A flight pipeline modeling method for cluster unmanned aerial vehicle control

Abstract

The application discloses a flight pipeline modeling method for cluster unmanned aerial vehicle control, and establishes a geometric pipeline model of unmanned aerial vehicle flight; a profile model of a convex polygon section is established based on rotation angles and lengths of convex vectors, and the convex polygon section is composed of a plurality of convex vectors; nodes are constructed by positions of the convex polygon section on a reference curve, the rotation angles and the lengths of the convex vectors, the geometric pipeline is discretized into a plurality of node sequences, a curve fitting algorithm is applied to the node sequences to generate a continuous three-dimensional pipeline model; a comprehensive cost function of the geometric pipeline is established according to characteristics of the geometric pipeline; and the optimal geometric flight pipeline model is obtained when the comprehensive cost function is the minimum. The flight pipeline is modeled by the reference curve and the convex polygon section, a polygon geometric pipeline is obtained, and safe flight space modeling in a complex three-dimensional scene is realized.

Description

Technical Field

[0001] This invention relates to flight safety control of unmanned aerial vehicles (UAVs), specifically to a flight pipeline modeling method for swarm UAV control. Background Technology

[0002] With the rapid development of drone technology, swarm drones have shown great potential in fields such as environmental monitoring, disaster relief, logistics transportation, formation display, and complex mission collaboration. Swarm drone systems achieve distributed mission execution and efficiency improvement through the coordinated flight of multiple drones. However, when conducting swarm flight in complex environments, it is necessary to consider multiple factors simultaneously, such as safe distance constraints between multiple drones, trajectory smoothness, obstacle avoidance constraints, and spatial collaborative control. Traditional path planning and formation control methods are difficult to meet these constraints in complex environments.

[0003] Currently, various UAV path planning and control methods have been proposed, but these methods are mostly designed for single-aircraft or formation flight, lacking a unified geometric model of the overall flight space. To improve flight safety and coordination, some studies have introduced virtual pipeline models to define the flight path of UAV swarms, ensuring safety during swarm flight through pipeline constraints. However, most existing virtual pipeline modeling methods use fixed-shape cross-sections or simple cylindrical cross-sections, making it difficult for the shape of the flight pipeline to adapt to changes in complex three-dimensional environments. Summary of the Invention

[0004] Purpose of the invention: To address the above shortcomings, this invention provides a flight pipeline modeling method for swarm UAV control that is adapted to complex environments.

[0005] Technical Solution: To solve the above problems, this invention adopts a flight pipeline modeling method for swarm UAV control, including the following steps:

[0006] (1) Establish a geometric pipeline model for UAV flight, wherein the geometric pipeline model includes a reference curve and several convex polygonal cross sections distributed along the reference curve;

[0007] (2) Based on the rotation angle and length of the convex vector, establish the contour model of the convex polygon cross section. The convex polygon cross section is composed of several convex vectors. The convex vectors are calculated by the basis vectors through the Rodriguez rotation formula. The basis vectors take the position of the convex polygon cross section on the reference curve as the starting point and the vertex of the convex polygon cross section as the ending point.

[0008] (3) Nodes are constructed using the position of the convex polygon cross section on the reference curve, the rotation angle and length of the convex vector. The geometric pipe is discretized into several node sequences. A curve fitting algorithm is applied to the node sequences to generate a continuous three-dimensional pipe model.

[0009] (4) In combination with the flight environment of the clustered UAVs, establish the comprehensive cost function of the geometric pipeline, which includes length cost, cross-sectional area cost, smoothness cost and safety cost;

[0010] (5) The optimal geometric flight pipeline model is obtained when the comprehensive cost function is minimized.

[0011] Furthermore, the geometric pipeline model for the drone's flight is as follows:

[0012] ;

[0013] in, For time, The start time of the drone's operation in the geometric pipe. The termination time of the drone's operation within the geometric pipe. As the baseline curve, Let be the set of lengths of the basis vectors. The set of angles of the basis vectors. A contour model representing a convex polygon cross section.

[0014] Furthermore, the contour model of the convex polygonal cross-section is as follows:

[0015] ;

[0016] ;

[0017] ;

[0018] in, Parameters describing the location of any point within the pipe cross-section, For the first A convex vector The number of convex vectors Angles of basis vectors For the vector that needs to be rotated, Let be the unit tangential vector of the reference curve. The unit normal vector of the reference curve. is the length of the basis vector.

[0019] Furthermore, the comprehensive cost function includes a length cost that minimizes the length of the pipeline reference curve, and the expression for the length cost is as follows:

[0020] ;

[0021] in, As a reference curve The first derivative, The L2 norm of a vector.

[0022] Furthermore, the comprehensive cost function also includes a cross-sectional area cost that maximizes the internal passable space of the pipeline, expressed as follows:

[0023] ;

[0024] in, This represents the theoretical maximum value of the cross-section. Represents a convex polygon cross section The size of the area.

[0025] Furthermore, the comprehensive cost function also includes a smoothing cost to make the pipeline baseline curve as smooth as possible, the expression for which is:

[0026] ;

[0027] in, As a reference curve The second derivative, Let represent the cube of the 2-norm of a vector.

[0028] Furthermore, the comprehensive cost function also includes a safety cost that ensures the pipeline is collision-free, and the expression for the safety cost is:

[0029] ;

[0030] in, The penalty distance between the geometric pipe and the obstacle. This represents the actual distance between the geometric pipe and the obstacle.

[0031] Furthermore, the comprehensive cost function is:

[0032] ;

[0033] in, There are four positive coefficients.

[0034] The present invention provides a computer device, including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the computer program to implement the steps of the above-described method.

[0035] The present invention provides a computer-readable storage medium having a computer program stored thereon, wherein the computer program, when executed by a processor, implements the steps of the above-described method.

[0036] Beneficial effects: Compared with the prior art, the significant advantage of this invention is that it models the flight pipe by using a reference curve and a convex polygonal cross section to obtain a polygonal geometric pipe, thereby realizing safe flight space modeling in complex three-dimensional scenes. At the same time, it considers spatial constraints, smoothness and environmental obstacle information, thus providing a safe, smooth and efficient geometric flight space modeling framework for the cooperative flight and path planning of swarm UAVs. Attached Figure Description

[0037] Figure 1 This is a schematic diagram of the modeling method of the present invention.

[0038] Figure 2 This is a rendering of the geometric flight pipe constructed for this invention. Detailed Implementation

[0039] like Figure 1 As shown in the figure, this embodiment presents a flight pipeline modeling method for swarm UAV control, which includes the following steps:

[0040] Step 1: Establish a geometric pipeline model for UAV flight. The geometric pipeline model includes a reference curve and several convex polygonal cross sections distributed along the reference curve. The reference curve is used to describe the reference position of the geometric flight pipeline.

[0041] Step 2: Generate a set of convex vectors for the cross section based on the rotation formula to determine the shape of the convex polygon cross section at different positions on the reference curve and establish the contour model of the convex polygon cross section.

[0042] Step 3: Construct nodes based on the position of the convex polygon cross section on the reference curve, the rotation angle and length of the convex vector, discretize the geometric pipe into several node sequences, and apply a curve fitting algorithm to the node sequences to generate a continuous three-dimensional pipe model.

[0043] Step 4: Based on the flight environment of the swarm of UAVs, establish a comprehensive cost function for the geometric pipeline. The cost function includes length cost, cross-sectional area cost, smoothness cost, and safety cost.

[0044] Step 5: Optimize the discrete node sequence, calculate the minimum value of the cost function, and obtain the optimal geometric flight path that satisfies both flight smoothness and safety.

[0045] Specifically, establishing a geometric pipeline model for swarm drone flight includes the following steps:

[0046] Geometric Flight Tube It is a continuous baseline curve The spatial set of convex polygonal cross-sections distributed along this curve is used to describe the spatial corridor structure for UAV swarm flight. The geometric flight channel can be represented as:

[0047]

[0048] in, For time, The start time, For the end time, Let be the set of lengths of the pipe along the directions of each vertex of the convex polygon. Let be the set of rotation angles of the pipe along the directions of each vertex of the convex polygon. Let be the convex polygonal cross-section of the pipe. The unit tangential vector of the reference curve is defined as... The unit normal vector is defined as auxiliary vector , for Components in the three-axis directions.

[0049] This geometric pipe model, through the joint definition of continuous curves and local cross sections, can effectively reflect the passable space for swarm drones to fly.

[0050] Specifically, the steps for generating the set of convex vectors of the cross section based on the rotation formula are as follows:

[0051] The convex polygonal cross-section of a pipe exists with a reference curve as its reference position. The convex polygonal cross-section of the pipe is perpendicular to the tangential direction of the reference curve, and the reference point of the convex polygon lies on the reference curve. The convex polygonal cross-section of the pipe is composed of... The set of lengths of the convex vectors is composed of basis vectors, which are calculated from the basis vectors using the Rodrigues rotation formula. The set of rotation angles of a convex vector is The angles in the set of rotation angles rotate counterclockwise, satisfying the following condition: Using the Rodriguez rotation formula, the first... Convex vectors Represented as:

[0052]

[0053] in, For the axis of rotation, For rotation angle, Let be the vector that needs to be rotated.

[0054] At this point, the convex polygonal cross-section of the pipe is defined as:

[0055]

[0056] in, Parameters describing the location of any point within the cross-section of the pipe.

[0057] The cross section is perpendicular to the tangential direction of the reference curve, and its reference point is located on the curve, thus ensuring spatial continuity and shape variability.

[0058] Specifically, the steps for geometric pipe discretization and curve fitting are as follows:

[0059] To fully describe the geometric flight pipe with a finite number of parameters, using A set of nodes This generates the pipeline. Each node contains position information, a length set, and an angle set. Each node is represented as:

[0060]

[0061] in, For position vectors, For a set of lengths, For the set of angles, the symbol is... This represents the transpose of a vector. For the first The length of a convex vector For the first The rotation angle of each convex vector. The position and shape of the cross-section are calculated based on the length set, angle set, and reference curve. The node set needs to be interpolated using a curve generation algorithm to generate the curve. Vectors formed by curves Ultimately, it can describe a complete pipeline model.

[0062] Specifically, the steps for establishing the geometric pipeline synthesis cost function are as follows:

[0063] To achieve comprehensive optimization of path smoothness, safety, and space utilization, the cost function of the geometric flight tunnel consists of length cost, cross-sectional area cost, smoothness cost, and safety cost.

[0064] Length cost The goal is to minimize the length of the pipeline reference curve, which is obtained by calculating the length of the pipeline reference curve, as shown below:

[0065]

[0066] in, As a reference curve The first derivative, sign The L2 norm of a vector.

[0067] Cross-sectional area cost The goal is to maximize the passable space inside the pipe, which can be achieved by calculating the area of ​​the convex polygon cross-section. The formula for calculating the cross-sectional area cost is shown below:

[0068]

[0069] in, Representing geometric cross section The size of the area, This represents the theoretical maximum value of the cross section.

[0070] Smoothing Cost The design aims to make the pipeline baseline curve as smooth as possible, facilitating the flight of swarm drones. The formula for calculating the smoothness cost is as follows:

[0071]

[0072] in, As a reference curve The second derivative of , where the superscript 3 indicates the cube of the value.

[0073] Security Cost The design aims to maintain a safe distance between the pipeline and obstacles. The safety cost is calculated as follows:

[0074]

[0075] in, The penalty distance between the designed pipe and the obstacle. This represents the actual distance between the pipe and the obstacle.

[0076] The overall cost function of the geometric flight tunnel is formed by the weighted sum of the above four terms, and can be expressed as:

[0077]

[0078] in, There are four positive coefficients.

[0079] By discretizing the time domain, the integral cost function is transformed into a computable discrete summation form. The node sequence is then searched and optimized using swarm intelligence algorithms or other numerical optimization algorithms to obtain the optimal solution. The resulting geometric flight pipeline enables efficient cooperative flight of swarm drones while ensuring path smoothness and safety. In this embodiment, five convex vectors are used as an example to construct the following... Figure 2 The geometric pipeline diagram shown depicts a flight pipeline constructed with four pentagonal sections 2 positioned relative to the reference curve 1.

[0080] The modeling method in this embodiment overcomes the problems of insufficient spatial constraint representation, poor path smoothness, and difficulty in ensuring safety in existing UAV swarm flight path planning. It proposes a geometric flight pipeline modeling method for swarm UAVs. This method achieves safe, smooth, and coordinated flight of UAV swarms in complex spaces by establishing a geometric pipeline model, constructing deformable sections, discretizing geometric parameters, and optimizing based on a cost function.

Claims

1. A flight pipeline modeling method for swarm UAV control, characterized in that, Includes the following steps: (1) Establish a geometric pipeline model for UAV flight, wherein the geometric pipeline model includes a reference curve and several convex polygonal cross sections distributed along the reference curve; (2) Based on the rotation angle and length of the convex vector, establish the contour model of the convex polygon cross section. The convex polygon cross section is composed of several convex vectors. The convex vectors are calculated by the basis vectors through the Rodriguez rotation formula. The basis vectors take the position of the convex polygon cross section on the reference curve as the starting point and the vertex of the convex polygon cross section as the ending point. (3) Nodes are constructed using the position of the convex polygon cross section on the reference curve, the rotation angle and length of the convex vector. The geometric pipe is discretized into several node sequences. A curve fitting algorithm is applied to the node sequences to generate a continuous three-dimensional pipe model. (4) In combination with the flight environment of the clustered UAVs, establish the comprehensive cost function of the geometric pipeline, which includes length cost, cross-sectional area cost, smoothness cost and safety cost; (5) The optimal geometric flight pipeline model is obtained when the comprehensive cost function is minimized.

2. The flight pipeline modeling method for swarm UAV control according to claim 1, characterized in that, The geometric pipeline model for the drone's flight is as follows: ； in, For time, The start time of the drone's operation in the geometric pipe. The termination time of the drone's operation within the geometric pipe. As the baseline curve, Let be the set of lengths of convex vectors. Let be the set of rotation angles of a convex vector. A contour model representing a convex polygon cross section.

3. The flight pipeline modeling method for swarm UAV control according to claim 2, characterized in that, The contour model of the convex polygonal cross section is as follows: ； ； ； in, Parameters describing the location of any point within the pipe cross-section, For the first A convex vector The number of convex vectors. Let be the rotation angle of the convex vector. For the vector that needs to be rotated, Let be the unit tangential vector of the reference curve. The unit normal vector of the reference curve. is the length of the basis vector.

4. The flight pipeline modeling method for swarm UAV control according to claim 3, characterized in that, The comprehensive cost function includes a length cost that minimizes the length of the pipeline baseline curve, and the expression for the length cost is as follows: ； in, As a reference curve The first derivative, The L2 norm of a vector is used to represent the vector's L2 norm.

5. The flight pipeline modeling method for swarm UAV control according to claim 4, characterized in that, The comprehensive cost function also includes a cross-sectional area cost that maximizes the internal passable space of the pipeline, and the expression for the cross-sectional area cost is as follows: ； in, This represents the theoretical maximum value of the cross-section. Represents a convex polygon cross section The size of the area.

6. The flight pipeline modeling method for swarm UAV control according to claim 5, characterized in that, The comprehensive cost function also includes a smoothing cost to make the pipeline baseline curve as smooth as possible. The expression for the smoothing cost is as follows: ； in, As a reference curve The second derivative, Let represent the cube of the 2-norm of a vector.

7. The flight pipeline modeling method for swarm UAV control according to claim 6, characterized in that, The comprehensive cost function also includes a safety cost that prevents pipeline collisions, and the expression for the safety cost is: ； in, The penalty distance between the geometric pipe and the obstacle. This represents the actual distance between the geometric pipe boundary and the obstacle.

8. The flight pipeline modeling method for swarm UAV control according to claim 7, characterized in that, The comprehensive cost function is: ； in, There are four positive coefficients.

9. A computer device, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that, When the processor executes the computer program, it implements the steps of the method according to any one of claims 1 to 8.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by a processor, it implements the steps of the method according to any one of claims 1 to 8.

Images