment only considers the case where the second variable T max is less than the first variable T opt . However, in some cases, it may occur that the second variable T max is equal to the first variable T optThe case where the first reference durations of two pairing schemes are equal. If only the first reference duration is used as the sole evaluation criterion, it is impossible to compare which of the two pairing schemes is better. In this embodiment, with the first reference duration as the primary evaluation criterion, a second reference duration is further introduced as the secondary evaluation criterion. When the first reference durations of the two pairing schemes are equal, the scheme with the smaller second reference duration is selected as the best pairing scheme.

[0092] Figure 4 FIG. 4 shows a schematic flow chart of step S12 in an embodiment of the present invention.

[0093] Referring to Figure 4 , in one embodiment, after step S12 starts, the real-time pose information of each unmanned aerial vehicle obtained through channels such as a broadcast network and the position information of each target point determined in the scheduling task are input, and a pairing scheme is determined to pair all unmanned aerial vehicles with all target points one by one. The predicted scheduling duration of each unmanned aerial vehicle and the corresponding target point is calculated, and the first reference duration and the second reference duration of the current pairing scheme are further analyzed. Whether the current pairing scheme is the best pairing scheme among all the pairing schemes determined so far is evaluated according to the first reference duration and the second reference duration. After all possible pairing schemes have been analyzed, the best pairing scheme is output.

[0094] Further, in one embodiment, according to the real-time position information, real-time speed information of the unmanned aerial vehicle r i and the position information of the target point g j , the steps for calculating the predicted scheduling duration t ij include:

[0095] Based on the heuristic algorithm formula, the predicted scheduling duration t ij is calculated. The heuristic algorithm formula is:

[0096]

[0097] where L ij is the Euclidean space distance between the position of the unmanned aerial vehicle r i and the target point g j , v m is the maximum unmanned aerial vehicle speed, a m is the maximum unmanned aerial vehicle acceleration, v ij is the speed component of the unmanned aerial vehicle r i in the direction of the target point g j , and v ij_ver is the speed component of the unmanned aerial vehicle r i perpendicular to the direction of the target point g j .

[0098] In this embodiment, the time taken for the UAV to fly in a complex obstacle environment without pre-modeling is affected by multiple factors, such as the distance from the target, the obstacle density, whether it is necessary to avoid the flight trajectories of friendly UAVs, and so on. However, since other factors except the real-time position information, real-time speed information of each UAV and the position information of each target point are unpredictable, a heuristic algorithm is adopted in this embodiment to calculate the predicted scheduling duration. After determining the optimal pairing scheme with the help of the predicted scheduling duration, the flight trajectory is then planned.

[0099] Specifically, v m and a m are the upper limits of speed and acceleration restricted by hardware and can be set according to the actual situation. v ij and v ij_ver are calculated based on the real-time position information, real-time speed information of the UAV and the position information of the corresponding target point.

[0100] Specifically, the Euclidean space distance calculation formula is:

[0101]

[0102] where the spatial coordinates of the UAV r i are A = {x ri , y ri , z ri}, and the spatial coordinates of the target point g j are B = {x gj , y gj , z gj}.

[0103] Furthermore, in one embodiment, the B-spline curve is described by p degrees, N + 1 control points {Q0,..., Q N}, M + 1 knots {u0,..., u M}, and M = N + p + 1, and there is the same time interval Δt = u i+1 - u i ;

[0104] The representation form of the B-spline curve is:

[0105]

[0106] where C(u) is the trajectory position represented at time u, Q i is the three-dimensional coordinate where the i-th control point is located, u i is the i-th knot, and N i,p (u) is the p-th order B-spline basis function.

[0107] Further, in one embodiment, the trajectory optimization operation is based on a soft constraint optimization method, and each requirement constraint of the UAV trajectory is expressed as a soft constraint cost:

[0108]

[0109] Among them, λ s 、λ c 、λ f 、λ sw 、λ v are the weight parameters corresponding to each part of the cost, J S is the trajectory smoothness cost calculated using the jerk of the UAV; J c is the collision cost calculated from the distance between the control points and the obstacles; J f is the cost of the dynamic feasibility of the trajectory; J sw is the swarm collision avoidance cost calculated by the UAV spacing, J v is the speed optimization cost guided by the local path points.

[0110] In this embodiment, λ s 、λ c 、λ f 、λ sw 、λ v are formulated based on experience. Exemplarily, each weight parameter is set as: λ s = 1.0, λ c = 0.5, λ f = 0.1, λ sw = 0.5, λ v = 0.3. It should be noted that when the planned trajectory has a collision with an obstacle or there is a collision risk between UAVs, resulting in the failure of the plan, the values of λ c and λ sw should be increased to avoid risks. The output of the trajectory optimization is the B-spline curve trajectory represented by variables such as the optimized control points Q = {Q0,..., Q N}.

[0111] In the second aspect, the embodiment of the present invention also provides a multi-UAV fast scheduling device.

[0112] In one embodiment, the multi-UAV fast scheduling device includes:

[0113] A determination module, used in step S11 to determine the number of UAVs in the UAV group, where the number of UAVs is equal to the number of target points in the scheduling task;

[0114] The pairing module, for step S12, determines the optimal pairing scheme according to the real-time position information, real-time speed information of each unmanned aerial vehicle (UAV) and the position information of each target point. Among them, the first reference duration of the optimal pairing scheme is the smallest among all pairing schemes, and the first reference duration is the maximum predicted scheduling duration in a pairing scheme. The predicted scheduling duration is calculated according to the real-time position information, real-time speed information of the UAV and the position information of the corresponding target point, and is used to estimate the time taken for the UAV to reach the corresponding target point.

[0115] The motion planning module, for step S13, based on the optimal pairing scheme, according to the real-time local environment modeling information and the position information of the target point, generates the preliminary paths of each UAV through the A* algorithm, fits the preliminary paths into B-spline curves, performs trajectory optimization operations on the B-spline curves, and publishes the optimized trajectories to each UAV.

[0116] The callback module, for step S14, if there is no UAV docked at any target point, repeats step S12 and step S13 at a preset time interval.

[0117] Further, in one embodiment, the pairing module is used for:

[0118] Set the first variable T opt , which is used to record the first reference duration of the optimal pairing scheme;

[0119] Set the second variable T max , which is used to record the first reference duration of the current pairing scheme;

[0120] Set an array X, which is used to record the target point allocation sequence corresponding to the optimal pairing scheme;

[0121] Select the UAV r from the group of unassigned UAVs according to the UAV number, where 1 ≤ i ≤ n; i Select the target point g randomly from the unassigned target points

[0122] where p ≤ j ≤ q, and the number of unassigned target points is equal to the number of unassigned UAVs; j Calculate the predicted scheduling duration t according to the real-time position information of the UAV r

[0123]

[0124] i , real-time speed information and the position information of the target point g j ; ij

[0124] If the predicted scheduling duration t ij is greater than the second variable T max , then overwrite the second variable T with the value of the predicted scheduling duration t ij ; maxThe original value;

[0125] If not all drones are paired with all target points, modify the unassigned drone swarm to Modify the unassigned target points Perform the next level of recursion for drone r i+1 Assign a target point;

[0126] If all drones and all target points are paired, and the second variable T max is less than the first variable T opt , then overwrite the original value of the first variable t max with the value of the second variable T opt , and overwrite the original assignment sequence of array X with the target point assignment sequence of the current pairing scheme;

[0127] If all possible pairing schemes have been analyzed, determine the pairing scheme corresponding to array X as the optimal pairing scheme.

[0128] Furthermore, in one embodiment, the pairing module is further configured to:

[0129] If the predicted scheduling duration t ij is greater than the first variable T opt , then abandon the pairing possibility of drone r i and target point g j : If then randomly select a target point g from the unassigned target points m to pair with drone r i , and calculate the predicted scheduling duration t im , where p ≤ m ≤ q and m ≠ j; if then return to the previous level of recursion, modify the unassigned drone swarm to reassign a target point for drone r i-1 .

[0130] Furthermore, in one embodiment, the pairing module is further configured to:

[0131] Set a third variable T opt_total to record the second reference duration of the optimal pairing scheme, where the second reference duration is the sum of the predicted scheduling durations of all drones in a pairing scheme;

[0132] Set a fourth variable T total to record the second reference duration of the current pairing scheme;

[0133] In the case where not all drones are paired with all target points, modify the unassigned drone swarm to Modify the unassigned target points Perform the next-level recursion for the drone r i+1 After the step of allocating target points to the drones, it further includes:

[0134] If all drones and all target points have completed pairing, the second variable T max is equal to the first variable T opt , and the fourth variable T total is less than the third variable T opt_total , then the value of the fourth variable T total overwrites the original value of the third variable T opt_total , and the target point allocation sequence of the current pairing scheme overwrites the original allocation sequence of the array X.

[0135] Furthermore, in one embodiment, the pairing module is used for:

[0136] Based on the heuristic algorithm formula, calculate the predicted scheduling duration t ij , and the heuristic algorithm formula is:

[0137]

[0138] where L ij is the Euclidean space distance between the position of the drone r i and the target point g j , v m is the maximum drone speed, a m is the maximum drone acceleration, v ij is the velocity component of the drone r i in the direction of the target point g j , v ij_ver is the velocity component of the drone r i perpendicular to the direction of the target point g j .

[0139] Among them, the function implementation of each module in the above multi-drone fast scheduling device corresponds to each step in the above multi-drone fast scheduling method embodiment, and its function and implementation process will not be elaborated here one by one.

[0140] In a third aspect, an embodiment of the present invention provides a multi-drone fast scheduling device, which can be a device with data processing functions such as a personal computer (PC), a laptop, a server, etc.

[0141] Figure 5 Shows the hardware structure schematic diagram of the multi-drone fast scheduling device in an embodiment of the present invention.

[0142] Refer to Figure 5, in the embodiments of the present invention, the multi-UAV rapid scheduling device may include a processor 1001 (such as a Central Processing Unit, CPU), a communication bus 1002, a user interface 1003, a network interface 1004, and a memory 1005. Among them, the communication bus 1002 is used to realize the connection and communication between these components; the user interface 1003 may include a display screen (Display) and an input unit such as a keyboard (Keyboard); the network interface 1004 may optionally include a standard wired interface and a wireless interface (such as a Wireless-FIdelity, WI-FI interface); the memory 1005 may be a high-speed random access memory (random access memory, RAM), or a stable memory (non-volatile memory), such as a disk memory. Optionally, the memory 1005 may also be a storage device independent of the aforementioned processor 1001. Those skilled in the art can understand that Figure 5 the hardware structure shown in

[0143] does not constitute a limitation to the present invention, and may include more or fewer components than shown in the figure, or combine some components, or have different component arrangements. Figure 5 , Figure 5 in the memory 1005, as a computer storage medium, may include an operating system, a network communication module, a user interface module, and a multi-UAV rapid scheduling program. Among them, the processor 1001 may call the multi-UAV rapid scheduling program stored in the memory 1005 and execute the multi-UAV rapid scheduling method provided by the embodiments of the present invention.

[0144] In a fourth aspect, the embodiments of the present invention further provide a readable storage medium.

[0145] The multi-UAV rapid scheduling program is stored on the readable storage medium of the present invention. When the multi-UAV rapid scheduling program is executed by a processor, the steps of the multi-UAV rapid scheduling method as described above are implemented.

[0146] Among them, the method implemented when the multi-UAV rapid scheduling program is executed may refer to the various embodiments of the multi-UAV rapid scheduling method of the present invention, which will not be elaborated here.

[0147] It should be noted that in this text, the term "including", "comprising" or any other variant thereof is intended to cover non-exclusive inclusion, so that a process, method, article or system including a series of elements not only includes those elements, but also includes other elements not explicitly listed, or further includes elements inherent to such process, method, article or system. Without further limitation, an element defined by the statement "including one..." does not exclude the existence of additional identical elements in the process, method, article or system including that element.

[0148] The serial numbers of the embodiments of the present invention above are only for description and do not represent the superiority or inferiority of the embodiments.

[0149] Through the description of the above embodiments, those skilled in the art can clearly understand that the above embodiment methods can be implemented by means of software plus a necessary general hardware platform. Of course, they can also be implemented by hardware, but in many cases the former is a better implementation method. Based on such an understanding, the technical solution of the present invention, in essence, or the part that contributes to the prior art, can be embodied in the form of a software product. This computer software product is stored in a storage medium (such as ROM / RAM, magnetic disk, optical disk) as described above and includes several instructions to enable a terminal device to execute the methods described in various embodiments of the present invention.

[0150] The above are only the preferred embodiments of the present invention and do not limit the patent scope of the present invention accordingly. Any equivalent structural or equivalent process transformation made by using the description of the present invention specification and the drawings, or directly or indirectly applied to other related technical fields, shall be equally included in the patent protection scope of the present invention.

Claims

1. A multi-UAV rapid scheduling method, characterized in that, The multi-UAV rapid scheduling method includes: Step S11: Determine the number of UAVs in the UAV group, where the number of UAVs is equal to the number of target points in the scheduling task; Step S12: Determine the optimal pairing scheme according to the real-time position information, real-time speed information of each UAV and the position information of each target point. Among them, the first reference duration of the optimal pairing scheme is the smallest among all pairing schemes. The first reference duration is the maximum predicted scheduling duration in a pairing scheme, and the predicted scheduling duration is calculated according to the real-time position information, real-time speed information of the UAV and the position information of the corresponding target point, and is used to estimate the time taken for the UAV to reach the corresponding target point; Step S13: Based on the optimal pairing scheme, generate the preliminary paths of each UAV through the A* algorithm according to the real-time local environment modeling information and the position information of the target points, fit the preliminary paths into B-spline curves, perform trajectory optimization operations on the B-spline curves, and publish the optimized trajectories to each UAV; Step S14: If there is no UAV docked at any target point, repeat Step S12 and Step S13 at a preset time interval; The trajectory optimization operation is based on the soft constraint optimization method, and represents each requirement constraint of the UAV trajectory as a soft constraint cost: Among them, λ s , λ c , λ f , λ sw , λ v are the weight parameters corresponding to the costs of each part, J S is the trajectory smoothness cost calculated using the jerk of the UAV; J c is the collision cost calculated from the distance between the control points and the obstacles; J f is the cost of the dynamic feasibility of the trajectory; J sw is the swarm collision avoidance cost calculated by the UAV spacing, J v is the speed optimization cost guided by local path points.

2. The multi-UAV fast scheduling method according to claim 1, wherein Step S12 specifically includes: Set the first variable T opt , which is used to record the first reference duration of the optimal pairing scheme; Set a second variable T max , which is used to record the first reference duration of the current pairing scheme; Set an array X to record the target point allocation sequence corresponding to the optimal pairing scheme; Never arrange a drone swarm Select drone r by drone serial number in i , where 1 ≤ i ≤ n; No target point has been arranged Randomly select the target point g j , where p ≤ j ≤ q, and the number of unarranged target points is equal to the number of unarranged UAVs; According to the real-time position information and real-time speed information of the drone r i and the position information of the target point g j calculate the predicted scheduling duration t ij ; If the predicted scheduling duration t ij is greater than the second variable T max , then the value of the predicted scheduling duration t ij is used to overwrite the original value of the second variable T max ; If not all drones are paired with all target points, modify the unassigned drone swarm to Modify the unassigned target points Perform the next level of recursion to assign a target point to drone r i+1 ; If all the drones and all the target points are paired, and the second variable T max is less than the first variable T opt , then the value of the second variable T max overwrites the original value of the first variable T opt , and the target point assignment sequence of the current pairing scheme overwrites the original assignment sequence of array X; If the analysis of all possible pairing schemes is completed, determine the pairing scheme corresponding to the array X as the optimal pairing scheme.

3. The multi-UAV rapid scheduling method according to claim 2, wherein, If the predicted scheduling duration t ij is greater than the second variable T max , then before the step of overwriting the original value of the second variable T with the value of the predicted scheduling duration t ij , it further includes: max ​ If the predicted scheduling duration t ij is greater than the first variable T opt , then abandon the pairing possibility of the drone r i and the target point g j : If then randomly select a target point g again from the unassigned target points and pair it with the drone r m , and calculate the predicted scheduling duration t i , where p ≤ m ≤ q and m ≠ j; if im then return to the previous level of recursion, modify the unassigned drone swarm to reassign a target point for the drone r i-1 .​ 4. The multi-UAV fast scheduling method according to claim 2, characterized in that After the step of setting the second variable T max , which is used to record the first reference duration of the current pairing scheme, the following steps are further included: Set a third variable T opt_total , which is used to record the second reference duration of the optimal pairing scheme, where the second reference duration is the sum of the predicted scheduling durations of all drones in a pairing scheme; Set the fourth variable T total , which is used to record the second reference duration of the current pairing scheme; If not all the drones are paired with all the target points, modify the unassigned drone swarm to be Modify the unassigned target points Perform the next-level recursion for drone r i+1 After the step of allocating target points to the drones, it further includes: If all the drones and all the target points are paired, the second variable T max is equal to the first variable T opt , the fourth variable T total is less than the third variable T opt_total , then the value of the fourth variable T total is used to overwrite the original value of the third variable T opt_total , and the target point allocation sequence of the current pairing scheme is used to overwrite the original allocation sequence of the array X.

5. The multi-UAV fast scheduling method according to claim 2, characterized in that The predicted scheduling duration t calculated based on the real-time position information and real-time speed information of the drone r i and the position information of the target point g j comprises the following steps: ij The steps of Based on the heuristic algorithm formula, the predicted scheduling duration t is calculated ij , and the heuristic algorithm formula is as follows: Among them, L ij is the Euclidean space distance between the position of the drone r i and the target point g j , v m is the maximum drone speed, a m is the maximum drone acceleration, v ij is the velocity component of the drone r i in the direction of the target point g j , and v ij_ver is the velocity component of the drone r i perpendicular to the direction of the target point g j .

6. The multi-UAV rapid scheduling method according to any one of claims 1 to 5, characterized in that The B-spline curve is described by p-degree, N+1 control points {Q0,...,Q N}, M+1 knots {u0,...,u M}, and M = N + p + 1. There is an equal time interval Δt = u i+1 -u i ; The representation form of the B-spline curve is: Among them, C(u) represents the trajectory position at time u, and Q i is the three-dimensional coordinate where the i-th control point is located, and u i is the i-th node, and N i,p (u) is the p-th B-spline basis function.

7. A multi-UAV rapid scheduling device, characterized in that, The multi-UAV rapid scheduling device includes: A determination module, used for Step S11 to determine the number of UAVs in the UAV group, where the number of UAVs is equal to the number of target points in the scheduling task; A pairing module, used for Step S12 to determine the optimal pairing scheme according to the real-time position information, real-time speed information of each UAV and the position information of each target point. Among them, the first reference duration of the optimal pairing scheme is the smallest among all pairing schemes. The first reference duration is the maximum predicted scheduling duration in a pairing scheme, and the predicted scheduling duration is calculated according to the real-time position information, real-time speed information of the UAV and the position information of the corresponding target point, and is used to estimate the time taken for the UAV to reach the corresponding target point; A motion planning module, used for Step S13 to generate the preliminary paths of each UAV through the A* algorithm based on the optimal pairing scheme according to the real-time local environment modeling information and the position information of the target points, fit the preliminary paths into B-spline curves, perform trajectory optimization operations on the B-spline curves, and publish the optimized trajectories to each UAV; A callback module, used for Step S14 to repeat Step S12 and Step S13 at a preset time interval if there is no UAV docked at any target point; The trajectory optimization operation is based on the soft constraint optimization method, and represents each requirement constraint of the UAV trajectory as a soft constraint cost: where, λ s , λ c , λ f , λ sw , λ v are the weight parameters corresponding to the costs of each part, J S is the trajectory smoothness cost calculated using the jerk of the UAV; J c is the collision cost calculated from the distance of the control point from the obstacle; J f is the cost of the dynamic feasibility of the trajectory; J sw is the swarm collision avoidance cost calculated by the UAV spacing, J v is the velocity optimization cost guided by local path points.

8. A multi-UAV rapid scheduling device, characterized in that, The multi-UAV rapid scheduling device includes a processor, a memory, and a multi-UAV rapid scheduling program stored on the memory and executable by the processor. When the multi-UAV rapid scheduling program is executed by the processor, the steps of the multi-UAV rapid scheduling method according to any one of claims 1 to 6 are implemented.

9. A readable storage medium, characterized in that, A multi-UAV rapid scheduling program is stored on the readable storage medium. When the multi-UAV rapid scheduling program is executed by a processor, the steps of the multi-UAV rapid scheduling method according to any one of claims 1 to 6 are implemented.

Images