Efficient autonomous three-dimensional path planning method and system for heavy-load low-altitude aircraft

Abstract

The invention discloses an efficient autonomous three-dimensional path planning method and system for a heavy-load low-altitude aircraft, path planning is realized through an improved A * algorithm, firstly, a planning area is rasterized, an obstacle avoidance area is defined, and node search and cost value calculation are only performed in the obstacle avoidance area, so that the calculation amount is reduced. The node cost value is jointly determined by heuristic estimation of a known operation distance and an unknown operation distance and motion energy consumption cost, and the path searching efficiency and quality of the aircraft are optimized. In the path post-processing stage, redundant steering nodes are removed through a key path node screening mechanism, and a more efficient path is obtained. In addition, an efficient learning quantification method of non-structural pruning optimization is introduced, and accurate prediction of the energy consumption of the aircraft is realized through an improved radial basis function neural network. The method has the advantages that the method is not only suitable for autonomous navigation of low-altitude aircrafts, but also can be widely applied to path planning and energy consumption optimization in various fields of unmanned aerial vehicles, automatic piloting aircrafts and the like.

Description

Technical Field

[0001] The present invention relates to the technical field of unmanned aerial vehicles, and particularly to an efficient autonomous three-dimensional path planning method and system for heavy-load low-altitude aircraft. Background Art

[0002] With the rapid development of unmanned aerial vehicle technology, low-altitude aircraft are increasingly widely used in various fields, especially in tasks such as heavy-load transportation, inspection, environmental monitoring, and mapping. However, when performing tasks, low-altitude aircraft face a complex flight environment, including the presence of obstacles, restrictions on the flight space, and issues such as minimizing flight time and energy consumption. These factors make the path planning of low-altitude aircraft more complex and challenging.

[0003] Currently, the path planning methods for low-altitude aircraft mainly include graph-based algorithms, sampling-based methods, and optimization-based algorithms, etc. Among them, the A* algorithm, as a classic graph search algorithm, has been widely used in path planning. The A* algorithm can effectively find the shortest path from the starting point to the ending point by combining a heuristic estimation function and the actual path cost. However, when dealing with complex environments, the traditional A* algorithm has the following problems:

[0004] Large computational amount: During the path search process of the traditional A* algorithm, it is necessary to comprehensively expand and search nodes in the entire space. Even in areas with fewer obstacles, a large amount of ineffective calculations will be performed, resulting in a large computational amount and low efficiency.

[0005] Insufficient path quality: When calculating the path, the traditional A* algorithm only considers the total cost of the path, ignoring the actual motion energy consumption and dynamic change factors of the aircraft, resulting in the obtained path may not be the optimal energy consumption path and cannot fully meet the energy-saving requirements of the aircraft.

[0006] Path redundancy: During the aircraft path planning process, especially in complex environments, the traditional A* algorithm may generate a large number of redundant turning nodes, resulting in too many broken lines and too long distance of the path, affecting the flight efficiency of the aircraft.

[0007] Poor dynamic obstacle avoidance ability: When dealing with obstacles in a dynamic environment, the traditional A* algorithm often cannot quickly adapt to environmental changes, resulting in poor flexibility and real-time performance of path planning.

[0008] Regarding the above problems, although many studies have made attempts to optimize path planning algorithms, there are still the following deficiencies: existing improved algorithms usually focus on one aspect (such as optimizing path quality or computational efficiency), and it is difficult to balance multiple objectives. Especially in the three-dimensional path planning of heavy-load low-altitude aircraft, it is necessary to meet the obstacle avoidance requirements while considering various factors such as energy consumption, path smoothness, and real-time performance. Therefore, how to optimize the path quality, especially energy consumption and redundant turning problems, while ensuring computational efficiency, remains a technical problem to be solved urgently. Summary of the Invention

[0009] In view of the defects of the prior art, the present invention provides an efficient autonomous three-dimensional path planning method and system for heavy-load low-altitude aircraft. By combining the aircraft motion model and energy consumption prediction, through methods such as segmented search, definition of obstacle avoidance areas, and screening of key path nodes, the efficiency, quality of path planning, and energy efficiency of the aircraft are improved to meet the actual needs of heavy-load low-altitude aircraft.

[0010] In order to achieve the above invention objectives, the technical solutions adopted by the present invention are as follows:

[0011] An efficient autonomous three-dimensional path planning method for heavy-load low-altitude aircraft, which performs path planning through an improved A* algorithm. The method includes the following steps:

[0012] Step 1: Definition of the planning area. First, rasterize the target area and input the starting point and the ending point. Mark the intermediate area of the obstacles crossed by the ideal optimal path (the line connecting the starting point and the ending point), and define this area as the obstacle avoidance area. Only search for nodes and calculate the cost values within the obstacle avoidance area, and there is no need to calculate the node costs in the non-obstacle avoidance area.

[0013] Step 2: Calculation of node cost values. Design a segmented search function. According to the obstacle avoidance area defined in Step 1, search for path nodes segment by segment and calculate the cost value of each node, and finally select the path with the minimum total cost. The formula for calculating the node cost value is:

[0014] F(n) = G(n) + H(n) + E(n)

[0015] where G(n) is the known travel distance, H(n) is the heuristic estimate of the unknown travel distance, and E(n) is the motion energy consumption cost.

[0016] Step 3: Path post-processing. Screen the redundant turning nodes of the planned path, and remove unnecessary turns through the screening mechanism of key path nodes to finally obtain a shorter and more efficient path.

[0017] Furthermore, the definition of the obstacle avoidance area is based on the intermediate area of the obstacles crossed by the ideal optimal path, and the accuracy of the obstacle avoidance area is controlled by setting a weight factor.

[0018] Further, the segmented search function is used to perform segmented calculations on the obstacle avoidance area, and the total cost of the calculation includes the actual cost from the starting point to the node, the heuristic cost from the node to the end point, and the motion energy consumption cost. The selection of each node depends on the minimum total cost criterion.

[0019] Further, the formula of the segmented search function is as follows:

[0020]

[0021] where nodes is the node with the minimum cost value; S is the starting point; P is the end point; node 1 ,…,node n are the adjacent nodes of multiple obstacles traversed by the ideal optimal path, used to define the obstacle avoidance area (the intermediate area of the adjacent nodes of the obstacle), that is, the obstacle avoidance area (node 1 ,node 2 ),…,(node n-1 ,node n );G(n 1 ),…,G(n n-1 ) are the known operating distance costs of the aircraft; H(n 1 ),…,H(n n-1 ) are the unknown operating distance costs of the aircraft; E(n 1 ),…,E(n n-1 ) are the motion energy consumption costs of the aircraft.

[0022] Further, in step 2, the calculation of the selection of the adjacent nodes of the obstacle is based on the following formula:

[0023] Cost=a(D S→node +D node→P )+bE(node)+cM

[0024] where D S→node and D node→P respectively represent the Euclidean distances from the starting point to the node and from the node to the end point, E(node) is the energy consumption value of the node, M is the number of obstacles traversed by the connection line of the starting point, the node and the end point, and a, b, c are weight factors.

[0025] Further, the heuristic estimation function H(n) in step 2 uses the Euclidean distance calculation, representing the shortest path distance from the current node to the end point.

[0026] Further, the energy consumption calculation E(n) in step 2 is based on the dynamic model and motion attitude analysis of the aircraft, uses a non-linear mapping relationship to quantify the motion energy consumption of the aircraft, and improves the energy consumption estimation accuracy through physical experiment data. The specific calculation is as follows:

[0027] E(n n ) = F[E(n n ), N(n n )] + ΔE(n n )

[0028] Among them, F[E(n n ), N(n n )] represents the non-linear mapping relationship between the energy consumption and the rotational speed of the flight drive motor, and ΔE(n n ) is the error between the estimated energy consumption and the actual energy consumption.

[0029] Further, in the critical path node screening mechanism in step 3, by judging the spatial relationship of consecutive nodes in the path, if there is no obstacle between adjacent nodes, they are merged into a straight line, so as to screen out the critical nodes, reduce redundant turning nodes, and improve the path quality.

[0030] Further, in the process of calculating the node cost value in step 2, the radial basis function neural network (RBFNN) is used to predict the motion energy consumption of the aircraft at different nodes, and the prediction accuracy and calculation efficiency are improved through the unstructured pruning optimization method.

[0031] Further, the hidden layer activation function of the RBFNN after unstructured pruning optimization is shown as follows:

[0032]

[0033] In the formula, N(n n-1 ) represents the rotational speed of the flight drive motor of node n n-1 , [c j - 3b j c j + 3b j is the central range of the radial basis function of the hidden layer neuron, c j is the center position of the jth radial basis function, b j is the width of the jth radial basis function, and ‖N(n n-1 ) - c j ‖ represents the Euclidean distance.

[0034] The present invention also discloses an efficient autonomous three-dimensional path planning system for a heavy-duty low-altitude aircraft, which can be used to implement the above-mentioned efficient autonomous three-dimensional path planning method for a heavy-duty low-altitude aircraft. Specifically, it includes:

[0035] The planning area definition module is used to rasterize the target area, input the starting point and the ending point, calculate the ideal optimal path (the straight line connecting the starting point and the ending point), mark the intermediate area of the obstacles crossed by the path, define this area as the obstacle avoidance area, and only search for nodes and calculate the cost value within the obstacle avoidance area. There is no need to calculate the node cost in the non-obstacle avoidance area;

[0036] The node cost value calculation module is used to, according to the raster map of the planning area and the obstacle avoidance area, by designing a segmented search function, search for path nodes segment by segment and calculate the cost value of each node, and finally select the path with the minimum total cost;

[0037] The path post-processing module is used to screen redundant turning nodes in the planned path, remove unnecessary turns through the screening mechanism of key path nodes, and finally obtain a shorter and more efficient path;

[0038] The result display module is used to display the optimal path planned, show the key nodes on the path and the relevant information of the path, provide the user with an intuitive path planning result, and can perform further path adjustment as needed.

[0039] The present invention also discloses a computer device, including a memory, a processor, and a computer program stored on the memory and executable on the processor. When the processor executes the program, it implements the above-mentioned efficient autonomous three-dimensional path planning method for a heavy-load low-altitude aircraft.

[0040] The present invention also discloses a computer-readable storage medium, on which a computer program is stored. When the program is executed by a processor, it implements the above-mentioned efficient autonomous three-dimensional path planning method for a heavy-load low-altitude aircraft.

[0041] Compared with the prior art, the advantages of the present invention are as follows:

[0042] 1. By improving the A* algorithm, using the ideal optimal path as a reference, and only searching for nodes and calculating the cost value within the obstacle avoidance area, the calculation amount of irrelevant areas is significantly reduced, thereby improving the efficiency of path search.

[0043] 2. The segmented search function and the node cost value calculation mechanism are introduced. Considering the known running distance, the heuristic estimation of the unknown running distance, and the motion energy consumption comprehensively, it can effectively plan the path with the minimum total cost, ensuring that the aircraft can efficiently and accurately avoid obstacles in a complex environment.

[0044] 3. Through the screening mechanism of key path nodes in the path post-processing module, unnecessary turns are removed, redundant path nodes are reduced, ensuring that the flight path of the aircraft is smoother and more intuitive, thereby improving the flight efficiency and reducing the energy consumption.

[0045] 4. The present invention provides a more practical energy consumption prediction for the flight planning of low-altitude aircraft by introducing the calculation of motion energy consumption cost, accurately quantifying the law of energy consumption change based on the motion model and energy consumption data of the aircraft, and optimizing the energy efficiency of the aircraft.

[0046] 5. The present invention can independently perform path planning, obstacle avoidance, and subsequent path optimization, reducing manual intervention and improving the adaptability and autonomous decision-making ability of the aircraft in complex and dynamic environments. BRIEF DESCRIPTION OF THE DRAWINGS

[0047] Figure 1 is a flowchart of the efficient autonomous three-dimensional path planning method for the heavy-duty low-altitude aircraft according to the embodiment of the present invention;

[0048] Figure 2 is a schematic diagram of path post-processing according to the embodiment of the present invention, where (a) is a polyline obstacle avoidance path and (b) is the screening of key path nodes;

[0049] Figure 3 is a schematic diagram of adjacent nodes of obstacles according to the embodiment of the present invention;

[0050] Figure 4 is the original RBFNN structure diagram;

[0051] Figure 5 is the RBFNN structure diagram optimized by non-structural pruning according to the embodiment of the present invention. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0052] To make the objectives, technical solutions, and advantages of the present invention clearer and more understandable, the following further describes the present invention in detail with reference to the drawings and by way of examples.

[0053] The present invention aims at the three-dimensional path planning problem of heavy-duty low-altitude aircraft and proposes an efficient autonomous planning method. This method improves the A* algorithm. Taking the ideal optimal path (the line connecting the starting point and the ending point) as a reference, it searches for path nodes in segments. A segmented search function is designed, and the search and cost value calculation of nodes only occur in the intermediate area of multiple obstacles crossed by the ideal optimal path, further reducing the node calculation amount and improving the efficiency and quality of path search. In addition, this method considers two indicators of the operating distance and motion energy consumption of the aircraft, and proposes an efficient learning quantization method optimized by non-structural pruning to quantitatively analyze the error between the approximate energy consumption obtained by traditional mathematical derivation and the actual energy consumption, and obtain a more accurate law of energy consumption change of the aircraft. The specific process is as follows Figure 1 .

[0054] Specifically, it is divided into three steps:

[0055] Step1: Definition of the planning area. The map is rasterized, and the starting point and the ending point are input. The ideal optimal path is marked, and the intermediate area passing through multiple obstacles on this path is defined as the obstacle avoidance area. The search for nodes and the calculation of cost values only occur in the newly defined obstacle avoidance area. In the non-obstacle avoidance area, there is no need to search for nodes and calculate the node cost.

[0056] Step2: Calculation of node cost values. Set a segmented search function. Corresponding to the multiple obstacle avoidance areas in Step1, segmentally search for path nodes and calculate the corresponding cost values, and finally plan a broken-line obstacle avoidance path with the minimum total cost. The segmented function is as follows:

[0057]

[0058] In the formula: nodes is the node with the minimum cost value; S is the starting point; P is the ending point; node 1 ,…,node n are the adjacent nodes of multiple obstacles passed by the ideal optimal path, which are used for the definition of the obstacle avoidance area in Step1 (the intermediate area between adjacent obstacle nodes), that is, the obstacle avoidance area (node 1 ,node 2 ),…,(node n-1 ,node n );G(n 1 ),…,G(n n-1 ) are the known operating distance costs of the aircraft; H(n 1 ),…,H(n n-1 ) are the unknown operating distance costs of the aircraft; E(n 1 ),…,E(n n-1 ) are the motion energy consumption costs of the aircraft.

[0059] Step3: Path post-processing. The path planning is completed in the raster map, and there are many unnecessary turning nodes in the planned path. In order to obtain a high-quality path with fewer redundant turning path nodes and a shorter distance, this method also introduces a screening mechanism for key path nodes, as Figure 2 shown.

[0060] Figure 2 In a, the finally planned broken-line obstacle avoidance path includes the following path nodes: S→node 1 →node 2 →node 3 →node 4 →node 5 →node 6 →node 7 →P. Among them, node 3 and node 4It is a redundant turning path node, generating unnecessary polyline turning distances.

[0061] Figure 2 In b, the screened key path nodes are S → node 1 → node 2 → node 5 → node 6 → node 7 → P. The specific screening mechanism is as follows: In Figure a, there are a total of nine path nodes from S → P. Let the corresponding coordinates be (node xi , node yi , node zi ), i = 1, …, 9. If there are no obstacles in the cube area composed of (node xi : node x(i+2) , node yi : node y(i+2) , node zi : node z(i+2) ), then (node xi , node yi , node zi ) is directly connected to (node x(i+2) , node y(i+2) , node z(i+2) ). Screen them in turn, retain the key nodes, and update the final path.

[0062] Regarding the above Step2, the following supplementary explanations are provided:

[0063] 1. For the obstacle adjacent nodes node 1 , …, node n in the piecewise function expression of Step2, the selection process is as follows:

[0064] As Figure 3 shown, along the forward direction of the ideal optimal path, each obstacle has three adjacent nodes (top surface / two side surfaces), such as node 1 , node′ 1 , and node″ 1 . The specific selection is completed with reference to the following formula.

[0065] Cost = a(D S→node + D node→P ) + bE(node) + cM

[0066]

[0067] In the formula: (node x , node y, node z ), are the coordinates of node node; (S x , S y , S z ), are the starting point coordinates; (P x , P y , P z ), are the ending point coordinates; M is the number of obstacles crossed by the three lines connecting S, node, and P; a, b, and c are weight factors; The node with the minimum Cost is selected as the final adjacent node in the piecewise function expression.

[0068] 2. In the piecewise function expression of Step2, H(n 1 ), …, H(n n-1 ) can be calculated using the Euclidean distance. Taking H(n n-1 ), n n-1 ∈ (node n-1 , node n ) as an example, the calculation is as follows:

[0069]

[0070] Where: are the coordinates of node n n-1 .

[0071] 3. For E(n 1 ), …, E(n n-1 ) in the piecewise function expression of Step2, the solution process is as follows:

[0072] Based on the ducted fan dynamics model, combined with the analysis of the aerial motion attitude, the non - linear mapping relationship between the aircraft motion energy consumption and the flight drive motor speed is obtained; With the help of physical prototype tests, the flight drive motor speed data and the aircraft motion energy consumption data are collected, and an efficient learning quantization method with unstructured pruning optimization is proposed to quantitatively analyze the error between the estimated energy consumption obtained from the above - mentioned mapping relationship and the actual test energy consumption, further improving the mapping accuracy and obtaining a more accurate aircraft energy consumption change law. Taking E(n n-1 ), n n-1 ∈ (node n-1 , node n ) as an example, the calculation is as follows:

[0073] The specific expression is as follows:

[0074] E(n n-1 ) = F[E(n n-1 ), N(n n-1 )] + ΔE(n n-1 )

[0075] Where: F[E(n n-1 ), N(n n-1 )] represents the non-linear mapping relationship between the aircraft motion energy consumption E and the rotational speed N of the flight drive motor; ΔE(n n-1 ) is the quantization error between the estimated energy consumption and the actual test energy consumption.

[0076] 4. For the above-mentioned efficient learning quantization method with unstructured pruning optimization, the specific explanation is as follows:

[0077] An efficient learning quantization method with unstructured pruning optimization is proposed by unstructured pruning of redundant connected neurons in a Radial Basis Function Neural Network (RBFNN). The RBFNN includes an input layer, a hidden layer, and an output layer. In the original RBFNN, each neuron in the input layer is connected to all neurons in the hidden layer, as Figure 4 shown. However, for the hidden layer, the activation function (radial basis function) of its neurons responds locally to the input. When an input approaches the central range of the radial basis function of a neuron, the hidden layer neuron will produce a large output. When the input is far from the central range, the output will decay exponentially. The activation function of the hidden layer in the original RBFNN is shown as follows:

[0078]

[0079] Where: w j is the connection weight between the hidden layer neuron and the network output; h j (N(n n-1 )) is the hidden layer activation function; c j is the central position of the activation function; b j is the width of the activation function.

[0080] In the original RBFNN, the connection between neurons in the input layer and the hidden layer is too redundant. As described above, when an input is far from the central range of the radial basis function of a hidden layer neuron, the output of the hidden layer is no longer obvious, and the connection effect between the input layer neuron corresponding to the above input and the above hidden layer neuron is not significant, and unstructured pruning can be performed. After pruning, the hidden layer neuron is only connected to the input layer neurons within the central range of its radial basis function, reducing the network calculation amount and improving the quantization efficiency. The RBFNN after unstructured pruning optimization is as Figure 5 shown. The optimized activation function of the hidden layer is shown as follows:

[0081]

[0082] Where, [c j - 3b j cj +3b j is the central range of the radial basis function of the hidden layer neurons.

[0083] In another embodiment of the present invention, an efficient autonomous three-dimensional path planning system for a heavy-duty low-altitude aircraft is provided. This system can be used to implement the efficient autonomous three-dimensional path planning method for the heavy-duty low-altitude aircraft described above. Specifically, it includes:

[0084] A planning area definition module, which is used to rasterize the target area, input the starting point and the ending point, calculate the ideal optimal path (the line connecting the starting point and the ending point), and mark the intermediate area of the obstacles crossed by the path. This area is defined as the obstacle avoidance area, and node search and cost value calculation are only performed within the obstacle avoidance area. There is no need to calculate the node cost in the non-obstacle avoidance area;

[0085] A node cost value calculation module, which is used to search for path nodes segment by segment and calculate the cost value of each node according to the raster map of the planning area and the obstacle avoidance area by designing a segmented search function, and finally select the path with the minimum total cost;

[0086] A path post-processing module, which is used to screen redundant turning nodes in the planned path, remove unnecessary turns through the screening mechanism of key path nodes, and finally obtain a shorter and more efficient path;

[0087] A result display module, which is used to display the optimal path planned, show the key nodes on the path and the relevant information of the path, provide the user with an intuitive path planning result, and can perform further path adjustment according to needs.

[0088] In another embodiment of the present invention, a terminal device is provided. The terminal device includes a processor and a memory. The memory is used to store a computer program, and the computer program includes program instructions. The processor is used to execute the program instructions stored in the computer storage medium. The processor may be a Central Processing Unit (CPU), or may also be other general-purpose processors, Digital Signal Processors (DSPs), Application Specific Integrated Circuits (ASICs), Field-Programmable Gate Arrays (FPGAs), or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. It is the computing core and control core of the terminal, and is suitable for implementing one or more instructions. Specifically, it is suitable for loading and executing one or more instructions to implement the corresponding method flow or corresponding function. The processor described in the embodiment of the present invention can be used to perform the operations of the efficient autonomous three-dimensional path planning method for a heavy-duty low-altitude aircraft.

[0089] In another embodiment of the present invention, a storage medium is also provided, specifically a computer-readable storage medium (Memory). The computer-readable storage medium is the memory device in the terminal device and is used to store programs and data. It can be understood that the computer-readable storage medium here can include both the built-in storage medium in the terminal device and, of course, the extended storage medium supported by the terminal device. The computer-readable storage medium provides a storage space, and this storage space stores the operating system of the terminal. And, in this storage space, one or more instructions suitable for being loaded and executed by the processor are also stored. These instructions can be one or more computer programs (including program codes). It should be noted that the computer-readable storage medium here can be a high-speed RAM memory or a non-volatile memory, such as at least one disk memory.

[0090] One or more instructions stored in the computer-readable storage medium can be loaded and executed by the processor to implement the corresponding steps of the efficient autonomous three-dimensional path planning method for a heavy-duty low-altitude aircraft in the above embodiment; one or more instructions in the computer-readable storage medium are loaded and executed by the processor.

[0091] Those skilled in the art should understand that the embodiments of the present invention can be provided as a method, a system, or a computer program product. Therefore, the present invention can take the form of a complete hardware embodiment, a complete software embodiment, or an embodiment combining software and hardware aspects. Moreover, the present invention can take the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to disk memory, CD-ROM, optical memory, etc.) that contain computer-usable program code.

[0092] The present invention is described with reference to the flowcharts and / or block diagrams of methods, apparatuses (systems), and computer program products according to the embodiments of the present invention. It should be understood that each flow and / or block in the flowchart and / or block diagram can be implemented by computer program instructions, and the combination of the flows and / or blocks in the flowchart and / or block diagram can also be implemented by computer program instructions. These computer program instructions can be provided to the processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing devices to generate a machine, so that the instructions executed by the processor of the computer or other programmable data processing devices generate means for implementing the functions specified in one Figure 1 one flow or multiple flows and / or blocks Figure 1 one block or multiple blocks.

[0093] These computer program instructions can also be stored in a computer-readable memory that can direct a computer or other programmable data processing device to work in a specific manner, so that the instructions stored in the computer-readable memory generate a manufactured article including instruction means, and the instruction means implement the functions specified in one Figure 1 one flow or multiple flows and / or blocks Figure 1 one block or multiple blocks.

[0094] These computer program instructions can also be loaded onto a computer or other programmable data processing device, so that a series of operation steps are executed on the computer or other programmable device to generate a computer-implemented process. Therefore, the instructions executed on the computer or other programmable device provide steps for implementing the functions specified in one Figure 1 one flow or multiple flows and / or blocks Figure 1 one block or multiple blocks.

[0095] Those of ordinary skill in the art will realize that the embodiments described herein are for helping readers understand the implementation methods of the present invention, and it should be understood that the protection scope of the present invention is not limited to such specific statements and embodiments. Those of ordinary skill in the art can make various other specific deformations and combinations that do not deviate from the essence of the present invention based on the technical revelations disclosed in the present invention, and these deformations and combinations are still within the protection scope of the present invention.

Claims

1. An efficient autonomous three-dimensional path planning method for a heavy-load low-altitude aircraft, characterized in that: The following steps are involved: Step 1: Define the planning area. First, rasterize the target area and enter the starting point and end point. Mark the middle area of ​​obstacles that the path passes through according to the ideal optimal path, and define this area as the obstacle avoidance area; search for nodes and calculate cost values ​​only in the obstacle avoidance area, and no node cost calculation is required in the non-obstacle avoidance area; Step 2: Calculate the node cost value. Design a segmented search function. According to the obstacle avoidance area defined in step 1, search the path nodes segment by segment and calculate the cost value of each node. Finally, select the path with the minimum total cost. The node cost value calculation formula is: F(n)=G(n)+H(n)+E(n) Among them, G(n) is the known running distance, H(n) is the heuristic estimate of the unknown running distance, and E(n) is the energy cost of movement; Step 3: Path post-processing: screening redundant turning nodes on the planned path, removing unnecessary turns through the screening mechanism of key path nodes, and finally obtaining a shorter and more efficient path.

2. The efficient autonomous three-dimensional path planning method for a heavy-loaded low-altitude aircraft according to claim 1 is characterized by: The definition of the obstacle avoidance area is based on the middle area of ​​the obstacle traversed by the ideal optimal path, and the accuracy of the obstacle avoidance area is controlled by setting the weight factor.

3. The efficient autonomous three-dimensional path planning method for a heavy-loaded low-altitude aircraft according to claim 1 is characterized by: The segmented search function is used to perform segmented calculations on the obstacle avoidance area. The total cost calculated includes the actual cost from the starting point to the node, the heuristic cost from the node to the end point, and the motion energy consumption cost. The selection of each node depends on the minimum total cost criterion.

4. The efficient autonomous three-dimensional path planning method for a heavy-loaded low-altitude aircraft according to claim 1 is characterized in that: The segmented search function formula is as follows: Among them, nodes is the node with the minimum cost value; S is the starting point; P is the end point; node1,…,node n It is the adjacent nodes of multiple obstacles traversed by the ideal optimal path, which is used to define the obstacle avoidance area, that is, the obstacle avoidance area (node1, node2),…, (node n-1 ,node n ); G(n1),…,G(n n-1 ) is the known operating distance cost of the aircraft; H(n1),…,H(n n-1 ) is the unknown operating distance cost of the aircraft; E(n1),…,E(n n-1 ) is the energy consumption cost of the aircraft.

5. The efficient autonomous three-dimensional path planning method for a heavy-load low-altitude aircraft according to claim 1 is characterized by: In step 2, the calculation of obstacle neighbor node selection is based on the following formula: Cost=a(D S→node +D node→P )+bE(node)+cM Among them, D S→node and D node→P They represent the Euclidean distance from the starting point to the node and from the node to the end point respectively, E(node) is the energy consumption value of the node, M is the number of obstacles traversed by the line connecting the starting point, the node and the end point, and a, b, and c are weight factors.

6. The efficient autonomous three-dimensional path planning method for a heavy-load low-altitude aircraft according to claim 1 is characterized by: The heuristic estimation function H(n) in step 2 is calculated using the Euclidean distance, which represents the shortest path distance from the current node to the end point.

7. The efficient autonomous three-dimensional path planning method for a heavy-load low-altitude aircraft according to claim 1 is characterized by: The energy consumption calculation E(n) in step 2 is based on the dynamic model and motion posture analysis of the aircraft, uses nonlinear mapping relationships to quantify the motion energy consumption of the aircraft, and improves the accuracy of energy consumption estimation through physical experimental data; the specific formula is: E(n n )=F[E(n n ),N(n n )]+ΔE(n n ) Among them, F[E(n n ),N(n n )] represents the nonlinear mapping relationship between energy consumption and flight drive motor speed, ΔE(n n ) is the error between the estimated energy consumption and the actual energy consumption.

8. The efficient autonomous three-dimensional path planning method for a heavy-load low-altitude aircraft according to claim 1 is characterized by: The key path node screening mechanism in step 3 determines the spatial relationship of consecutive nodes in the path. If there are no obstacles between the nodes, they are merged into a straight line to screen out the key nodes.

9. The efficient autonomous three-dimensional path planning method for a heavy-load low-altitude aircraft according to claim 1, characterized in that: In the process of calculating the node cost value in step 2, the radial basis function neural network RBFNN is used to predict the motion energy consumption of the aircraft at different nodes, and the prediction accuracy and calculation efficiency are improved through the unstructured pruning optimization method.

10. The method for efficient autonomous three-dimensional path planning of a heavy-load low-altitude aircraft according to claim 9, characterized in that: The hidden layer activation function of RBFNN after optimization by unstructured pruning is as follows: In the formula, N(n n-1 ) represents node n n-1 The flight drive motor speed, [c j -3b j c j +3b j ] is the central range of the radial basis function of the hidden layer neurons, c j is the center position of the jth radial basis function, b j is the width of the jth radial basis function, ‖N(n n-1 )-c j ‖ represents the Euclidean distance.

Images