Distributed UAV swarm autonomous formation flight method, device and storage medium

Abstract

The present invention provides a distributed unmanned aerial vehicle cluster autonomous formation flight method, device and storage medium, which relates to unmanned aerial vehicle navigation, including: constructing a three-dimensional map; obtaining an initial path through path search; extracting a safe flight corridor; obtaining trajectory information of other unmanned aerial vehicles, and calculating the optimal formation-keeping trajectory of the unmanned aerial vehicle; calculating the actual flight trajectory of the unmanned aerial vehicle according to the optimal formation-keeping trajectory and airborne sensor data; executing the actual flight trajectory and publishing the actual flight trajectory to a wireless broadcast network for other unmanned aerial vehicles to obtain. The distributed unmanned aerial vehicle cluster autonomous formation flight method proposed by the present invention has formation-keeping awareness, generates a theoretical optimal formation trajectory that considers the scaling of the formation scale, and guides the optimization of the actual flight trajectory of the unmanned aerial vehicle based on this, thereby improving the formation-keeping effect of the unmanned aerial vehicle cluster, and enabling it to quickly plan a safe, efficient and actively formation-keeping flight trajectory in a complex and dense obstacle environment.

Description

Technical Field

[0001] The present invention relates to the field of unmanned aerial vehicle navigation, and in particular to a distributed unmanned aerial vehicle cluster autonomous formation flight method, device and storage medium. Background Art

[0002] Multi-rotor drones are often used in environmental inspections, low-altitude patrols and other tasks to assist in completing a series of complex tasks. Through the coordination between multiple drones, drone swarms can help people complete more complex tasks. So far, many drone formation motion planning methods have been proposed and put into practical use, but the flexibility and environmental adaptability of existing methods are still low, making it difficult to directly apply them to multi-drone formation flight missions in complex environments. In the process of implementing the present application, the applicant found that the existing multi-UAV formation flight planning methods can be mainly divided into pilot-follower, virtual structure and behavior-based methods: the pilot-follower method designates one or several aircraft in the formation as the pilot responsible for trajectory planning, and other UAVs as followers follow the pilot at a certain control rate. Although this method has strong scalability, the centralized planning framework has high requirements for the pilot, resulting in poor robustness of the system; different from the pilot-follower method, the virtual structure method regards the entire formation as a rigid virtual structure, and realizes the formation flight of UAVs by planning the trajectory for the center of the virtual structure and letting the UAVs track the corresponding points on the structure. This method still belongs to the centralized planning framework, and the rigid formation structure makes it difficult to apply to complex environments with dense obstacles; the behavior-based formation maintenance method maintains the formation by designing a variety of basic behaviors such as formation maintenance and obstacle avoidance, and maintains the formation according to the behavior coordination criterion and the basic behavior of perceptual information fusion, thereby completing the formation task with multiple competitive objectives; the disadvantage is that this distributed planning framework relies heavily on the design of the behavior coordination criterion, and it is difficult to quantitatively evaluate the stability of the formation.

[0003] Therefore, how to improve the real-time operation and formation maintenance effect of formation flying in complex environments to ensure flight efficiency and safety has become a technical problem to be solved. Summary of the invention

[0004] The present invention aims to solve at least one of the technical problems existing in the prior art or related technology, and discloses a distributed unmanned aerial vehicle cluster autonomous formation flight method, device and storage medium, which have formation maintenance awareness, and are conducive to large-scale unmanned aerial vehicle clusters to perform safe and efficient autonomous formation flight in an environment with complex structures and dense obstacles.

[0005] The first aspect of the present invention discloses a distributed unmanned aerial vehicle cluster autonomous formation flight method, comprising: obtaining onboard sensor data and flight trajectory information of multiple unmanned aerial vehicles in the unmanned aerial vehicle cluster through a wireless broadcast network to construct a three-dimensional occupancy map; given a task starting point and a task end point, performing a path search on the three-dimensional occupancy map based on an A* path search algorithm to obtain an initial path of the unmanned aerial vehicle unit; extracting a safe flight corridor: for each unmanned aerial vehicle unit, traversing the corresponding initial path in order, determining whether to add a new convex geometric body by judging whether each point in the initial path is within the last convex geometric body of the current flight corridor, and adding the newly generated convex geometric body to the flight corridor until every point on the initial path is traversed to generate a safe flight corridor. ; Obtain the trajectory information of other UAVs other than this aircraft, and calculate the theoretical optimal formation keeping trajectory of this aircraft: construct the optimal formation keeping problem as an unconstrained convex optimization problem considering the formation keeping effect and formation scale regularization, and obtain the theoretically optimal formation keeping trajectory by solving the analytical solution; According to the theoretical optimal formation keeping trajectory and the onboard sensor data, calculate the actual flight trajectory of this aircraft in real time: Taking flight smoothness, motion efficiency, and formation keeping effect as the goals, construct the trajectory planning problem as a nonlinear optimization problem that meets flight safety, multi-aircraft collision avoidance, and kinematic and dynamic feasible constraints, and obtain the actual flight trajectory by solving it; Execute the actual flight trajectory, and publish the actual flight trajectory of this aircraft to the wireless broadcast network so that other UAVs can obtain the actual flight trajectory.

[0006] In this technical solution, in view of the problem that formation flight in existing complex environments is difficult to balance the real-time operation and the formation maintenance effect, the present invention proposes a distributed UAV cluster autonomous flight method with formation maintenance awareness. The proposed method belongs to a distributed planning algorithm. Each UAV only needs to transmit its own trajectory information to achieve formation maintenance during formation flight, and has low requirements for multi-machine communication load. For each UAV, the method first uses the known trajectory information of other surrounding UAVs to calculate the optimal formation maintenance trajectory without considering obstacles, that is, the theoretical optimal formation trajectory. The proposed planning algorithm first generates the optimal formation trajectory considering the scaling of the formation scale, and guides the trajectory optimization based on this, which not only improves the formation maintenance effect of the UAV cluster, but also enables it to quickly plan a safe, efficient and active formation maintenance trajectory in a complex and dense obstacle environment. The above theoretical optimal formation trajectory is used to guide the self-trajectory planning of each UAV. In addition, each UAV perceives environmental obstacles in real time according to the onboard sensor, and combines the optimal formation trajectory guidance to quickly plan a safe, efficient and active formation maintenance flight trajectory.

[0007] According to the distributed UAV swarm autonomous formation flight method disclosed in the present invention, preferably, the mission starting point is the current position of the UAV, the mission end point is the position of the mission target point or the local target point, and the initial path is a collision-free safe path.

[0008] According to the distributed UAV swarm autonomous formation flight method disclosed in the present invention, preferably, the expression of the unconstrained convex optimization problem considering the formation maintenance effect and the formation scale regularization is:

[0009]

[0010] The decision variables are and , For its own trajectory, For the formation scale with respect to time The function of is the total trajectory time, and They are respectively the objective function of formation preservation and the objective function of regular items, is the regularization term weight.

[0011] According to the distributed UAV swarm autonomous formation flight method disclosed in the present invention, preferably, the formation maintenance degree objective function is designed to minimize the formation error distance:

[0012]

[0013] in, is the number of drones in the formation, It is The trajectory of the drone, For the team The formation target of the first UAV is The relative position of the formation target of the drones.

[0014] According to the distributed UAV swarm autonomous formation flight method disclosed in the present invention, preferably, the regularization objective function is designed as a custom desired formation maintenance scale:

[0015]

[0016] For an unconstrained quadratic convex optimization problem at each moment, the optimal solution at each moment is analytically obtained to obtain the theoretically optimal formation trajectory. .

[0017] According to the distributed UAV swarm autonomous formation flight method disclosed in the present invention, preferably, the optimization problem corresponding to the actual flight trajectory is:

[0018]

[0019] in, For the smoothing target item, For time target items, Keep the target item for the formation, is the collision penalty term, is the multi-machine conflict penalty term, is the kinematic and dynamic out-of-bounds penalty term. The decision variables in the optimization problem include the trajectory of the UAV And the total trajectory time ; Solve the optimization problem by gradient descent method to obtain the Actual flight path of the drone .

[0020] According to the distributed UAV cluster autonomous formation flight method disclosed in the present invention, preferably, the smoothing target item Designed to minimize jerk:

[0021] Time target item Designed to minimize time:

[0022]

[0023] Formation keeping target item Keep the difference between the trajectory and the optimal formation:

[0024]

[0025] in, Keep track for optimal formation.

[0026] According to the distributed UAV swarm autonomous formation flight method disclosed in the present invention, preferably, the collision penalty term Used to punish those who exceed the safe flight corridor:

[0027]

[0028] in, is the number of convex geometric bodies in the flight corridor, For the The number of faces of a convex body, and For the The first convex geometry Parameters of each face;

[0029] Multi-machine conflict penalty Used to punish drones that do not meet the safety distance:

[0030]

[0031] in, is the minimum safe distance between drones;

[0032] Kinematic and dynamic out-of-bounds penalties Used to penalize those that exceed the maximum speed or acceleration:

[0033]

[0034] in, is the speed curve, is the acceleration curve, and are the maximum speed and maximum acceleration respectively.

[0035] The second aspect of the present invention discloses a distributed unmanned aerial vehicle cluster autonomous formation flight device, including: a memory for storing program instructions; a processor for calling the program instructions stored in the memory to implement a distributed unmanned aerial vehicle cluster autonomous formation flight method such as any of the above-mentioned technical solutions.

[0036] A third aspect of the present invention discloses a computer-readable storage medium storing a program code, wherein the program code is used to implement a distributed unmanned aerial vehicle cluster autonomous formation flight method as described in any of the above technical solutions.

[0037] The beneficial effects of the present invention include at least: the present invention proposes a distributed planning algorithm with formation maintenance awareness, and each drone only needs to transmit its own trajectory information to achieve formation maintenance during formation flight, and the communication load requirements between multiple drones are relatively low. For each drone, the present invention uses the known trajectory information of other surrounding drones to calculate the optimal formation maintenance trajectory without considering obstacles, that is, the optimal formation trajectory, and guides the trajectory planning of each drone based on the above optimal formation trajectory; in addition, each drone perceives environmental obstacles in real time according to the airborne sensor, and combines the optimal formation trajectory guidance to quickly plan a safe, efficient and active formation-maintaining flight trajectory. In general, the present invention generates an optimal formation trajectory that considers the scaling of the formation scale, and based on this guided trajectory optimization, not only improves the formation maintenance effect of the drone cluster, but also enables it to quickly plan a safe, efficient and active formation-maintaining trajectory in a complex and dense obstacle environment. It is conducive to large-scale drone clusters to perform safe and efficient autonomous formation flight in complex and dense obstacle environments such as post-disaster ruins and wild forests. BRIEF DESCRIPTION OF THE DRAWINGS

[0038] Figure 1A schematic flow chart of a distributed UAV swarm autonomous formation flight method according to an embodiment of the present invention is shown.

[0039] Figure 2 A schematic diagram of the system framework of a distributed UAV swarm autonomous formation flight method according to an embodiment of the present invention is shown.

[0040] Figure 3 A schematic diagram of the optimal formation trajectory of a distributed UAV swarm autonomous formation flight method according to an embodiment of the present invention is shown.

[0041] Figure 4 A schematic diagram of trajectory optimization based on optimal formation trajectory guidance of a distributed UAV swarm autonomous formation flight method according to an embodiment of the present invention is shown.

[0042] Figure 5 A schematic block diagram of a distributed UAV swarm autonomous formation flying device according to an embodiment of the present invention is shown. DETAILED DESCRIPTION

[0043] In order to more clearly understand the above-mentioned objects, features and advantages of the present invention, the present invention is further described in detail below with reference to the accompanying drawings and specific embodiments.

[0044] In the following description, many specific details are set forth to facilitate a full understanding of the present invention. However, the present invention may also be implemented in other ways different from those described herein. Therefore, the present invention is not limited to the specific embodiments disclosed below.

[0045] like Figure 1As shown, according to one embodiment of the present invention, the distributed UAV cluster autonomous formation flight method disclosed in the present invention includes: step S101, constructing a three-dimensional occupancy map: obtaining the airborne sensor data and flight trajectory information of multiple UAVs in the UAV cluster through a wireless broadcast network to construct a three-dimensional occupancy map; step S102, searching for an initial path: given a task start point and a task end point, a path search is performed on the three-dimensional occupancy map based on an A* path search algorithm to obtain an initial path of a UAV unit, wherein the task start point is the current position of the UAV, the task end point is the position of the task target point or the local target point, and the initial path is a collision-free safe path; step S103, extracting a safe flight corridor: for each UAV unit, traverse the corresponding initial path in order, determine whether to add a new convex geometric body by judging whether each point in the initial path is within the last convex geometric body of the current flight corridor, and add the newly generated The convex geometric body is added to the flight corridor until every point on the initial path is traversed to generate a safe flight corridor; step S104, obtaining the trajectory information of UAVs other than the aircraft itself, and calculating the theoretical optimal formation keeping trajectory of the aircraft: the optimal formation keeping problem is constructed as an unconstrained convex optimization problem considering the formation keeping effect and the formation scale regularization, and the optimal formation keeping trajectory is obtained by solving the analytical solution; step S105, according to the optimal formation keeping trajectory and the onboard sensor data, the actual flight trajectory of the aircraft is calculated in real time: with flight smoothness, motion efficiency, and formation keeping effect as the goals, the trajectory planning problem is constructed as a nonlinear optimization problem that satisfies flight safety, multi-aircraft collision avoidance, and kinematic and dynamic feasible constraints, and the actual flight trajectory is obtained by solving the problem; step S106, execution and broadcasting: executing the actual flight trajectory, and publishing the actual flight trajectory of the aircraft to the wireless broadcast network so that other UAVs can obtain the actual flight trajectory.

[0046] According to the above embodiment, preferably, step S104 specifically includes: the expression of the unconstrained convex optimization problem considering the formation preservation effect and the formation scale regularization is:

[0047]

[0048] The decision variables are and , For its own trajectory, For the formation scale with respect to time The function of is the total trajectory time, and They are respectively the objective function of formation preservation and the objective function of regular items, is the regularization term weight.

[0049] The formation keeping degree objective function is designed to minimize the formation error distance:

[0050]

[0051] in, is the number of drones in the formation, It is The trajectory of the drone, For the team The formation target of the first UAV is The relative position of the formation target of the drones.

[0052] The regularization objective function is designed to be a custom desired shape preserving metric:

[0053]

[0054] For an unconstrained quadratic convex optimization problem at each moment, the optimal solution at each moment is analytically obtained to obtain the theoretically optimal formation trajectory. .

[0055] According to the above embodiment, preferably, step S105 specifically includes: the optimization problem corresponding to the actual flight trajectory is:

[0056]

[0057] in, For the smoothing target item, For time target items, Keep the target item for the formation, is the collision penalty term, is the multi-machine conflict penalty term, is the kinematic and dynamic out-of-bounds penalty term. The decision variables in the optimization problem include the trajectory of the UAV And the total trajectory time ;

[0058] The optimization problem is solved by gradient descent method to obtain the Actual flight path of the drone . Specifically:

[0059] Smooth Target Item Designed to minimize jerk:

[0060] Time target item Designed to minimize time:

[0061]

[0062] Formation keeping target item Keep the difference between the trajectory and the optimal formation:

[0063]

[0064] in, Keep track for optimal formation.

[0065] Collision Penalty Used to punish those who exceed the safe flight corridor:

[0066]

[0067] in, is the number of convex geometric bodies in the flight corridor, For the The number of faces of a convex geometry, and For the The first convex geometry Parameters of each face;

[0068] Multi-machine conflict penalty Used to punish drones that do not meet the safety distance:

[0069]

[0070] in, is the minimum safe distance between drones;

[0071] Kinematic and dynamic out-of-bounds penalties Used to penalize those that exceed the maximum speed or acceleration:

[0072]

[0073] in, is the speed curve, is the acceleration curve, and are the maximum speed and maximum acceleration respectively.

[0074] like Figure 2 As shown, according to another embodiment of the present invention, a specific implementation method of the distributed UAV cluster autonomous formation flight method of the above embodiment is also disclosed. Figure 2 The framework diagram of the distributed UAV swarm autonomous flight system is shown. The figure shows that multiple UAVs (UAV 0, UAV 1, and other UAVs) fly in formation together. They communicate in real time through wireless networks. Each UAV unit runs the autonomous flight method shown in the figure. The specific methods include:

[0075] The first step is to use wireless broadcast networks and drone-mounted sensors. Each drone will obtain the trajectory information of other drones and the environmental information perceived by itself: The airborne sensors (such as depth cameras, lidar, etc.) on the drones obtain information about the surrounding environment and build a three-dimensional map based on it. Then, the position of obstacles in the world coordinate system is obtained according to the drone's position, and other drone obstacles are deleted using the trajectory information of other drones to build a three-dimensional occupancy map for autonomous navigation.

[0076] In the second step, given the starting and ending points of the task, the A* path search algorithm is run on the three-dimensional occupancy map to obtain the first The initial path of the drone: Set the starting position of the mission on the three-dimensional occupancy map, where the starting position For the The current position of the drone and the final position is the location of the mission target point or local target point. Then, perform A* path search on the occupied map to obtain a collision-free safe path .

[0077] The third step is to follow Extracting safe flight corridors from the initial path of a UAV: ​​Traversing safe paths in order Every point , by judging Is it in the current flight corridor? The last convex geometry of To decide whether to add new convex geometry, if Not Available Internally, Generate a convex geometry in a safe space for the center , and added to the flight corridor, i.e. , until the path is traversed Every point on.

[0078] The fourth step is Figure 3 As shown, based on the trajectory information of other drones around, the first The theoretical optimal formation-keeping trajectory of a UAV is the optimal formation trajectory: without considering obstacles, the optimal formation-keeping problem is constructed as an unconstrained convex optimization problem that considers the formation-keeping effect and the regularization of the formation scale, and the theoretically optimal optimal formation trajectory is obtained by solving its analytical solution. .

[0079] The above formation preservation problem is constructed as an unconstrained convex optimization problem that simultaneously considers the formation preservation effect and formation scale regularization:

[0080]

[0081] The decision variables are and , For its own trajectory, For the formation scale with respect to time The function of is the total trajectory time, and are the formation keeping degree objective function and the regular term function respectively, is the regularization term weight.

[0082] (1) The objective function of formation preservation is designed to minimize the formation error distance:

[0083]

[0084] in, is the number of drones in the formation, It is The trajectory of the drone, For the team The formation target of the first UAV is The relative position of the formation target of the drones.

[0085] (2) The regularization objective function is designed to preserve the desired formation size specified by the user:

[0086]

[0087] For each moment, the above optimization problem is an unconstrained quadratic convex optimization problem, so the optimal solution at each moment can be analytically obtained, thereby obtaining the theoretically optimal formation trajectory. .

[0088] Step 5: Figure 4 As shown in the figure, based on the above optimal formation trajectory and the perception data of the drone’s airborne sensors, the first The actual flight trajectory of the UAV: ​​The UAV trajectory planning problem is constructed as a nonlinear optimization problem that takes into account the smoothness, motion efficiency, and formation keeping effect objectives, as well as safety, multi-aircraft collision avoidance, and kinematic and dynamic feasibility constraints:

[0089]

[0090] in, For the smoothing target item, For time target items, Keep the target item for the formation, is the collision penalty term, is the multi-machine conflict penalty term, is the kinematic and dynamic out-of-bounds penalty term. The decision variables in the above optimization problem include the trajectory of the UAV And the total trajectory time .

[0091] (1) Smoothing target item Designed to minimize jerk:

[0092]

[0093] (2) Time target item Designed to minimize time:

[0094]

[0095] (3) Formation maintenance target item is the difference between the trajectory and the optimal formation trajectory:

[0096]

[0097] in is the optimal formation trajectory corresponding to the fourth step.

[0098] (4) Obstacle collision penalty Penalties will be imposed for exceeding the safe flight corridor:

[0099]

[0100] in, is the number of convex geometric bodies in the flight corridor, For the The number of faces of a convex geometry, and For the The first convex geometry Parameters of each face.

[0101] (5) Multi-machine conflict penalty Will penalize being too close to other drones:

[0102]

[0103] in, is the minimum safe distance between drones.

[0104] (6) Kinematic and dynamic out-of-bounds penalty Exceeding the maximum speed or acceleration will be penalized:

[0105]

[0106] in, is the speed curve, is the acceleration curve, and are the maximum speed and maximum acceleration respectively.

[0107] The above nonlinear optimization problem is solved by gradient descent method to obtain A safe, efficient and proactive flight path for drones .

[0108] Step 6: Execute the resulting trajectory , and publish it to the wireless broadcast network: The drone will execute a new trajectory , and publish it to the wireless broadcast network, so that other drones can obtain the trajectory information of the drone.

[0109] like Figure 5 As shown, according to another embodiment of the present invention, a distributed UAV cluster autonomous formation flight device 500 is also disclosed, including: a memory 501, used to store program instructions; a processor 502, used to call the program instructions stored in the memory to implement the distributed UAV cluster autonomous formation flight method as the above embodiment.

[0110] According to another embodiment of the present invention, a computer-readable storage medium is disclosed, which stores program code, and the program code is used to implement the distributed drone cluster autonomous formation flight method as described in the above embodiment.

[0111] All or part of the steps in the various methods of the above embodiments can be completed by controlling the relevant hardware through a program, and the program can be stored in a readable storage medium, and the storage medium includes a read-only memory (ROM), a random access memory (RAM), a programmable read-only memory (PROM), an erasable programmable read-only memory (EPROM), a one-time programmable read-only memory (OTPROM), an electronically erasable programmable read-only memory (EEPROM), a compact disc read-only memory (CD-ROM) or other optical disc storage, magnetic disk storage, magnetic tape storage, or any other readable medium that can be used to carry or store data.

[0112] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. For those skilled in the art, the present invention may have various modifications and variations. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present invention shall be included in the protection scope of the present invention.

Claims

1. A distributed UAV swarm autonomous formation flight method, characterized in that: include: Obtain sensor data and flight trajectory information from multiple drones in a drone cluster through a wireless broadcast network to build a three-dimensional occupancy map; Given a mission start point and a mission end point, a path search is performed on the three-dimensional occupancy map based on an A* path search algorithm to obtain an initial path of the UAV unit; Extracting safe flight corridors: For each drone unit, the corresponding initial path is traversed in order, and whether to add a new convex body is determined by judging whether each point in the initial path is within the last convex body of the current flight corridor. The newly generated convex body is added to the flight corridor until every point on the initial path is traversed to generate a safe flight corridor. Obtain the trajectory information of other drones other than the aircraft itself, and calculate the theoretical optimal formation-keeping trajectory of the aircraft itself: The optimal formation-keeping problem is constructed as an unconstrained convex optimization problem that considers the formation-keeping effect and the regularization of the formation scale, and the theoretically optimal formation-keeping trajectory is obtained by solving the analytical solution; According to the theoretical optimal formation-keeping trajectory and the onboard sensor data, the actual flight trajectory of the aircraft is calculated in real time: with the flight smoothness, motion efficiency, and formation-keeping effect as the goals, the trajectory planning problem is constructed as a nonlinear optimization problem that satisfies flight safety, multi-aircraft collision avoidance, and kinematic and dynamic feasible constraints, and the actual flight trajectory is obtained by solving it; Execute the actual flight trajectory and publish the actual flight trajectory of the aircraft to the wireless broadcast network so that other drones can obtain the actual flight trajectory; The expression of the unconstrained convex optimization problem considering the formation preservation effect and formation scale regularization is: min∫0 T J form (p(t),s(t))+λ norm J norm (p(t),s(t))dt Among them, the decision variables are p(t) and s(t), p(t) is the own trajectory, s(t) is the function of the formation scale with respect to time t, T is the total trajectory time, J form and J norm are the objective function of formation preservation and the objective function of regular items, respectively, norm is the regularization term weight; The formation keeping degree objective function is designed to minimize the formation error distance: Where N is the number of drones in the formation, p o (t),o≠i is the trajectory of the oth UAV, f i,o is the relative position of the formation target of the i-th UAV and the formation target of the o-th UAV in the formation; The regularization objective function is designed to customize the desired shape preserving metric: J norm (p(t),s(t))=(s(t)-1) 2 For an unconstrained quadratic convex optimization problem at each moment, the optimal solution at each moment is analytically obtained to obtain the theoretically optimal formation trajectory p * (t); The optimization problem corresponding to the actual flight trajectory is: minJ=J s +λ t J t +λ f J f +λ o J o +λ c J c +λ d J d Among them, J s is the smoothing target term, J t is the time target term, J f For the formation maintenance objective, J o is the collision penalty term, J c is the multi-machine conflict penalty term, J d is the kinematic and dynamic out-of-bounds penalty term, and the decision variables in the optimization problem include the UAV trajectory p(t) and the total trajectory time T; The optimization problem is solved by the gradient descent method to obtain the actual flight trajectory p(t) of the i-th UAV; The smoothing target term J s Designed to minimize jerk: ∫0 T p (3) (t)dt The time target item J t Designed to minimize time: ∫0 T 1dt The formation maintaining target item J f Keep the difference between the trajectory and the optimal formation: Among them, p * (t) Maintaining track for optimal formation; The collision penalty term J o Used to punish those who exceed the safe flight corridor: Where n′ is the number of convex geometric bodies in the flight corridor, m j is the number of faces of the j-th convex body, C j,h With d j,h is the parameter of the hth face of the jth convex body; The multi-machine conflict penalty term J c Used to punish drones that do not meet the safety distance: Among them, d min is the minimum safe distance between drones; The kinematic and dynamic out-of-bounds penalty term J d Used to penalize those that exceed the maximum speed or acceleration: Among them, v(t) is the velocity curve, a(t) is the acceleration curve, and v max with a max are the maximum speed and maximum acceleration respectively.

2. The distributed UAV swarm autonomous formation flight method according to claim 1, characterized in that: The mission starting point is the current position of the UAV, the mission end point is the position of the mission target point or the local target point, and the initial path is a collision-free safe path.

3. A distributed UAV swarm autonomous formation flight device, characterized in that: include: A memory for storing program instructions; A processor is used to call the program instructions stored in the memory to implement the distributed unmanned aerial vehicle cluster autonomous formation flight method as described in claim 1 or 2.

4. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores program code, and the program code is used to implement the distributed drone cluster autonomous formation flight method as described in claim 1 or 2.

Images