Low-complexity obstacle avoidance method and system for quad-rotor unmanned aerial vehicle

Abstract

The invention relates to the technical field of unmanned aerial vehicle navigation, in particular to a low-complexity obstacle avoidance method and system for a quad-rotor unmanned aerial vehicle, and the method comprises the steps: representing the geometric shapes of the quad-rotor unmanned aerial vehicle and an obstacle as a union set of convex geometries; the obstacle avoidance requirement between the unmanned aerial vehicle and the obstacle can be converted into a mathematical problem that the intersection between two union sets is an empty set, and the implicit obstacle avoidance mathematical expression with the intersection being the empty set is converted into continuous differentiable explicit obstacle avoidance inequality constraints by further using the hyperplane separation theorem, so that efficient obstacle avoidance between the unmanned aerial vehicle and the obstacle is realized. According to the method, the separation hyperplane between the unmanned aerial vehicle and the obstacle is creatively and directly constructed, the constructed obstacle avoidance inequality is micro-continuous and can be directly adapted to a universal nonlinear solver, the efficiency of the unmanned aerial vehicle for rapidly planning the flight path is effectively improved, the accurate and real-time obstacle avoidance requirement of the unmanned aerial vehicle can be met, and the method is suitable for large-scale popularization and application. Practical technical support is provided for rapid flight path planning of the quad-rotor unmanned aerial vehicle.

Description

Technical Field

[0001] This invention relates to the field of unmanned aerial vehicle (UAV) navigation technology, and in particular to a low-complexity obstacle avoidance method and system for quadcopter UAVs. Background Technology

[0002] In recent years, the low-altitude economy has flourished, becoming an important engine for industrial upgrading and technological innovation. Against this backdrop, quadcopter drones, with their advantages of flexibility, maneuverability, and vertical takeoff and landing, have shown broad application prospects in logistics, emergency rescue, and agricultural plant protection. However, as application scenarios continue to expand, the obstacle avoidance capability of drones in complex and confined environments has become a key bottleneck restricting their large-scale deployment. Researching precise and rapid obstacle avoidance technology for quadcopter drones in confined environments can not only improve their mission execution capabilities in restricted spaces such as mine shafts and indoor warehouses, but also provide core technological support for the safe and efficient operation of the low-altitude economy.

[0003] In real-world applications, quadcopter drones often need to perform tasks in complex and ever-changing environments, such as urban canyons, indoor warehouses, or dense jungles. These spaces are not only narrow but also filled with irregular dynamic obstacles (such as pedestrians, vehicles, and building structures), making safe trajectory planning for drones extremely challenging. Traditional methods rely on prior environmental information or static obstacle avoidance strategies, which are difficult to adapt to dynamic changes. Therefore, how to achieve rapid and accurate obstacle avoidance for drones in confined spaces has become a key issue.

[0004] Currently, methods for solving the obstacle avoidance motion planning problem for unmanned aerial vehicles (UAVs) mainly include graph search methods, sampling methods, machine learning methods (such as reinforcement learning), and numerical optimization methods. Among these, numerical optimization methods have become an important choice for high-precision obstacle avoidance planning due to their advantages such as accurate modeling, controllable constraints, and flexible optimization. However, the main challenges faced by these methods are high computational complexity and insufficient real-time performance, making it difficult to meet the rapid response requirements of UAVs in dynamic environments.

[0005] In numerical optimization frameworks, obstacle avoidance problems are typically modeled as constrained optimization problems, where the distance constraint between the UAV and obstacles is crucial. Traditional methods ensure obstacle avoidance by solving distance optimization problems, such as ensuring the distance is non-negative, but this is computationally expensive and difficult to meet real-time requirements. To improve efficiency, some researchers use signed distance functions and leverage the strong duality of convex optimization to reconstruct obstacle avoidance constraints, making them continuously differentiable and easier to solve. Furthermore, methods based on Farkas' lemma have also been used to construct explicit obstacle avoidance constraints, avoiding direct distance calculation. However, these methods still have significant bottlenecks. Both methods based on duality theory and Farkas' lemma require the introduction of numerous auxiliary variables, leading to an expansion of the optimization problem size. Moreover, the more complex the obstacle shape (e.g., a non-convex polyhedron), the more variables are required, further reducing solution efficiency. In dynamic environments, planning algorithms need millisecond-level responses, while existing optimization methods, due to their excessive computational burden, struggle to meet the demands of fast obstacle avoidance.

[0006] Therefore, how to reduce the complexity of the optimization problem and improve computational efficiency while ensuring obstacle avoidance accuracy has become a key research direction for quadcopter UAVs to achieve fast and accurate obstacle avoidance in narrow environments. Summary of the Invention

[0007] Therefore, the technical problem to be solved by the present invention is that the computational complexity of existing UAV obstacle avoidance methods is too high, which leads to a decrease in solution efficiency and makes it difficult to meet the requirements of real-time dynamic obstacle avoidance.

[0008] To address the aforementioned technical problems, this invention provides a low-complexity obstacle avoidance method for quadcopter drones, comprising: Construct geometric representations of drones and obstacles; Obstacle avoidance conditions for drones are constructed based on the geometric representations of drones and obstacles; By using the separating hyperplane to represent the obstacle avoidance conditions of the UAV, the obstacle avoidance inequality of the UAV is obtained; The system acquires the real-time position information of the drone and obstacles, solves for the optimal separating hyperplane based on the drone obstacle avoidance inequality, and constructs a path planning model for the drone based on the optimal separating hyperplane. Solve the path planning model to generate control commands for the UAV.

[0009] Preferably, constructing a geometric representation of the drone and obstacles includes: The drone and obstacles are represented by polyhedra, and decomposed into a union of convex geometries.

[0010] Preferably, the convex geometry is represented as follows: The convex geometry can be represented using linear matrix inequalities, with the following formula: ; in, For the convex geometry of the drone, For the index of the convex geometry of the UAV, The convex geometry of the obstacle For the index of the convex geometry of the obstacle, Let be the coordinate vector of a point in three-dimensional space. Let be the constraint matrix of the convex geometry of the UAV. This serves as the upper bound of the constraint on the convex geometry of the UAV. Let be the constraint matrix of the convex geometry of the obstacle. It is the upper bound of the constraint on the convex geometry of the obstacle.

[0011] Preferably, the obstacle avoidance conditions for the UAV are constructed based on the geometric representations of the UAV and the obstacles, including: Based on the convex decomposition relationship, the obstacle avoidance condition for UAVs is that no two convex geometries of the UAV and the obstacle intersect.

[0012] Preferably, the obstacle avoidance conditions of the UAV are represented using a separating hyperplane, including: Construct a separating hyperplane to transform the disjointness between any two convex geometries of the UAV and the obstacle into the following: the vertex position matrix of all UAV convex geometries is located in one half-space obtained by the separating hyperplane, and the vertex position matrix of all obstacle convex geometries is located in the other half-space. Based on the separated hyperplane, the obstacle avoidance inequality for UAVs is derived: ; in, To separate the normal vectors of the hyperplane, This represents the position offset of the separating hyperplane. It is a row vector where all elements have a value of 1. Let be the vertex position matrix of the convex geometry of the UAV. For the index of the convex geometry of the UAV, Let be the vertex position matrix of the convex geometry of the obstacle. The index of the convex geometry of the obstacle.

[0013] Preferably, the method for acquiring the location information of the drone and obstacles in real time is as follows: The environmental point cloud data is acquired in real time using airborne sensors, and the vertex position matrix of the obstacle is obtained after filtering. Calculate the vertex position matrix of the UAV based on the UAV attitude estimation values.

[0014] Preferably, the optimal separating hyperplane is solved based on the UAV obstacle avoidance inequality, including: Based on the obstacle avoidance inequality for UAVs, the optimal separating hyperplane can be solved using the following formula: ; in, and These are all parameters of the separating hyperplane. Let be the vertex position matrix of the drone. Let be the vertex position matrix of the obstacle. This is the safety threshold.

[0015] Preferably, a path planning model for the UAV is constructed based on the optimal separating hyperplane, using the following method: With trajectory smoothness and energy consumption minimization as objective functions, a path planning model for the UAV is constructed based on the optimal separating hyperplane, as expressed by the following equation: ; ; in, Let t be the position of the drone at time t. Total time for The first derivative, for The second derivative, These are the weighting coefficients. and These are all parameters of the optimal separating hyperplane. Let be the vertex position matrix of the convex geometry of the UAV. Let be the vertex position matrix of the obstacle. This is the safety threshold.

[0016] Preferably, the method for solving the path planning model is as follows: the path planning model is solved using a sequential quadratic programming solver.

[0017] The present invention also provides a low-complexity obstacle avoidance system for a quadcopter unmanned aerial vehicle, comprising: The geometry building block is used to construct geometric representations of drones and obstacles. The obstacle avoidance condition construction module is used to construct obstacle avoidance conditions for the UAV based on the geometric representations of the UAV and obstacles. An obstacle avoidance inequality establishment module is used to represent the obstacle avoidance conditions of the UAV using a separating hyperplane, and to obtain the UAV obstacle avoidance inequality. The path planning modeling module is used to acquire the position information of the UAV and obstacles in real time, solve for the optimal separating hyperplane based on the UAV obstacle avoidance inequality, and construct the path planning model of the UAV based on the optimal separating hyperplane. The solution module is used to solve the path planning model and generate control commands for the UAV.

[0018] Compared with the prior art, the above-described technical solution of the present invention has the following advantages: This invention presents a low-complexity obstacle avoidance method for quadrotor drones, addressing the challenge of accurate and rapid obstacle avoidance in confined environments such as mines. First, the geometry of the quadrotor drone and the obstacle is represented as the union of convex geometric bodies. The obstacle avoidance requirement between the drone and the obstacle can be transformed into a mathematical problem where the intersection of the two unions is an empty set. To meet the requirement of differentiable inequality constraints in the optimization problem for general nonlinear solvers, this invention further uses the hyperplane separation theorem to transform the implicit obstacle avoidance mathematical expression of the empty intersection into a continuously differentiable explicit obstacle avoidance inequality constraint, achieving efficient obstacle avoidance between the drone and the obstacle. This invention innovatively uses the hyperplane separation theorem to directly construct the separating hyperplane between the drone and the obstacle. The constructed obstacle avoidance inequality is differentiable and continuous, directly adaptable to general nonlinear solvers, and effectively improves the efficiency of rapid drone trajectory planning. It enables precise and real-time obstacle avoidance for quadrotor drones, providing practical technical support for rapid trajectory planning of quadrotor drones. Attached Figure Description

[0019] To make the content of this invention easier to understand, the invention will be further described in detail below with reference to specific embodiments and accompanying drawings, wherein: Figure 1 This is a flowchart of a low-complexity obstacle avoidance method for a quadcopter drone according to the present invention; Figure 2 This is a schematic diagram of the geometric shape of a quadcopter drone, in which... Figure 2 (a) in the image is a 3D oblique view of the drone. Figure 2 (b) in the image is a top view of the drone. Detailed Implementation

[0020] The present invention will be further described below with reference to the accompanying drawings and specific embodiments, so that those skilled in the art can better understand and implement the present invention. However, the embodiments described are not intended to limit the present invention.

[0021] Reference Figure 1 As shown, this invention provides a low-complexity obstacle avoidance method for quadcopter drones, comprising:

[0022] S1: Construct the geometric representation of the drone and obstacles.

[0023] Preferably, this embodiment uses polyhedra to represent the drone and obstacles, and decomposes them into a union of convex geometries, specifically including the following steps:

[0024] S11: Defining Quadcopter Drones and obstacles The geometric shape.

[0025] In real-world scenarios, due to the quadcopter drone (referred to as...) ) and the obstacles around them (denoted as In practical applications, polyhedra are typically non-convex and irregular in shape. Polyhedra possess excellent mathematical properties, and for ease of subsequent modeling and solving, quadcopter drones and obstacles can be approximated using polyhedra.

[0026] S12: According to the principle of convex decomposition in geometry, non-convex geometry can be represented as the union of several convex geometric solids. Furthermore, the polyhedra of the drone and the obstacle can be decomposed into the union of convex geometric solids, as shown in the formula: ; in, This refers to a quadcopter drone. Indicates an obstacle. and All are convex geometries.

[0027] In this embodiment, and All are convex polyhedra and ellipsoids. (Refer to...) Figure 2 As shown, where Figure 2 (a) in the image is a 3D oblique view of the drone. Figure 2 (b) in the diagram is a top view of the UAV. The body of a quadcopter UAV is represented by a cuboid, and the four rotors are represented by ellipsoids. Therefore, a quadcopter UAV can be accurately represented by the union of a cuboid and four ellipsoids.

[0028] S13: Based on the representation method of convex sets, convex geometry can be represented using linear matrix inequalities, the specific formula of which is: ; in, For the convex geometry of the drone, For the index of the convex geometry of the UAV, The convex geometry of the obstacle For the index of the convex geometry of the obstacle, Let be the coordinate vector of a point in three-dimensional space. For the set of real numbers, Let be the constraint matrix of the convex geometry of the UAV. This serves as the upper bound of the constraint on the convex geometry of the UAV. Let be the constraint matrix of the convex geometry of the obstacle. The upper bound of the constraint on the convex geometry of the obstacle. For the convex geometry of the obstacle The number of linear inequality constraints (i.e., the number of faces of a polyhedron). Represents the convex geometry of the drone The number of linear inequality constraints.

[0029] This step involves selecting a convex geometry that can be expressed using explicit mathematical formulas, specifically for quadcopter drones. and obstacles The irregular geometric shape is represented as the union of convex geometric bodies, which facilitates subsequent modeling and calculation.

[0030] S2: Construct obstacle avoidance conditions for the UAV based on the geometric representation of the UAV and obstacles.

[0031] Quadrone drones as defined in S1 and obstacles Based on the polyhedron, the obstacle avoidance conditions for UAVs are constructed, that is, the two polyhedra are studied. and Obstacle avoidance conditions between them.

[0032] S21: The drone must not collide with obstacles during flight. This embodiment uses set theory and mathematical modeling to construct the obstacle avoidance requirement between two polyhedra as a mathematical problem where the intersection of the two sets is an empty set, expressed as: .

[0033] S22: Combining the convex decomposition relation of S12, a non-convex polyhedron. and It can be represented as the union of their respective convex geometries, therefore the obstacle avoidance condition This can be achieved by ensuring that the intersection between their subsets is empty, that is, that no two convex geometries of the drone and the obstacle intersect, as expressed by the formula: ;

[0034] This method decomposes the complex intersection conditions between non-convex polyhedra into an intersection problem between multiple convex geometries, significantly reducing the model complexity.

[0035] S3: Using the separated hyperplane to represent the obstacle avoidance conditions of the UAV, the obstacle avoidance inequalities of the UAV are obtained, including: Construct a separating hyperplane to transform the non-intersection of any two convex geometries of the UAV and the obstacle into the following: the vertex position matrix of all UAV convex geometries is located in one half-space obtained by the separating hyperplane, and the vertex position matrix of all obstacle convex geometries is located in the other half-space. Based on the separated hyperplane, the obstacle avoidance inequality for UAVs is derived: ; in, To separate the normal vectors of the hyperplane, This represents the position offset of the separating hyperplane. It is a row vector where all elements have a value of 1. Let be the vertex position matrix of the convex geometry of the UAV. For the index of the convex geometry of the UAV, Let be the vertex position matrix of the convex geometry of the obstacle. The index of the convex geometry of the obstacle.

[0036] The detailed derivation process of S3 is as follows:

[0037] S31: Obtained from S2 The relationship between two convex geometries was studied. According to the relevant theory of convex optimization, a convex polyhedron can be represented by the convex hull of its vertices. Let... and The vertex position matrices are represented as follows: ; in, and They are respectively and The vertex position matrix.

[0038] S32: According to the hyperplane separation theorem, the necessary and sufficient condition for two closed convex sets to be disjoint is that there exists a separating hyperplane that can separate the two closed convex sets.

[0039] Define the separating hyperplane for: ; in, , To separate the normal vectors of the hyperplane, This represents the position offset of the separating hyperplane.

[0040] This separating hyperplane divides space into two half-spaces. and , represented as: .

[0041] S33: Obstacle Avoidance Conditions It can be constructed and Separating hyperplanes between ,make and Placed separately in a semi-space and To achieve this. Therefore, and The condition for disjointness can be expressed as: .

[0042] S34: Since the convex polyhedron is the convex hull of its vertices, based on convex optimization and the properties of the convex hull, the above obstacle avoidance conditions... By ensuring and Each vertex in half space and This is achieved internally, meaning that if all vertices of a polyhedron are within their respective half-spaces, then the entire polyhedron lies within those half-spaces. Therefore, the obstacle avoidance condition described above can be transformed into: ; in, It is required that there exists such a separating hyperplane. It is a row vector where all elements have a value of 1. .

[0043] S35: In summary, the initial implicit obstacle avoidance conditions for drones This can be replaced by constructing an explicit differentiable inequality, i.e.: .

[0044] S4: The UAV acquires environmental point cloud data in real time, and solves for the optimal separating hyperplane based on the UAV obstacle avoidance inequality; based on the optimal separating hyperplane, a path planning model for the UAV is constructed. The specific steps are as follows:

[0045] S41: Real-time acquisition of the location information of the drone and obstacles, using the following method: Using airborne sensors (such as LiDAR and vision cameras) to acquire environmental point cloud data in real time, the resulting data is filtered to obtain the vertex position matrix composed of multiple convex geometric shapes of the obstacle. (M is the number of vertices of the obstacle); Based on the UAV attitude estimation values, calculate the vertex position matrix formed by multiple convex geometries of the UAV. (N is the number of drone vertices).

[0046] S42: Solving for the optimal separating hyperplane based on the UAV obstacle avoidance inequality, including: Extracting drone data from sensor data Location ,speed and the vertex coordinates of the obstacles Considering the safety distance, based on the UAV obstacle avoidance inequality, the optimal separating hyperplane is solved using the following formula: ; in, and These are all parameters of the separating hyperplane. Let be the vertex position matrix of the drone. Let be the vertex position matrix of the obstacle. This is the safety threshold.

[0047] S43: Based on the optimal separating hyperplane, a path planning model for the UAV is constructed using the following method: With trajectory smoothness and energy consumption minimization as objective functions, a path planning model for the UAV is constructed based on the optimal separating hyperplane, as expressed by the following equation: ; ; in, Let t be the position of the drone at time t. Total time for The first derivative of , i.e., the speed of the drone at time t, represents the energy consumption of the drone. for The second derivative of , i.e., the acceleration of the UAV at time t, represents the smoothness of the UAV's trajectory. These are the weighting coefficients. and These are all parameters of the optimal separating hyperplane. Let be the vertex position matrix of the drone. Let be the vertex position matrix of the obstacle. This is the safety threshold.

[0048] S5: Solve the path planning model to generate control commands for the UAV.

[0049] The method for solving the path planning model in this embodiment is as follows: A sequential quadratic programming solver (such as IPOPT) is used to solve the path planning model, generating control commands for the UAV (including speed). angular velocity This drives the drone to execute obstacle avoidance trajectories.

[0050] First, this invention proposes a weakly conservative geometric modeling method to describe the geometry of a quadcopter drone and obstacles. (Drone) and obstacles The shape of a polyhedron is usually non-convex and irregular. According to geometric theory, polyhedra have good mathematical properties, therefore... and The shape of a drone is accurately described by a non-convex polyhedron, and a non-convex polyhedron can be represented using the union of convex polyhedra. and obstacles It can be modeled as , .

[0051] Secondly, this invention cleverly uses set theory to construct the obstacle avoidance relationship between the drone and obstacles. (Drone) When flying, you need to avoid obstacles in the environment. According to mathematical set theory, this obstacle avoidance safety requirement can be modeled as two sets whose intersection is an empty set, i.e. ,gather and It is a non-convex polyhedron because convex polyhedra can be represented by explicit expressions (such as linear quadratic inequalities), and sets. and Obstacle avoidance requirements where the intersection is an empty set can be studied. and This is achieved through the relationships between subsets of a convex polyhedron, i.e. .

[0052] Finally, this invention further proposes using the hyperplane separation theorem to construct explicit obstacle avoidance constraints. This addresses the obstacle avoidance requirements between the UAV and obstacles. It is implicit; according to convex hull theory, a convex polyhedron... and Each vertex can be used , To explicitly state, further according to the hyperplane separation theorem, the necessary and sufficient condition for two closed convex sets to be unconnected is that there exists a separating hyperplane between them. Therefore, obstacle avoidance requirements... It can be constructed and Separating hyperplanes between This allows us to eventually construct an explicit obstacle avoidance inequality. .

[0053] To verify the effectiveness of the proposed method, a comparison of the complexity between the proposed method and existing methods is shown in Table 1, where the number of vertices of the polyhedral obstacle is M, the number of vertices of the polyhedral UAV is N, and the prediction time domain of the constrained optimization problem is K. Table 1 shows that when the number of vertices of the polyhedral obstacle is M and the number of vertices of the polyhedral UAV is N, the proposed method for constructing an obstacle avoidance inequality requires only four variables: a three-dimensional vector representing the normal direction of the separating hyperplane. It contains three scalar variables. and a scalar variable representing the positional offset of the hyperplane in space. Existing methods require M+N variables. The simplest polyhedron in three-dimensional space is the tetrahedron, which has 4 vertices, therefore M+N≥8. When considering a constrained optimization problem for trajectory planning, the proposed method needs to introduce 4×K, while existing methods require (M+N)×K. Furthermore, as the geometric complexity of obstacles and UAVs increases, i.e., M and N increase, the number of variables required by the proposed method remains constant at 4, while the number required by existing methods, M+N, also increases. Therefore, the proposed method can reduce the number of optimization variables and has lower complexity.

[0054] Table 1. Comparison of the complexity between the method of the present invention and existing methods.

[0055] In summary, the low-complexity obstacle avoidance method for quadrotor UAVs described in this invention addresses the challenge of accurate and rapid obstacle avoidance in confined environments such as mines. Firstly, the geometry of the quadrotor UAV and the obstacle is represented as the union of convex geometric bodies. The obstacle avoidance requirement between the UAV and the obstacle can be transformed into a mathematical problem where the intersection of the two unions is an empty set. To meet the requirement of differentiable inequality constraints in the optimization problem for general nonlinear solvers, this invention further uses the hyperplane separation theorem to transform the implicit obstacle avoidance mathematical expression of the empty intersection into a continuously differentiable explicit obstacle avoidance inequality constraint, achieving efficient obstacle avoidance between the UAV and the obstacle. This invention innovatively uses the hyperplane separation theorem to directly construct the separating hyperplane between the UAV and the obstacle. The constructed obstacle avoidance inequality is differentiable and continuous, directly adaptable to general nonlinear solvers, and effectively improves the efficiency of rapid UAV trajectory planning. It can achieve the precise and real-time obstacle avoidance requirement of the UAV, providing practical technical support for rapid trajectory planning of quadrotor UAVs.

[0056] Based on the aforementioned low-complexity obstacle avoidance method for quadcopter drones, this invention also provides a low-complexity obstacle avoidance system for quadcopter drones, comprising: The geometry building block is used to construct geometric representations of drones and obstacles. The obstacle avoidance condition construction module is used to construct obstacle avoidance conditions for the UAV based on the geometric representations of the UAV and obstacles. An obstacle avoidance inequality establishment module is used to represent the obstacle avoidance conditions of the UAV using a separating hyperplane, and to obtain the UAV obstacle avoidance inequality. The path planning modeling module is used to acquire the position information of the UAV and obstacles in real time, solve for the optimal separating hyperplane based on the UAV obstacle avoidance inequality, and construct the path planning model of the UAV based on the optimal separating hyperplane. The solution module is used to solve the path planning model and generate control commands for the UAV.

[0057] Those skilled in the art will understand that embodiments of this application can be provided as methods, systems, or computer program products. Therefore, this application can take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, this application can take the form of a computer program product embodied on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0058] This application is described with reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of this application. It will be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, special-purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, generate instructions for implementing the flowchart... Figure 1 One or more processes and / or boxes Figure 1 A device that provides the functions specified in one or more boxes.

[0059] These computer program instructions may also be stored in a computer-readable storage medium that can direct a computer or other programmable data processing device to function in a particular manner, such that the instructions stored in the computer-readable storage medium produce an article of manufacture including instruction means, which are implemented in a process Figure 1 One or more processes and / or boxes Figure 1 The function specified in one or more boxes.

[0060] These computer program instructions may also be loaded onto a computer or other programmable data processing equipment to cause a series of operational steps to be performed on the computer or other programmable equipment to produce a computer-implemented process, thereby providing instructions that execute on the computer or other programmable equipment for implementing the process. Figure 1 One or more processes and / or boxes Figure 1 The steps of the function specified in one or more boxes.

[0061] Obviously, the above embodiments are merely illustrative examples for clear explanation and are not intended to limit the implementation. Those skilled in the art will recognize that other variations or modifications can be made based on the above description. It is neither necessary nor possible to exhaustively list all possible implementations here. However, obvious variations or modifications derived therefrom are still within the scope of protection of this invention.

Claims

1. A low-complexity obstacle avoidance method for quadcopters, characterized in that, The method comprises the following steps: constructing a geometric representation of the UAV and the obstacles; constructing an obstacle avoidance condition of the UAV based on the geometric representation of the UAV and the obstacles; representing the obstacle avoidance condition of the UAV by using a separating hyperplane to obtain a UAV obstacle avoidance inequality; obtaining position information of the UAV and the obstacles in real time, and solving an optimal separating hyperplane based on the UAV obstacle avoidance inequality; constructing a path planning model of the UAV based on the optimal separating hyperplane; solving the path planning model to generate a control instruction of the UAV.

2. The low-complexity obstacle avoidance method for quadcopters of claim 1, wherein, The method of constructing a geometric representation of the UAV and the obstacles comprises the following steps: representing the UAV and the obstacles by using polyhedrons and decomposing the polyhedrons into a union set of convex geometries.

3. The low-complexity obstacle avoidance method for quadcopters of claim 2, wherein, The representation of the convex geometries is as follows: representing the convex geometries by using a linear matrix inequality, and the specific formula is as follows: ； wherein, is a convex geometry of the UAV, is an index of the convex geometry of the UAV, is a convex geometry of the obstacle, is an index of the convex geometry of the obstacle, is a point coordinate vector of the three-dimensional space, is a constraint matrix of the convex geometry of the UAV, is an upper bound of the constraint of the convex geometry of the UAV, is a constraint matrix of the convex geometry of the obstacle, is an upper bound of the constraint of the convex geometry of the obstacle.

4. The low-complexity obstacle avoidance method for quadcopters of claim 1 or 2, wherein, The method of constructing an obstacle avoidance condition of the UAV based on the geometric representation of the UAV and the obstacles comprises the following steps: based on the decomposition of the UAV and the obstacles into a union set of convex geometries, the obstacle avoidance condition of the UAV is that any two convex geometries of the UAV and the obstacles are not intersected.

5. The low-complexity obstacle avoidance method for quadcopters of claim 4, wherein, The method of representing the obstacle avoidance condition of the UAV by using a separating hyperplane comprises the following steps: constructing a separating hyperplane to convert the non-intersection between any two convex geometries of the UAV and the obstacles into: all vertex position matrices of the convex geometries of the UAV are located in one half space divided by the separating hyperplane, and all vertex position matrices of the convex geometries of the obstacles are located in the other half space; obtaining a UAV obstacle avoidance inequality based on the separating hyperplane: ； wherein, is a normal vector of the separating hyperplane, is a position offset of the separating hyperplane, is a row vector with element values of 1, is a vertex position matrix of the convex geometry of the UAV, is an index of the convex geometry of the UAV, is a vertex position matrix of the convex geometry of the obstacle, is an index of the convex geometry of the obstacle.

6. The low-complexity obstacle avoidance method for quadcopters of claim 1, wherein, The method of obtaining position information of the UAV and the obstacles in real time comprises the following steps: obtaining point cloud data of an environment in real time by using an on-board sensor, and obtaining vertex position matrices of the obstacles after filtering; calculating vertex position matrices of the UAV according to an attitude estimation value of the UAV.

7. The low-complexity obstacle avoidance method for quadcopters of claim 1, wherein, The method of solving an optimal separating hyperplane based on the UAV obstacle avoidance inequality comprises the following steps: solving the optimal separating hyperplane by using the following formula based on the UAV obstacle avoidance inequality: ； wherein, and are parameters of the separating hyperplane, is a matrix of vertex positions of the UAV, is a matrix of vertex positions of the obstacle, is a safety threshold.

8. The low-complexity obstacle avoidance method for quadcopters of claim 1, wherein, The method of constructing a path planning model of the UAV based on the optimal separating hyperplane comprises the following steps: taking trajectory smoothness and energy minimization as objective functions, constructing a path planning model of the UAV based on the optimal separating hyperplane, and the following formula represents the path planning model. ； ； wherein, is the position of the UAV at time t, is the total time, is the first derivative of is the second derivative of is the weight coefficient, is the second derivative of is the weight coefficient, and are parameters of the optimal separating hyperplane, is the matrix of vertex positions of the convex geometry of the UAV, is the matrix of vertex positions of the obstacle, is the safety threshold.

9. The low-complexity obstacle avoidance method for quadcopters of claim 1, wherein, The method of solving the path planning model comprises the following step: solving the path planning model by using a sequential quadratic programming solver.

10. A low complexity obstacle avoidance system for a quadcopter drone, characterized in that, The method comprises the following steps: a geometric construction module, configured to construct a geometric representation of the UAV and the obstacles; an obstacle avoidance condition construction module, configured to construct an obstacle avoidance condition of the UAV based on the geometric representation of the UAV and the obstacles; an obstacle avoidance inequality establishment module, configured to represent the obstacle avoidance condition of the UAV by using a separating hyperplane to obtain a UAV obstacle avoidance inequality; a path planning modeling module, configured to obtain position information of the UAV and the obstacles in real time, and solve an optimal separating hyperplane based on the UAV obstacle avoidance inequality; construct a path planning model of the UAV based on the optimal separating hyperplane; a solving module, configured to solve the path planning model to generate a control instruction of the UAV.

Images