A method and device for real-time planning of high-speed obstacle avoidance trajectories for a multi-machine hoisting system
Through dynamic modeling and multi-objective optimization, combined with A* algorithm and parallel solution, the problems of unsmooth trajectory and safety hazards of multi-machine lifting systems in high-speed maneuvering flight are solved, and high-speed obstacle avoidance and collision-free path planning are realized. Improves the overall performance and real-time performance of the system.
Patent Information
- Application Number
- CN202411148852.5
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- Filing Date
- 2024-08-21
- Publication Date
- 2024-11-15
- Estimated Expiration
- 2044-08-21
AI Technical Summary
The existing multi-machine lifting system has problems such as unsmooth trajectory, state oscillation, safety hazards, dynamic constraints and insufficient real-time performance during high-speed maneuvering flight, especially in complex environments, the high-speed obstacle avoidance of drones cannot be guaranteed. and collision-free path planning.
Through dynamic modeling, building a scalable positive pyramid configuration, indirect characterizing the position of the drone, optimizing the rope direction and angle, and building multi-objective optimization problems, combined with A* algorithm and parallel solution, optimize the trajectory to ensure the smoothness and energy efficiency of the trajectory. , dynamic enforceability and safety.
The high-speed obstacle avoidance trajectory planning of multi-machine lifting system has been realized, which improves the smoothness of the system and high-speed flight capabilities, ensures the safety and real-timeness of the drone, avoids the cross-entangle of ropes, and improves the efficiency of trajectory optimization. and scope of application.
Smart Images

Figure CN118672296B_ABST
Abstract
Description
Technical Field
[0001] The present invention belongs to the field of logistics transportation of unmanned aerial vehicles, and in particular relates to a method and device for real-time planning of high-speed obstacle avoidance trajectories of a multi-machine hoisting system. Background Art
[0002] In recent years, with the continuous growth of logistics and transportation demand, fixed-wing unmanned aerial vehicles cannot be further expanded to cargo transportation tasks in jungles, urban spaces, and inside buildings due to their high flight speed and large turning radius. Multi-rotor drones have the advantages of small size, low cost, vertical take-off and landing, and strong maneuverability. They are the most widely used type of aircraft in military and commercial applications in recent years, and can perform a variety of cargo transportation tasks.
[0003] Multi-rotor drones with high-speed maneuverable flight capabilities can significantly improve mission efficiency, but existing commercial cargo transportation solutions using a single multi-rotor drone (referred to as a single-machine hoisting system) still rely on manual operation and external positioning. The hoisting flight speed is often very slow, and even skilled operators find it difficult to efficiently perform high-speed maneuverable flight missions in complex environments. The terminal delivery capability is poor (personnel support is required), and is only suitable for slow cargo transportation scenarios with open environments and simple tasks.
[0004] The single-machine lifting system has inherent defects: 1) The load-bearing capacity is limited. The load-bearing range that a single drone can adapt to is small, and the drone cannot carry heavy objects that exceed the upper limit of the load-bearing capacity; 2) The posture controllability of the heavy object is poor. It is impossible to provide torque through only one rope, so the posture of the heavy object cannot be controlled; 3) The system's emergency obstacle avoidance capability is weak. After the drone makes an emergency brake, the existence of inertia causes the heavy object to continue to swing a certain distance in the direction of movement, and drive the drone to swing forward together, posing a safety hazard.
[0005] Compared with a single-machine lifting system, a multi-machine lifting system composed of multiple multi-rotor drones has the following advantages: 1) Strong load capacity. The multi-machine lifting system can take advantage of the large number of drones to improve the load capacity and improve the system's adaptability to the size and weight of heavy objects. There is no need to design platforms with different load capacities due to different sizes and weights of heavy objects; 2) The posture of heavy objects is highly controllable and can be used in scenarios with special requirements for the posture of heavy objects, such as using multiple drones to lift minesweeping equipment. By adjusting the position of the drones, the minesweeping equipment is always parallel to the uneven terrain; 3) The system has strong emergency obstacle avoidance capabilities. The multi-machine lifting system can provide the necessary instantaneous braking control force for heavy objects through multiple ropes in different directions to avoid large swings in the ropes, and has a stronger ability to respond to scenarios such as emergency stops and emergency obstacle avoidance. Therefore, the multi-machine lifting system has attracted widespread attention at home and abroad.
[0006] At present, the existing multi-machine hoist system motion planning scheme still has the following problems:
[0007] 1. Existing research on multi-drone systems generally selects flat output variables that characterize the system state from the perspective of the system as a whole. However, this flat output selection method cannot provide independent flat output variables for all the drones in the system to directly characterize their trajectories. There are always a few drones in the system whose positions depend on the flat output variables of all other drones. Therefore, the energy of these drone trajectories cannot be taken into account in the optimization objective function, which may cause the trajectories of these drones to be uneven and the states to have high-frequency and large oscillations, restricting the overall high-speed maneuvering flight capability of the system.
[0008] 2. Existing solutions often select the three components of the rope force vector as flat output variables. In this type of solution, the position of the suspended drone is often represented by the unit direction vector of the rope force vector. However, for this representation method, when the magnitude of the force approaches zero, it will cause singular points in the position of the suspended drone.
[0009] 3. In existing research, there are many fast and safe trajectory search schemes for particle models, but there is a lack of fast, safe and collision-free path planning schemes for complex configurations of multi-machine hoisting systems.
[0010] 4. Most multi-machine lifting system motion planning only considers the trajectory planning of heavy objects. The control force required by the heavy objects is obtained through closed-loop control of the heavy objects, and the control force is distributed to each drone according to the principle of minimum force norm. For this method, the motion of the lifting drone is uniquely determined by the motion of the heavy objects and the force distribution process. In the heavy object motion planning and subsequent control process, the obstacle avoidance requirements of the lifting drone cannot be considered, so the safety of the lifting drone cannot be guaranteed.
[0011] 5. Existing trajectory optimization schemes cannot guarantee the dynamics of high-speed maneuvering trajectories. UAVs are limited by factors such as system delay, response time, physical upper bounds, and their own dynamics. Some states of the UAV (maximum thrust, etc.) cannot exceed their upper bounds, and the UAV cannot adapt to excessively rapid changes in some system states (acceleration, angular velocity, etc.). The limitations of UAV capabilities cause the system to be unable to keep up with the planned trajectory, resulting in safety hazards such as system crashes. In order to ensure the dynamics of the trajectory are executable, existing schemes often only consider rough and conservative dynamic constraints such as maximum speed and maximum acceleration. However, rough dynamic constraints often make the trajectory too conservative, resulting in slow and less maneuverable optimized trajectories. Therefore, it is not suitable for high-speed maneuverable flight of multi-machine hoisting systems.
[0012] 6. For the multi-machine hoisting system of the rigid body model, the existing research has not considered the possible cross-entanglement between ropes during the optimization process.
[0013] 7. A few full-state motion planning schemes for multi-machine hoisting systems based on nonlinear programming are difficult to meet the real-time planning requirements due to the limitation of the solution speed of general solvers. Summary of the invention
[0014] In view of the problems existing in the prior art, the purpose of the embodiments of the present application is to provide a method and device for real-time planning of high-speed obstacle avoidance trajectories for a multi-machine hoisting system.
[0015] According to a first aspect of an embodiment of the present application, a method for real-time planning of high-speed obstacle avoidance trajectories for a multi-machine hoisting system is provided, comprising:
[0016] Performing dynamic modeling on a multi-machine hoisting system, wherein the multi-machine hoisting system includes a plurality of hoisting drones and a heavy object connected to each of the hoisting drones via a rope;
[0017] The configuration of the multi-machine hoisting system is constructed as a retractable regular pyramid, and a collision-free path is quickly planned based on the configuration;
[0018] The position of the suspended drone is indirectly represented by the position of the weight and the vector direction of the rope;
[0019] The optimized variables representing the rope direction angle with constraints are converted into unconstrained dummy variables, so as to eliminate the rope direction angle constraints;
[0020] Construct a multi-objective optimization problem for high-speed obstacle avoidance trajectory optimization of a multi-machine hoisting system, wherein the multi-objective optimization problem includes an energy function that characterizes the smoothness of the trajectory and the required energy, a total time function that characterizes the trajectory execution time, fine dynamics constraints, safety constraints, and coupled dynamics constraints;
[0021] Based on the multi-objective optimization problem, trajectory optimization is performed with the collision-free path as the initial trajectory, wherein the energy function characterizing the smoothness of the trajectory and the required energy, the fine dynamics constraints, the safety constraints and the coupled dynamics constraints are solved in parallel.
[0022] Furthermore, the multi-machine hoisting system includes N multi-rotor drones as hoisting drones, and the dynamic model of the heavy objects in the system is as follows:
[0023]
[0024] in are the mass of the weight and its moment of inertia about its center of mass, are the position and velocity of the weight in the world coordinate system, is the rotation matrix between the weight coordinate system and the world coordinate system, is the angular velocity of the weight, for The corresponding antisymmetric matrix is, is the tension exerted by the i-th rope on the weight, is the unit vector of the rope direction, pointing from the weight to the drone, is the vector between the center of mass and the point where the rope is connected, pointing from the center of mass to the point where the rope is connected. is the antisymmetric matrix corresponding to the vector, ;
[0025] The dynamic model of a single suspended UAV is:
[0026]
[0027] The subscript Representing the A suspended drone. are the mass of the mounted drone and its moment of inertia around the center of mass, are the position and speed of the mounted UAV in the world coordinate system, is the rotation matrix between the body axis system of the mounted UAV and the world coordinate system, for The corresponding quaternion is, is the angular velocity of the mounted UAV, is the antisymmetric matrix corresponding to the angular velocity, is the angular acceleration of the mounted UAV. They are the total thrust and attitude control torque of the suspended UAV respectively. is the tension exerted by the rope on the body. is the unit vector of the rope direction, pointing from the weight to the drone, and from the pitch angle of the rope and azimuth The only certainty,
[0028]
[0029] in , is the tension of the rope on the suspended drone.
[0030] Furthermore, the position of the weight , describing the elevation angle of the regular pyramid space posture , azimuth and the radius of the circumscribed circle of the regular triangle at the base of the pyramid As the configuration parameter of the multi-machine lifting system.
[0031] Furthermore, the obstacle avoidance trajectory is planned based on the A* algorithm, wherein the configuration parameters of the pyramid are aligned during the trajectory search. Discretize the configuration parameters and use them as search variables in the algorithm. And the shape parameters of the regular pyramid To calculate the position of the system as a whole in space, and thus determine whether it collides with obstacles in the map.
[0032] Furthermore, the position of the suspended drone is indirectly characterized by the position of the weight and the vector direction of the rope, specifically:
[0033]
[0034] in is the position of the i-th mounted UAV, is the position of the weight, is the unit vector representing the direction of the i-th rope, pointing from the weight to the hanging drone, is the vector between the center of mass and the point where the rope is connected, pointing from the center of mass to the point where the rope is connected. is the length of the rope.
[0035] Furthermore, the optimized variable representing the rope direction angle with constraints is equivalently converted into an unconstrained dummy variable, where the pitch angle The equivalent conversion process is:
[0036] Introducing dummy optimization variables ;
[0037] Create virtual optimization variables and actual rope direction angle variables Diffeomorphisms between :
[0038]
[0039] in is the pitch angle The upper and lower bounds of
[0040] Thus, the constrained pitch angle is equivalently converted into an unconstrained dummy variable.
[0041] Furthermore, in the multi-objective optimization problem,
[0042] The energy function characterizing the smoothness of the trajectory and the required energy ,in, is the weight position, let , but The trajectory of the force vector optimization variable of the rope connected to the i-th suspended drone, To adjust the coefficient of the two energy functions, T is the total duration of the trajectory, The number of drones being hoisted;
[0043] The total time function representing the execution time of the trajectory ,in, is the time of one segment of the multi-segment polynomial trajectory;
[0044] The fine dynamics constraints include a maximum velocity constraint violation function for the suspended UAV , maximum / minimum thrust constraint violation function , maximum axis angle constraint violation function , angular rate constraint violation function , angular acceleration constraint violation function , where i represents the i-th suspended UAV, j represents the j-th segment of the trajectory, and k represents the k-th sampling point on the j-th segment of the trajectory. is the speed of the mounted UAV in the world coordinate system, is the maximum permissible speed of the suspended drone, is the total thrust of the suspended UAV, are the maximum and minimum allowable thrust, respectively. is the angle between the z-axis of the body axis system of the suspended UAV and the direction of gravity acceleration, is the maximum allowable shaft angle, is the angular velocity of the mounted UAV, is the maximum permissible angular velocity, is the angular acceleration of the mounted UAV, is the maximum permissible angular acceleration;
[0045] The safety constraints include a system obstacle avoidance safety constraint violation function and the mutual collision avoidance constraint violation function within the multi-machine hoisting system ,in is the number of sampling points on the rope, is the maximum allowable safe distance from obstacles, is the kth sampling point on the jth trajectory of the heavy object;
[0046] The coupled dynamics constraints include a heavy object translation dynamics constraint violation function and the mass rotational dynamics constraint violation function ,in are the mass of the weight and its moment of inertia about its center of mass, is the velocity of the mass in the world coordinate system, is the tension exerted by the i-th rope on the weight, is the unit vector of the rope direction, , is the angular velocity of the weight, is the angular acceleration of the mass, is the antisymmetric matrix corresponding to the vector between the center of mass and the point where the rope is connected, is the rotation matrix between the weight coordinate system and the world coordinate system, is the number of mounted drones.
[0047] According to a second aspect of an embodiment of the present application, a high-speed obstacle avoidance trajectory real-time planning device for a multi-machine hoisting system is provided, comprising:
[0048] A dynamic modeling module, used for performing dynamic modeling on a multi-machine hoisting system, wherein the multi-machine hoisting system includes a plurality of hoisting drones and a heavy object connected to each hoisting drone via a rope;
[0049] A trajectory planning module, used to construct the configuration of the multi-machine hoisting system into a retractable regular pyramid, and to perform rapid planning of a collision-free path based on the configuration;
[0050] The position representation module is used to indirectly represent the position of the suspended drone through the position of the weight and the vector direction of the rope;
[0051] The constraint elimination module is used to convert the optimization variables representing the rope direction angle with constraints into unconstrained virtual variables, so as to eliminate the rope direction angle constraints;
[0052] An optimization problem construction module is used to construct a multi-objective optimization problem for high-speed obstacle avoidance trajectory optimization of a multi-machine hoisting system, wherein the multi-objective optimization problem includes an energy function representing the smoothness of the trajectory and the required energy, a total time function representing the trajectory execution time, fine dynamics constraints, safety constraints, and coupled dynamics constraints;
[0053] A trajectory optimization module is used to perform trajectory optimization based on the multi-objective optimization problem and taking the collision-free path as the initial trajectory, wherein the energy function characterizing the smoothness of the trajectory and the required energy, the fine dynamics constraints, the safety constraints and the coupled dynamics constraints are solved in parallel.
[0054] According to a third aspect of an embodiment of the present application, there is provided an electronic device, including:
[0055] one or more processors;
[0056] A memory for storing one or more programs;
[0057] When the one or more programs are executed by the one or more processors, the one or more processors implement the method as described in the first aspect.
[0058] According to a fourth aspect of an embodiment of the present application, a computer-readable storage medium is provided, on which computer instructions are stored, and when the instructions are executed by a processor, the steps of the method described in the first aspect are implemented.
[0059] The technical solution provided by the embodiments of the present application may have the following beneficial effects:
[0060] 1. The present invention selects a flat output variable representing the state of each suspended drone from the perspective of a single drone in the system. The flat output variable of the i-th drone is the position of the heavy object. , attitude quaternion , heading angle , and the pitch angle representing the rope force vector , azimuth , the magnitude of the force And the heading angle of the drone On this basis, the overall expanded flat output space of the system is selected as , expanding the flat output space can ensure that each suspended UAV in the system has an independent flat output variable to directly represent its position, which can ensure the smoothness of the trajectory of each suspended UAV, improve the high-speed obstacle avoidance flight capability of the multi-drone suspended system, and solve the problem that the existing flat output selection method from the perspective of the overall system cannot ensure the smoothness of the trajectory of some UAVs and limit the high-speed obstacle avoidance flight capability of the system;
[0061] 2. The present invention proposes a novel method for indirectly characterizing the position of a UAV in a multi-machine lifting system, which uses the position of the UAV and the pitch angle and azimuth angle of the rope to uniquely characterize the trajectory of the UAV, thereby eliminating the kinematic constraints between the UAV and the heavy object and avoiding the singular point problem of the UAV position that may be caused by other characterization methods.
[0062] 3. The present invention proposes a safe and collision-free path planning method for a multi-machine hoisting system. The complex configuration space of the multi-machine hoisting system is simplified to a configuration space with only 6 degrees of freedom, so as to realize fast safe and collision-free path planning in a complex obstacle environment.
[0063] 4. This paper proposes a lightweight multi-objective optimization framework suitable for high-speed maneuvering trajectory optimization of a multi-machine hoisting system. The fine dynamic constraints of the hoisted UAV are considered to ensure that the optimized high-speed maneuvering trajectory dynamics are executable, and the obstacle avoidance / collision avoidance constraints of the system are further considered to ensure the safety of the entire multi-machine hoisting system including heavy objects and UAVs.
[0064] 5. The present invention proposes a method for eliminating the topological constraints of rope optimization variables. The variables representing the rope optimization are subjected to differential homeomorphism mapping to eliminate the topological constraints of the rope direction angle and avoid the possible cross-entanglement of the rope.
[0065] 6. Parallel evaluation of constraint violations of each UAV and solution of polynomial trajectory parameter gradients ensures that trajectory optimization time does not increase linearly with the number of UAVs in the system, ensuring the real-time performance of trajectory optimization while improving the scalability of the number of UAVs.
[0066] It should be understood that the foregoing general description and the following detailed description are exemplary and explanatory only and are not restrictive of the present application. BRIEF DESCRIPTION OF THE DRAWINGS
[0067] The accompanying drawings, which are incorporated in and constitute a part of this specification, illustrate embodiments consistent with the present application and, together with the description, serve to explain the principles of the present application.
[0068] Figure 1 The present invention is a flowchart of a real-time trajectory planning method for a multi-machine hoisting system according to an exemplary embodiment.
[0069] Figure 2 It is a force analysis diagram of a multi-machine hoisting system according to an exemplary embodiment.
[0070] Figure 3 It is a schematic diagram of a simplified configuration space definition of a multi-machine hoisting system according to an exemplary embodiment.
[0071] Figure 4 It is a schematic diagram of the rope azimuth angle restriction area of a multi-machine hoisting system according to an exemplary embodiment.
[0072] Figure 5 It is a spatiotemporal deformation framework diagram of the trajectory of a multi-machine hoisting system according to an exemplary embodiment.
[0073] Figure 6 is a schematic diagram of a high-speed obstacle avoidance trajectory of a multi-machine hoisting system according to an exemplary embodiment.
[0074] Figure 7 It is a schematic diagram of the obstacle avoidance effect of a high-speed obstacle avoidance trajectory of a multi-machine hoisting system according to an exemplary embodiment.
[0075] Figure 8 It is a block diagram of a real-time trajectory planning device for a multi-machine hoisting system according to an exemplary embodiment.
[0076] Fig. 9 The diagram is a schematic diagram of an electronic device according to an exemplary embodiment. DETAILED DESCRIPTION
[0077] Here, exemplary embodiments are described in detail, and examples thereof are shown in the accompanying drawings. When the following description refers to the drawings, unless otherwise indicated, the same numbers in different drawings represent the same or similar elements. The implementations described in the following exemplary embodiments do not represent all implementations consistent with the present application.
[0078] The terms used in this application are for the purpose of describing specific embodiments only and are not intended to limit this application. The singular forms of "a", "said" and "the" used in this application and the appended claims are also intended to include plural forms unless the context clearly indicates other meanings. It should also be understood that the term "and / or" used herein refers to and includes any or all possible combinations of one or more associated listed items.
[0079] The present invention is aimed at a multi-machine hanging system composed of multiple multi-rotor hanging UAVs, and proposes a real-time planning method for high-speed obstacle avoidance trajectories suitable for a multi-machine hanging system in a complex obstacle environment. Since the planning method of the present invention takes into account the fine dynamic constraints during high-speed maneuvering flight, the dynamic feasibility of the system during high-speed maneuvering flight of the multi-machine hanging system can be guaranteed. Therefore, unlike the traditional trajectory planning method of the multi-machine hanging system, which can only plan low-speed trajectories by restricting rough and conservative dynamic constraints, the trajectory planning method proposed in the present invention can realize both low-speed trajectory planning and high-speed maneuvering trajectory planning. In addition, the real-time performance of the trajectory planning method of the present invention is higher than that of the traditional trajectory planning method of the multi-machine hanging system, and the method has a wider range of application.
[0080] The method is described in detail below.
[0081] S1. Performing dynamic modeling on a multi-machine hoisting system, wherein the multi-machine hoisting system includes a plurality of hoisting drones and a heavy object connected to each of the hoisting drones via a rope;
[0082] The dynamic modeling of UAV considering the time-varying tension of the rope, the force analysis diagram of the multi-machine hanging system and some symbol definitions are as follows Figure 2 shown.
[0083] Assume that the multi-machine lifting system has A multi-rotor drone is used as a hanging drone. First, a dynamic model of the heavy objects in the system is established.
[0084]
[0085] in are the mass of the weight and its moment of inertia about its center of mass, are the position and velocity of the weight in the world coordinate system, is the rotation matrix between the weight coordinate system and the world coordinate system, and its corresponding quaternion is , is the angular velocity of the weight, is the antisymmetric matrix corresponding to the angular velocity, is the angular acceleration of the mass. is the tension exerted by the i-th rope on the weight, is the unit vector of the rope direction, pointing from the weight to the drone, It is the vector between the center of mass and the point where the rope is connected, pointing from the center of mass to the point where the rope is connected. is the antisymmetric matrix corresponding to the vector, .
[0086] Then, the dynamic model of a single suspended UAV is established.
[0087]
[0088] The subscript Representing the A suspended drone. are the mass of the mounted drone and its moment of inertia around the center of mass, are the position and speed of the mounted UAV in the world coordinate system, is the rotation matrix between the body axis system of the mounted UAV and the world coordinate system, for The corresponding quaternion is, is the angular velocity of the mounted UAV, is the antisymmetric matrix corresponding to the angular velocity, is the angular acceleration of the mounted UAV. They are the total thrust and attitude control torque of the suspended UAV respectively. is the tension exerted by the rope on the body. is the unit vector of the rope direction, pointing from the weight to the drone, and from the pitch angle of the rope and azimuth The only certainty,
[0089]
[0090] in , is the tension of the rope on the suspended drone.
[0091] S2, constructing the configuration of the multi-machine hoisting system into a retractable regular pyramid, and performing rapid planning of a collision-free path based on the configuration;
[0092] In order to reduce the complexity of the trajectory planning problem and improve the efficiency of the overall trajectory search of the system, the freedom of describing the position changes of each drone and heavy object in the system configuration space is ignored, and the multi-machine hanging system configuration is regarded as a scalable regular pyramid. , describing the elevation angle of the regular pyramid space posture , azimuth and the radius of the circumscribed circle of the regular triangle at the base of the pyramid As a system-simplified configuration space, such as Figure 3 As shown. The position of the weight describes the overall displacement of the system, and the elevation angle , azimuth And the radius It is used to improve the overall flexibility of the system in passing through complex and narrow environments. For example, for an inclined square channel, the elevation angle can be adjusted and azimuth To improve the safety probability of multi-machine hoisting system, for narrow passages, the radius can be reduced. To improve the overall safety passing probability of the system.
[0093] Finally, based on the constructed local grid map, a trajectory planning algorithm based on the A* algorithm is designed to achieve fast collision-free path planning for the simplified configuration space. The specific implementation plan of the trajectory planning algorithm is as follows: each point in the grid map is regarded as a node. The algorithm selects the next node to be expanded by evaluating the estimated cost of each node to the target node, thereby finding the shortest trajectory from the starting point to the target point. When searching for a path, the configuration parameters of the pyramid are aligned. discretize it and use it as a search variable in the algorithm, and pass it through the position of the weight And the shape parameters of the regular pyramid To calculate the position of the system as a whole in space, to determine whether it collides with obstacles in the map. For the collision-free path planned by the above method, its path points are discrete, the path is not smooth, and no fine dynamic constraints are considered. Compared with a continuous trajectory, this collision-free path is relatively rough. Therefore, in step S6, a high-quality trajectory that is continuous and smooth and meets safety constraints and fine dynamic constraints will be further solved by optimization.
[0094] S3, indirectly characterize the position of the suspended drone through the position of the weight and the vector direction of the rope;
[0095] For a multi-machine lifting system, the drone and the weight are connected by a rope. When the rope is taut, there is a kinematic constraint between the drone and the weight, that is, the relative distance between the drone and the weight is always the length of the rope. If the positions of the drone and the weight are represented independently, it is necessary to add a kinematic constraint between the drone and the weight, which will increase the complexity of the optimization problem.
[0096] Therefore, it is necessary to expand the flat output space so that each UAV in the system can have an independent flat output variable to directly represent its trajectory. The flat output variable of the i-th UAV is the position of the weight. , attitude quaternion , heading angle , represents the pitch angle of the rope force vector , azimuth , the magnitude of the force And the yaw angle of the drone .
[0097] The present invention proposes a new indirect representation method for the position of a suspended UAV, which uniquely determines the position of the suspended UAV through the position of the weight and the vector direction of the rope.
[0098]
[0099] in is the position of the i-th mounted UAV, is the position of the weight, is the unit vector representing the direction of the i-th rope, pointing from the weight to the hanging drone, is the vector between the center of mass and the point where the rope is connected, pointing from the center of mass to the point where the rope is connected. is the length of the rope. This representation method can effectively prevent the rope force from approaching zero, that is, The drone position singularity caused by
[0100] choose is the extended flat output variable of the system. The state and control input of the i-th suspended UAV in the system can be directly obtained by the flat output variable and its higher-order derivatives are uniquely solved. Determine the position of the i-th mounted drone in the system , the kinematic states of the UAV such as velocity, acceleration and jerk are further determined by continuous differentiation of spherical coordinates. Then, the total thrust of the UAV is solved in step S5 by differential flat mapping , shaft angle , angular velocity , angular acceleration etc. high-level status / control inputs.
[0101] S4, converting the optimization variable representing the rope direction angle with constraints into an unconstrained dummy variable, thereby eliminating the rope direction angle constraints;
[0102] Rope's azimuth Always Between, and the pitch angle Always Therefore, when selecting two angles representing the direction of the rope as optimization variables, it is necessary to ensure that the optimization variables are always within the above bounded range. In addition, when a multi-machine hoist system transports a rigid body, the ropes may be cross-entangled due to the different positions of the ropes fixed on the rigid body. Therefore, it is necessary to constrain the azimuth of different ropes to ensure the separation of azimuths between different ropes. Taking the four-machine hoist system as an example, The azimuth angle of the four UAVs is divided into four equal parts, and the azimuth angle of the rope of the four UAVs is limited to the four equally divided areas. The azimuth angle of the rope carrying UAV 2 is limited to the R2 area, that is, ,like Figure 4 As shown. According to actual needs, adjust the pitch angle of the rope Restricted to Within the angle range, that is .
[0103] In order to eliminate the directional angle constraint of the rope, the present invention proposes a method for eliminating the directional angle constraint. As an example, introduce virtual optimization variables , establish virtual optimization variables and the actual direction angle variables of the rope Diffeomorphisms between :
[0104]
[0105] in is the pitch angle The upper and lower bounds of . Thus, the optimization variable representing the rope direction angle with constraints is equivalently converted into an unconstrained dummy variable. From the above formula, it can be seen that no matter the dummy variable In real number space How to get the value in the actual direction angle of the rope Always satisfied , thereby eliminating the direction angle constraint.
[0106] S5. Based on the indirectly characterized position of the hoisted UAV, a multi-objective optimization problem for high-speed obstacle avoidance trajectory optimization of a multi-machine hoisted system is constructed, wherein the multi-objective optimization problem includes an energy function representing the smoothness of the trajectory and the required energy, a total time function representing the trajectory execution time, fine dynamics constraints, safety constraints, and coupled dynamics constraints;
[0107] The high-speed obstacle avoidance trajectory optimization problem of a multi-machine hoisting system is defined as a lightweight multi-objective optimization problem. The objective functions of this problem include the following five categories: an energy function that characterizes the smoothness of the trajectory and the required energy, a total time function of the trajectory, fine dynamic constraints of the drone, safety constraints, and coupled dynamic constraints between optimization variables. The present invention uses a polynomial trajectory class that can be characterized by sparse spatiotemporal parameters to describe the trajectory of the optimization variable. In fact, the polynomial trajectory class is the only solution to the unconstrained minimization energy optimization problem. The deformation of the trajectory is achieved by optimizing the spatiotemporal parameter gradients that characterize the trajectory, thereby completing the optimization of the trajectory. The constraint violation function at discrete time points is evaluated by uniformly sampling the trajectory in time. Finally, the evaluation result of the constraint violation function is converted into a sparse parameter gradient of the optimization variable trajectory.
[0108] The constraint violation function is not smooth at the zero point, that is, the gradient does not exist. In order to ensure that the optimization constraint violation function is three times continuously differentiable, the nonlinear transformation function of the constraint violation function is first defined.
[0109]
[0110] in , usually through To adjust the size of the interval at the zero point of the constraint violation function to smooth the objective function at the zero point. The smaller the value, the smaller the range of smooth gradient transition.
[0111] (1) Energy function that characterizes the smoothness of the trajectory and the required energy :This function consists of two parts, namely 1) the energy function of the weight trajectory, which describes the energy and smoothness required to execute the system trajectory; 2) the energy function of the rope force vector trajectory, which describes the smoothness of the drone trajectory. These two energy functions can be directly calculated using the third-order derivative integral of the trajectory:
[0112]
[0113] in, is the weight position, let , but The trajectory of the force vector of the rope connected to the i-th drone is optimized. is the coefficient to adjust the weight of the two energy functions, and T is the total duration of the trajectory.
[0114] (2) Total time function representing the trajectory execution time : This function is the sum of each time segment of multiple polynomial trajectories.
[0115]
[0116] in, is the first segment of the multi-segment polynomial trajectory The time of the trajectory.
[0117] (3) Fine dynamic constraints: This type of constraints includes the maximum speed constraint, maximum axis angle constraint, maximum angular velocity constraint, maximum angular acceleration constraint, maximum / minimum thrust constraint, etc. of the suspended UAV.
[0118] 1) The maximum speed constraint violation function of the suspended UAV is defined as:
[0119]
[0120] Where i represents the i-th suspended UAV, j represents the j-th segment of the trajectory, and k represents the k-th sampling point on the j-th segment of the trajectory. is the maximum speed allowed for the suspended UAV. The speed of the suspended drone can be indirectly expressed by a flat output variable:
[0121]
[0122] 2) The maximum / minimum thrust constraint violation function of the suspended UAV is defined as
[0123]
[0124] in are the maximum and minimum thrust allowed respectively, i represents the i-th suspended UAV, j represents the j-th segment of the trajectory, and k represents the k-th sampling point on the j-th segment of the trajectory.
[0125] Since the suspended drone has a differentially flat property, the first The z-axis direction vector of the drone It can be expressed as:
[0126]
[0127] No. The total thrust of a suspended UAV is calculated as follows:
[0128]
[0129] There is an upper limit on the rotation speed of the suspended UAV, so the maximum thrust limit ensures that the upper limit of the thrust of the suspended UAV is within a safe range.
[0130] 3) The maximum axis angle constraint violation function of the suspended UAV is defined as
[0131]
[0132] in is the maximum allowable axis angle, i represents the i-th suspended UAV, j represents the j-th segment of the trajectory, and k represents the k-th sampling point on the j-th segment of the trajectory.
[0133] Given yaw angle , we can solve the quaternion representing the attitude of the suspended drone
[0134]
[0135] in, They are The first, second, and third components of the vector. The angle between the z-axis of the drone's body coordinate system and the direction of gravity acceleration The cosine of is defined as:
[0136]
[0137] in is a quaternion In The component corresponding to the element.
[0138] 4) The angular rate constraint violation function of the suspended UAV is defined as
[0139]
[0140] in is the maximum angular velocity allowed. In the above formula, i represents the i-th suspended UAV, j represents the j-th segment of the trajectory, and k represents the k-th sampling point on the j-th segment of the trajectory. Angular velocity of the mounted drone The analytical expression is as follows:
[0141]
[0142] 5) Angular acceleration constraint violation function of the suspended drone
[0143]
[0144] Among them, i represents the i-th suspended UAV, j represents the j-th segment of the trajectory, k represents the k-th sampling point on the j-th segment of the trajectory, and Angular acceleration of the mounted drone The analytical expression is as follows:
[0145]
[0146]
[0147]
[0148] in is the maximum permissible angular acceleration, are the three components of the angular acceleration vector.
[0149] (4) Safety constraints
[0150] The multi-machine hoisting system needs to consider the obstacle avoidance requirements of the hoisted drone, rope and heavy objects at the same time. Therefore, the present invention takes uniform sampling points, determine this Obstacle avoidance safety constraints at sampling points, where is the number of sampling points, which can be adjusted by the user according to their own optimization needs To adjust the sampling density.
[0151] 1) System obstacle avoidance safety constraint violation function
[0152]
[0153] in is the maximum allowable safe distance from obstacles, is the kth sampling point on the jth trajectory of the heavy object. When evaluating whether the heavy object collides with the obstacle, When evaluating whether the suspended drone collides with an obstacle, When evaluating whether the sampling point in the middle of the rope collides with an obstacle. It is the Euclidean signed distance field function, which can determine the Euclidean signed distance between each sampling point and the obstacle.
[0154] 2) Collision avoidance constraint violation function within the multi-machine hoisting system
[0155]
[0156] Among them, i represents the i-th hoisted UAV, j represents the j-th segment of the trajectory, and k represents the k-th sampling point on the j-th segment of the trajectory. This constraint is used to avoid collisions between any two hoisted UAVs in the multi-machine hoisting system.
[0157] (5) Coupled dynamic constraints
[0158] Since the weight and the rope are characterized by independent optimization variables, the weight dynamics and the rope tension do not satisfy the weight translational dynamics and rotational dynamics constraints. The present invention provides the coupled dynamics constraints that need to be satisfied between the rope optimization variables and the weight optimization variables.
[0159] 1) Heavy object translation dynamic constraint violation function
[0160]
[0161] 2) Heavy object rotation dynamics constraint violation function
[0162]
[0163] Where i represents the i-th suspended UAV, j represents the j-th segment of the trajectory, and k represents the k-th sampling point on the j-th segment of the trajectory. The function makes the heavy object satisfy the translational dynamics constraint, The function makes the weight satisfy the rotational dynamics constraints.
[0164] S6. Based on the multi-objective optimization problem, the collision-free path is used as the initial trajectory to perform trajectory optimization, wherein the energy function characterizing the smoothness of the trajectory and the required energy, the fine dynamics constraint, the safety constraint and the coupled dynamics constraint are solved in parallel;
[0165] For the constructed multi-objective optimization problem, the computational complexity of a single iteration increases linearly with the increase in the number of drones in the system. Therefore, as the number of suspended drones increases, the time required for the traditional serial constraint evaluation and polynomial trajectory parameter gradient calculation process will continue to increase, resulting in an increase in the system trajectory optimization time, thereby reducing the real-time performance of the optimization and restricting the high-speed flight of the system. Since the flat output space that characterizes the state and control input of the suspended drones is expanded in step S3, the state and control input of different suspended drones can be uniquely represented by different flat output variables and their high-order derivatives. Therefore, the present invention proposes a parallel constraint evaluation and polynomial trajectory parameter gradient calculation framework, such as Figure 5 As shown in Figure 2, each suspended drone and its connected rope are considered as a subsystem, and the subsystem energy function defined in step S5 is evaluated in parallel. , safety constraint violation function , fine dynamics constraint violation function and the coupled dynamics constraint violation function . If these constraints are violated, that is, these constraint violation functions are greater than zero, then these constraint violation functions are used to further calculate the polynomial parameter gradients of the subsystem flat output variables. The time required for a single iteration of this parallel constraint evaluation and gradient calculation framework does not increase with the increase in the number of mounted drones, thereby avoiding the reduction in the real-time performance of trajectory optimization caused by the increase in the number of mounted drones in the system. After all subsystem constraint violation functions are solved in parallel with respect to their respective polynomial parameter gradients, the total time for solving the trajectory separately is About polynomial time gradient. Then add all the gradients to obtain the sparse space-time parameter gradient of the constraint violation function with respect to all extended flat output variables. Finally, the sparse space-time parameter gradient is solved by the polynomial parameter gradient, and the L-BFGS method is further used to solve the update direction and step size of the sparse space-time parameter to determine whether the exit condition is met. In one embodiment, the exit condition is that the cost function decreases less than 0.0000 for three consecutive iterations. If the exit condition is not met, the sparse spatiotemporal parameters are updated based on the update direction and step size, and the new trajectory is generated by the fast trajectory generation method given by the polynomial trajectory class, thereby completing a single trajectory spatiotemporal deformation process; the above process is repeated until the exit condition is met.
[0166] Through the above method, the high-speed obstacle avoidance trajectory of the multi-machine hoisting system is generated as follows: Figure 6 As shown, Figure 6 The blue curve in the middle is the optimized trajectory of the heavy object, and the three green trajectories are the optimized trajectories of the drone. Figure 7 This is a schematic diagram of the obstacle avoidance trajectory optimized for the multi-machine hoisting system. Both the heavy objects and the drone trajectories can flexibly avoid obstacles. In the actual flight experiment, the maximum acceleration of the drone in the trajectory was 14m / s. 2 , the maximum speed is 9m / s, the maximum axis angle (the angle between the z-axis of the body and the z-axis of the world system) , the total optimization time is 37.4ms.
[0167] Corresponding to the aforementioned embodiment of the method for real-time planning of high-speed obstacle avoidance trajectories for a multi-machine hoisting system, the present application also provides an embodiment of a device for real-time planning of high-speed obstacle avoidance trajectories for a multi-machine hoisting system.
[0168] Figure 8 The following is a block diagram of a high-speed obstacle avoidance trajectory real-time planning device for a multi-machine hoisting system according to an exemplary embodiment. Figure 8 , the device may include:
[0169] A dynamic modeling module 21 is used to perform dynamic modeling on a multi-machine hoisting system, wherein the multi-machine hoisting system includes a plurality of hoisting drones and a heavy object connected to each hoisting drone via a rope;
[0170] A collision-free path planning module 22, used to construct the configuration of the multi-machine hoisting system into a retractable regular pyramid, and to quickly plan an initial collision-free path based on the configuration;
[0171] The position characterization module 23 is used to indirectly characterize the position of the suspended UAV through the position of the weight and the vector direction of the rope;
[0172] The constraint elimination module 24 is used to convert the optimization variable representing the rope direction angle with constraints into an unconstrained virtual variable, so as to eliminate the rope direction angle constraint;
[0173] An optimization problem construction module 25 is used to construct a multi-objective optimization problem for high-speed obstacle avoidance trajectory optimization of a multi-machine hoisting system, wherein the multi-objective optimization problem includes an energy function representing the smoothness of the trajectory and the required energy, a total time function representing the trajectory execution time, fine dynamics constraints, safety constraints, and coupled dynamics constraints;
[0174] The trajectory optimization module 26 is used to perform trajectory optimization based on the multi-objective optimization problem and take the collision-free path as the initial trajectory, wherein the energy function characterizing the smoothness of the trajectory and the required energy, the fine dynamics constraints, the safety constraints and the coupled dynamics constraints are solved in parallel.
[0175] Regarding the device in the above embodiment, the specific manner in which each module performs operations has been described in detail in the embodiment of the method, and will not be elaborated here.
[0176] For the device embodiment, since it basically corresponds to the method embodiment, the relevant parts can refer to the partial description of the method embodiment. The device embodiment described above is only schematic, wherein the units described as separate components may or may not be physically separated, and the components displayed as units may or may not be physical units, that is, they may be located in one place, or they may be distributed on multiple network units. Some or all of the modules may be selected according to actual needs to achieve the purpose of the present application scheme. A person of ordinary skill in the art can understand and implement it without paying any creative work.
[0177] Correspondingly, the present application also provides a computer program product, including a computer program / instruction, which, when executed by a processor, implements the real-time planning method for high-speed obstacle avoidance trajectories of a multi-machine hoisting system as described above.
[0178] Accordingly, the present application also provides an electronic device, including: one or more processors; a memory for storing one or more programs; when the one or more programs are executed by the one or more processors, the one or more processors implement the high-speed obstacle avoidance trajectory real-time planning method for the multi-machine hoisting system as described above. Fig. 9 As shown, it is a hardware structure diagram of a device with data processing capability in which a high-speed obstacle avoidance trajectory real-time planning device for a multi-machine hoisting system provided by an embodiment of the present invention is located, except Fig. 9In addition to the processor, memory and network interface shown, any device with data processing capability in which the apparatus in the embodiment is located may also include other hardware, generally based on the actual functions of the device with data processing capability, which will not be described in detail.
[0179] Accordingly, the present application also provides a computer-readable storage medium on which computer instructions are stored, and when the instructions are executed by the processor, the high-speed obstacle avoidance trajectory real-time planning method for the multi-machine hoisting system as described above is implemented. The computer-readable storage medium can be an internal storage unit of any device with data processing capabilities described in any of the aforementioned embodiments, such as a hard disk or memory. The computer-readable storage medium can also be an external storage device, such as a plug-in hard disk, a smart memory card (Smart Media Card, SMC), an SD card, a flash card (Flash Card), etc. equipped on the device. Furthermore, the computer-readable storage medium can also include both an internal storage unit and an external storage device of any device with data processing capabilities. The computer-readable storage medium is used to store the computer program and other programs and data required by any device with data processing capabilities, and can also be used to temporarily store data that has been output or is to be output.
[0180] Those skilled in the art will readily appreciate other embodiments of the present application after considering the specification and practicing the contents disclosed herein. The present application is intended to cover any variations, uses or adaptations of the present application, which follow the general principles of the present application and include common knowledge or customary technical means in the art that are not disclosed in the present application.
Claims
1. A real-time planning method for high-speed obstacle avoidance trajectories of a multi-machine hoisting system, characterized in that: include: Performing dynamic modeling on a multi-machine hoisting system, wherein the multi-machine hoisting system includes a plurality of hoisting drones and a heavy object connected to each of the hoisting drones via a rope; The configuration of the multi-machine hoisting system is constructed as a retractable regular pyramid, and a collision-free path is quickly planned based on the configuration; The position of the suspended drone is indirectly represented by the position of the weight and the vector direction of the rope; The optimized variables representing the rope direction angle with constraints are converted into unconstrained dummy variables, so as to eliminate the rope direction angle constraints; Construct a multi-objective optimization problem for high-speed obstacle avoidance trajectory optimization of a multi-machine hoisting system, wherein the multi-objective optimization problem includes an energy function that characterizes the smoothness of the trajectory and the required energy, a total time function that characterizes the trajectory execution time, fine dynamics constraints, safety constraints, and coupled dynamics constraints; Based on the multi-objective optimization problem, trajectory optimization is performed with the collision-free path as the initial trajectory, wherein the energy function characterizing the smoothness of the trajectory and the required energy, the fine dynamics constraint, the safety constraint and the coupled dynamics constraint are solved in parallel; The fine dynamics constraints include a maximum speed constraint violation function for the suspended UAV , maximum / minimum thrust constraint violation function , maximum axis angle constraint violation function , angular rate constraint violation function , angular acceleration constraint violation function , where i represents the i-th suspended UAV, j represents the j-th segment of the trajectory, and k represents the k-th sampling point on the j-th segment of the trajectory. is the speed of the mounted UAV in the world coordinate system, is the maximum permissible speed of the suspended drone, is the total thrust of the suspended UAV, are the maximum and minimum allowable thrust, respectively. is the angle between the z-axis of the drone and the direction of gravity acceleration, , is the maximum allowable shaft angle, is the angular velocity of the mounted UAV, is the maximum permissible angular velocity, is the angular acceleration of the mounted UAV, is the maximum permissible angular acceleration; The coupled dynamics constraints include a heavy object translation dynamics constraint violation function and the mass rotational dynamics constraint violation function ,in are the mass of the weight and its moment of inertia about its center of mass, is the velocity of the mass in the world coordinate system, is the tension exerted by the i-th rope on the weight, is the unit vector of the rope direction, , is the angular velocity of the weight, is the angular acceleration of the mass, is the antisymmetric matrix corresponding to the vector between the center of mass and the point where the rope is connected, is the rotation matrix between the weight coordinate system and the world coordinate system, is the number of mounted drones.
2. The method according to claim 1, characterized in that: The multi-machine lifting system includes N multi-rotor drones as lifting drones. The dynamic model of the heavy objects in the system is as follows: , in are the mass of the weight and its moment of inertia about its center of mass, are the position and velocity of the weight in the world coordinate system, is the rotation matrix between the weight coordinate system and the world coordinate system, is the angular velocity of the weight, for The corresponding antisymmetric matrix is, is the tension exerted by the i-th rope on the weight, is the unit vector of the rope direction, pointing from the weight to the drone, is the vector between the center of mass and the point where the rope is connected, pointing from the center of mass to the point where the rope is connected. is the antisymmetric matrix corresponding to the vector, ; The dynamic model of a single suspended UAV is: , The subscript Representing the 10000 drones, including are the mass of the mounted drone and its moment of inertia around the center of mass, are the position and speed of the mounted UAV in the world coordinate system, is the rotation matrix between the body axis system of the mounted UAV and the world coordinate system, for The corresponding quaternion is, is the angular velocity of the mounted UAV, for The corresponding antisymmetric matrix is, are the total thrust and attitude control torque of the suspended UAV, is the tension exerted by the rope on the body. is the unit vector of the rope direction, pointing from the weight to the drone, and from the pitch angle of the rope and azimuth The only certainty is , in , is the tension of the rope on the suspended drone.
3. The method according to claim 1, characterized in that The position of the heavy object , describing the elevation angle of the regular pyramid space posture , azimuth and the radius of the circumscribed circle of the regular triangle at the base of the pyramid As the configuration parameter of the multi-machine lifting system.
4. The method according to claim 3, characterized in that The collision-free path planning is based on the A* algorithm, in which the configuration parameters of the pyramid are adjusted during the path search. Discretize the configuration parameters and use them as a search variable in the algorithm. And the shape parameters of the regular pyramid To calculate the position of the system as a whole in space, and thus determine whether it collides with obstacles in the map.
5. The method according to claim 1, characterized in that The position of the suspended drone is indirectly characterized by the position of the weight and the vector direction of the rope, specifically: , in is the position of the i-th mounted UAV, is the position of the weight, is the rotation matrix between the weight coordinate system and the world coordinate system, is the unit vector representing the direction of the i-th rope, pointing from the weight to the hanging drone, is the vector between the center of mass and the point where the rope is connected, pointing from the center of mass to the point where the rope is connected. is the length of the rope.
6. The method according to claim 1, characterized in that The optimized variable representing the rope direction angle with constraints is equivalently converted into an unconstrained dummy variable, where the pitch angle The equivalent conversion process is: Introducing dummy optimization variables ; Create virtual optimization variables and actual rope direction angle variables Diffeomorphisms between : , in is the pitch angle The upper and lower bounds of Thus, the constrained pitch angle is equivalently converted into an unconstrained dummy variable.
7. The method according to claim 1, characterized in that In the multi-objective optimization problem, The energy function characterizing the smoothness of the trajectory and the required energy ,in, is the weight position, let , but The trajectory of the force vector optimization variable of the rope connected to the i-th suspended drone, and are the pitch angle and azimuth angle representing the rope force vector, To adjust the coefficient of the two energy functions, T is the total duration of the trajectory, The number of drones being hoisted; The total time function representing the execution time of the trajectory ,in, is the first segment of the multi-segment polynomial trajectory The duration of the segment trajectory; The safety constraints include a system obstacle avoidance safety constraint violation function and the mutual collision avoidance constraint violation function within the multi-machine hoisting system ,in is the number of sampling points on the rope, is the maximum allowable safe distance from obstacles, is the kth sampling point on the jth trajectory of the heavy object, is the Euclidean signed distance field function, is the length of the rope.
8. A high-speed obstacle avoidance trajectory real-time planning device for a multi-machine hoisting system, characterized in that: include: A dynamic modeling module, used for dynamic modeling of a multi-machine hoisting system, wherein the multi-machine hoisting system includes a plurality of hoisting drones and a heavy object connected to each hoisting drone via a rope; A collision-free path planning module, used to construct the configuration of the multi-machine hoisting system into a scalable regular pyramid, and to perform rapid planning of a collision-free path based on the configuration; The position representation module is used to indirectly represent the position of the suspended drone through the position of the weight and the vector direction of the rope; The constraint elimination module is used to convert the optimization variables representing the rope direction angle with constraints into unconstrained virtual variables, so as to eliminate the rope direction angle constraints; An optimization problem construction module is used to construct a multi-objective optimization problem for high-speed obstacle avoidance trajectory optimization of a multi-machine hoisting system, wherein the multi-objective optimization problem includes an energy function representing the smoothness of the trajectory and the required energy, a total time function representing the trajectory execution time, fine dynamics constraints, safety constraints, and coupled dynamics constraints; A trajectory optimization module, configured to perform trajectory optimization based on the multi-objective optimization problem and taking the collision-free path as the initial trajectory, wherein the energy function characterizing the smoothness of the trajectory and the required energy, the fine dynamics constraint, the safety constraint and the coupled dynamics constraint are solved in parallel; The fine dynamics constraints include a maximum speed constraint violation function for the suspended UAV , maximum / minimum thrust constraint violation function , maximum axis angle constraint violation function , angular rate constraint violation function , angular acceleration constraint violation function , where i represents the i-th suspended UAV, j represents the j-th segment of the trajectory, and k represents the k-th sampling point on the j-th segment of the trajectory. is the speed of the mounted UAV in the world coordinate system, is the maximum permissible speed of the suspended drone, is the total thrust of the suspended UAV, are the maximum and minimum allowable thrust, respectively. is the angle between the z-axis of the drone and the direction of gravity acceleration, , is the maximum allowable shaft angle, is the angular velocity of the mounted UAV, is the maximum permissible angular velocity, is the angular acceleration of the mounted UAV, is the maximum permissible angular acceleration; The coupled dynamics constraints include a heavy object translation dynamics constraint violation function and the mass rotational dynamics constraint violation function ,in are the mass of the weight and its moment of inertia about its center of mass, is the velocity of the mass in the world coordinate system, is the tension exerted by the i-th rope on the weight, is the unit vector of the rope direction, , is the angular velocity of the weight, is the angular acceleration of the mass, is the antisymmetric matrix corresponding to the vector between the center of mass and the point where the rope is connected, is the rotation matrix between the weight coordinate system and the world coordinate system, is the number of mounted drones.
9. An electronic device, characterized in that: include: one or more processors; A memory for storing one or more programs; When the one or more programs are executed by the one or more processors, the one or more processors implement the method according to any one of claims 1 to 7.
10. A computer-readable storage medium having computer instructions stored thereon, characterized in that: When the instruction is executed by a processor, the steps of the method according to any one of claims 1 to 7 are implemented.
