Multi-algorithm unmanned aerial vehicle intelligent route planning method

Abstract

A multi-algorithm unmanned aerial vehicle intelligent route planning method, comprising: receiving a route planning task, determining the navigation target and constraint conditions of the unmanned aerial vehicle; analyzing the required route planning algorithm characteristics according to the navigation target and constraint conditions of the unmanned aerial vehicle; dynamically selecting one or more route planning algorithms suitable for the current task requirements from the preset algorithm library according to the analysis result; dynamically loading the selected route planning algorithm; generating a preliminary path of the route; defining the objective function and loss function for different navigation targets; optimizing the objective function according to the defined loss function, adjusting the weight value of the influence factors of the navigation target; determining the optimal route. The method solves the problem that a single algorithm is difficult to adapt to complex scenarios, and improves the accuracy and efficiency of route planning. Through the dynamic loading and switching mechanism, the optimal algorithm strategy can be selected according to the actual demand, the route can be dynamically adjusted, and the flight safety and task completion quality are improved.

Description

Technical Field

[0001] This invention relates to unmanned aerial vehicle (UAV) route planning technology, and in particular to a multi-algorithm intelligent route planning method for UAVs. Background Technology

[0002] With the development of drone technology, drones are increasingly widely used in various industries. Flight path planning is a crucial issue in drone applications, affecting flight efficiency, safety, and mission completion quality. Existing technologies offer various flight path planning algorithms for different application scenarios and needs, such as the Dixtra algorithm, A* algorithm, and genetic algorithms. However, different algorithms exhibit varying advantages and disadvantages in different scenarios, making it difficult for a single algorithm to adapt to diverse application requirements.

[0003] It should be noted that the information disclosed in the background section above is only for understanding the background of this application, and therefore may include information that does not constitute prior art known to those skilled in the art. Summary of the Invention

[0004] The main objective of this invention is to solve the problems existing in the background art mentioned above and to provide a multi-algorithm unmanned aerial vehicle (UAV) intelligent route planning method.

[0005] To achieve the above objectives, the present invention adopts the following technical solution:

[0006] In a first aspect of the present invention, a multi-algorithm unmanned aerial vehicle (UAV) intelligent route planning method includes the following steps:

[0007] S1: Receive the flight path planning task and determine the UAV's flight target and constraints;

[0008] S2: Analyze the characteristics of the required route planning algorithm based on the UAV's navigation objectives and constraints;

[0009] S3: Based on the analysis results of step S2, dynamically select one or more route planning algorithms suitable for the current task requirements from the preset algorithm library;

[0010] S4: Dynamically load the route planning algorithm selected in step S3;

[0011] S5: Generate a preliminary path for the air route using the algorithm loaded in step S4;

[0012] S6: Define the objective function and loss function for different navigation objectives;

[0013] S7: Apply machine learning methods to optimize the objective function based on the loss function defined in step S6, and adjust the weight values ​​of the factors affecting the navigation target;

[0014] S8: Based on the optimization results in step S7, determine the final optimal route.

[0015] In a second aspect of the invention, a computer program product includes a computer program that, when executed by a processor, implements the multi-algorithm unmanned aerial vehicle (UAV) intelligent route planning method.

[0016] The present invention has the following beneficial effects:

[0017] This invention provides a multi-algorithm intelligent flight path planning method for unmanned aerial vehicles (UAVs). Through a dynamic loading and switching mechanism, it can select different algorithm strategies according to actual needs, employing a multi-level algorithm strategy: first, an initial path is generated using a pre-selected algorithm such as the Dixtra algorithm; then, machine learning methods are used to optimize the path, ultimately generating the optimal flight path. This multi-algorithm intelligent flight path planning method solves the problem that a single algorithm is difficult to adapt to complex scenario requirements, improving the accuracy and efficiency of intelligent flight path planning for UAVs.

[0018] The method of this invention can use a unified API interface to integrate flight path planning algorithms, dynamically load and switch different algorithms according to actual needs, and support the application of multi-level algorithms, including a flight path planning layer and a local optimization layer of route weight under different objectives, thereby improving the accuracy and efficiency of UAV flight path planning.

[0019] Based on the method of this invention, a UAV route engine capable of integrating multiple algorithms and dynamically selecting the optimal algorithm according to specific circumstances can be provided. This is a multi-algorithm intelligent UAV route engine that integrates multiple intelligent route planning algorithms through a unified API interface and supports pluggable algorithm output. Through this engine, different algorithms can be dynamically loaded and switched according to actual needs, and an expansion interface can be reserved to support adding factors affecting the route and replacing different levels of algorithm applications.

[0020] Other beneficial effects of the embodiments of the present invention will be further described below. Attached Figure Description

[0021] Figure 1 This is an interactive diagram of the multi-algorithm UAV intelligent flight path engine according to an embodiment of the present invention.

[0022] Figure 2 This is a flowchart of the preliminary path planning (first-level path planning) according to an embodiment of the present invention.

[0023] Figure 3 This is a flowchart of path weight optimization (second-layer path optimization) according to an embodiment of the present invention.

[0024] Figure 4 This is a schematic diagram illustrating weight optimization using gradient descent in an embodiment of the present invention.

[0025] Figure 5 This is a data-driven optimization framework according to an embodiment of the present invention.

[0026] Figure 6 The following is pseudocode describing the implementation process of the Dixtra algorithm in an embodiment of the present invention.

[0027] Figure 7 This is a schematic diagram illustrating the evaluation of the machine learning objective function in an embodiment of the present invention. Detailed Implementation

[0028] The embodiments of the present invention will be described in detail below. It should be emphasized that the following description is merely exemplary and not intended to limit the scope and application of the present invention.

[0029] Furthermore, the terms "first" and "second" are used for descriptive purposes only and should not be construed as indicating or implying relative importance or implicitly specifying the number of technical features indicated. Thus, a feature defined as "first" or "second" may explicitly or implicitly include one or more of that feature. In the description of embodiments of the present invention, "a plurality of" means two or more, unless otherwise explicitly specified.

[0030] See Figures 1 to 4 This invention provides a multi-algorithm unmanned aerial vehicle (UAV) intelligent route planning method, comprising the following steps:

[0031] Step S1: Receive the flight path planning task and determine the UAV's flight target and constraints;

[0032] Step S2: Analyze the task requirements and, based on the UAV's navigation objectives and constraints, analyze the characteristics of the required flight path planning algorithm.

[0033] Step S3: Dynamically select an algorithm. Based on the analysis results of step S2, dynamically select one or more route planning algorithms suitable for the current task requirements from the preset algorithm library.

[0034] Step S4: Dynamically load the algorithm. The route planning algorithm selected in step S3 can be dynamically loaded through the algorithm manager.

[0035] Step S5: Preliminary path generation, using the algorithm loaded in step S4 to generate a preliminary route;

[0036] Step S6: Path optimization preparation, defining the objective function and loss function for different navigation objectives;

[0037] Step S7: Path optimization is performed by applying machine learning methods to optimize the objective function based on the loss function defined in step S6 and adjusting the weight values ​​of the factors influencing the navigation target.

[0038] Step S8: Generate the optimal route. Based on the optimization results in step S7, determine the final optimal route.

[0039] In some embodiments, the navigation objective includes optimal time, shortest route, or lowest risk; wherein, the characteristic parameters related to optimal time include horizontal distance, true altitude of the route, degree of congestion of the route, speed of the UAV, etc., and the weight of each route segment is related to the characteristic parameters related to optimal time; the characteristic parameters related to shortest route include horizontal distance, true altitude of the route, etc., and the weight of each route segment is related to the characteristic parameters related to shortest route; the characteristic parameters related to lowest risk include weather conditions, terrain complexity of the route, etc., and the weight of each route segment is related to the characteristic parameters related to lowest risk.

[0040] In a preferred embodiment, the generated flight path can be dynamically adjusted by combining real-time mission data and environmental perception data to improve flight safety and mission completion quality.

[0041] Figure 1 This invention illustrates a UAV flight path engine based on the method of this invention, capable of integrating multiple algorithms and dynamically selecting the optimal algorithm according to specific circumstances. It is a multi-algorithm intelligent flight path engine for UAVs that integrates multiple intelligent flight path planning algorithms through a unified API interface and supports pluggable algorithm output. This engine allows for the dynamic loading and switching of different algorithms according to actual needs, and reserves expansion interfaces to support adding factors influencing flight paths and replacing different levels of algorithm applications. As a further optimization, an extensible design is adopted, reserving expansion interfaces to support the addition of new influencing factors and the integration of new flight path planning algorithms.

[0042] See Figure 1 In a preferred embodiment, a target influencing factor library is pre-built to store the correspondence between different navigation targets and related influencing factors. In step S2, when performing the operation to analyze task requirements, the navigation target and related influencing factors are input through a unified API interface, and the influencing factors and weight values ​​are returned. A target influencing factor weight library is pre-built to store and optimize path weights. In step S7, when performing path optimization, the path weight optimization layer is called through a unified API interface to optimize the path weights. The path weight optimization layer is configured to adjust and optimize the weights of various influencing factors in the flight path.

[0043] See Figure 2In a preferred embodiment, in step S5, the preliminary path generation uses the Dixtra algorithm as the default base algorithm. Based on the user-defined target and the system-configured weight factors, the weight value of each route segment is calculated, and a preliminary path is generated. Specifically, the preliminary path generation includes: determining the navigation target and selecting one of the targets with the optimal time, shortest distance, or lowest risk; obtaining the influencing factors affecting the navigation target, including various parameters that may affect route planning; assigning initial weights to each influencing factor, where each weight represents the importance of each influencing factor in route planning; calculating the weight value of each route segment using a weight function based on the navigation target, influencing factors, and their weights; wherein the weights of the influencing factors evolve to an optimal value through an iterative process; and applying the Dixtra algorithm to generate a preliminary route from the starting point to the destination using the calculated weight values ​​of each route segment.

[0044] More preferably, during the execution of the Dixtra algorithm, the node closest to the starting node is selected from the priority queue for expansion at each step, and the shortest distances of other nodes are dynamically updated. The final result includes the shortest distance of each node and the predecessor node of each node, which can be used to backtrack to find the shortest path (see [link]). Figure 6 (The pseudocode shown).

[0045] See Figure 3 In a preferred embodiment, step S7 specifically includes: defining an objective function based on the navigation objective and relevant influencing factors; constructing a loss function to evaluate the merits of the objective function and determine the deviation between the objective function and the actual navigation objective; executing gradient descent to iteratively optimize the objective function and gradually adjust the weight values ​​of each influencing factor; evaluating the optimization results using the loss function to evaluate the optimization results after each iteration to determine whether it is close to the optimal solution; determining the optimal weight values ​​by using gradient descent to determine the optimal weight values ​​of each influencing factor to form the optimal solution function; and generating an optimized path by applying the optimal weight values ​​to adjust the initial path and generate an optimized path that conforms to the navigation objective.

[0046] The gradient descent method specifically includes the following process: calculating the gradient of the weights and biases relative to the loss function for each training data point; calculating the sum of the gradients of the weights of all training data points; calculating the sum of the gradients of the biases of all training data points; calculating the average gradient of the weights and biases of all samples; updating the weights and biases based on the average gradient; repeating the gradient calculation and weight update process until the loss function converges. Figure 4 This is a schematic diagram illustrating weight optimization using gradient descent in an embodiment of the present invention.

[0047] See Figure 5In some embodiments, the method of the present invention is implemented based on a data-driven optimization framework, including the following data-driven implementation process: Dataset segmentation: The collected data is divided into a training dataset and a test dataset for model training and validation; Preliminary analysis: The training data is preliminarily analyzed to identify key features affecting route planning; Feature engineering: Based on the results of the preliminary analysis, features are selected and constructed; Model selection: A suitable machine learning model is selected based on the navigation objective and the results of feature engineering; Parameter setting: Necessary parameters are set for the selected model to adapt to the specific route planning objective; Model training: The model is trained using the training dataset; Model testing: The trained model is tested using the test dataset; Model prediction: The trained and tested model is applied to predict and optimize the initially generated path to generate an optimized path that better matches the navigation objective; Model evaluation and iteration: The model performance is evaluated based on the test results, and if necessary, the model is reselected or the parameters are adjusted, and the training and testing process is repeated.

[0048] The multi-algorithm intelligent flight path planning method for unmanned aerial vehicles (UAVs) of this invention can adapt to changing application requirements and complex scenarios. This method integrates and dynamically loads multiple flight path planning algorithms, allowing the UAV to flexibly switch and select the most suitable algorithm based on actual flight conditions and mission requirements. The method employs a multi-level algorithm strategy: first, an initial path is generated using an initial selection algorithm such as the Dixtra algorithm; then, machine learning methods are used to optimize these paths to generate the optimal flight path. This method not only improves the accuracy and efficiency of flight path planning but also enhances the flight efficiency and safety of the UAV, while ensuring the quality of mission completion. The flight path planning engine based on this invention features an extensible interface, supporting the integration of future new technologies and the application of new algorithms. Overall, this invention, through intelligent flight path planning, significantly improves the flexibility and optimization of UAV applications across various industries, demonstrating broad application prospects and significant market value.

[0049] The following describes specific embodiments of the present invention.

[0050] This invention provides a multi-algorithm intelligent flight path planning method for unmanned aerial vehicles (UAVs), and designs a multi-algorithm intelligent flight path engine for UAVs, including a dynamic loading and switching mechanism, an extensibility design, and a multi-level algorithm application strategy. It can integrate multiple flight path planning algorithms through a unified API interface, achieving pluggable algorithm output. The dynamic loading and switching mechanism allows for dynamic loading and switching of corresponding flight path planning algorithms through configuration items. The extensibility design reserves an extension interface to support adding factors affecting the flight path and different levels of algorithms. The multi-level algorithm application strategy includes two levels: preliminary path generation and path optimization. The preliminary path generation layer preferably uses the Dixtra algorithm, calculating the weight value of each flight path segment based on the user-defined objective and the system-configured weight factors, and generating a preliminary path. The path optimization layer uses machine learning methods to optimize the objective function under different flight objectives and the optimal solution function among various factors, deriving the final optimal flight path.

[0051] 1. Dynamic Loading and Switching Mechanism: As this is a UAV route planning engine that integrates multiple algorithms, the corresponding algorithms can be dynamically loaded and switched through configuration options to observe the advantages and disadvantages of different algorithms in route calculation under different conditions. This invention defaults to configuring the Dixtra algorithm in the first-layer route planning layer (other path planning algorithms can also be switched), and the second-layer local optimization layer uses gradient descent to optimize the route weights W (other machine learning optimization algorithms can also be switched).

[0052] 2. Scalability: It has reserved expansion interfaces to support the addition of factors affecting the flight path and the addition of different levels of algorithms.

[0053] 3. Applicable to multi-level algorithms:

[0054] The first layer is used for initial path generation, employing the basic Dixtra algorithm. This layer first determines the navigation objective (optimal time / shortest distance / minimum risk), and generates a weight value W for each segment of the route based on the user-defined objective and the weighting factors configured by the system. The weighting function is as follows:

[0055] W = w1x1 + w2x 2+ w3x 3+......+ w n x n

[0056] Where x1, x2, ..., xn are the factors influencing the target, and w1, w2, ..., wn are the weights of these factors. The weights of the factors are initially given default values, and then the optimal values ​​are evolved iteratively. Then, the optimal route is calculated using the Dixtra algorithm to recommend a route.

[0057] The second layer is used for path optimization, primarily optimizing the objective function and the optimal solution function among various factors under different navigation objectives (optimal time / shortest route / minimum risk). Following general machine learning practices, the objective function is first initially derived based on factor x and objective y, and then the quality of the objective function is evaluated using a loss function. Gradient descent is then used to continuously approach the neighborhood of the optimal solution, accelerating convergence to the optimal solution, thus obtaining the optimal solution function and the weight values ​​of each factor.

[0058] Dynamic loading and switching mechanism

[0059] This engine features a dynamic loading and switching mechanism, allowing for the flexible loading of different route planning algorithms through configuration options. These options include algorithm name and parameters, which users can configure according to their needs. During runtime, the engine dynamically loads the corresponding algorithm based on the configuration options and performs route planning calculations. In different application scenarios, by switching between different algorithms and comparing their calculation results, the optimal algorithm is selected for practical application.

[0060] Extensibility Design

[0061] To accommodate potential new requirements and algorithms, the engine design includes reserved extension interfaces. These interfaces include interfaces for factors influencing flight paths and algorithm extension interfaces. The factors influencing flight paths interface allows users to add new influencing factors, such as wind speed and obstacle distribution. The algorithm extension interface supports users adding new route planning algorithms, which can be integrated into the engine through simple interface calls, achieving fusion with existing algorithms.

[0062] Multi-level algorithm application

[0063] The multi-algorithm UAV intelligent route engine employs a multi-layered algorithm application strategy during route planning, specifically including:

[0064] 1. Preliminary path generation

[0065] Using the Dixtra algorithm, the weight value W of each route segment is calculated based on the user-defined objectives (such as optimal time, shortest route, and minimum risk) and the weight factors configured by the system.

[0066] The initial weight values ​​are set based on empirical values ​​and optimized in subsequent iterations.

[0067] The initial path is calculated using the Dixtra algorithm, providing a foundation for subsequent optimization.

[0068] Figure 6The pseudocode shown illustrates the implementation of Dixtra's algorithm. During execution, the algorithm iteratively expands upon the node closest to the starting node from a priority queue, dynamically updating the shortest distances of other nodes. The final result includes the shortest distance of each node and its predecessor node, which can be used to backtrack and find the shortest path.

[0069] Using Dixtra's algorithm for initial path generation has the following advantages: it considers edge weights and selects the edge with the shortest distance from the current node to the starting node at each step. The introduction of weights provides a unified mathematical standard for weight optimization for different objectives in later stages. Furthermore, Dixtra's algorithm exhibits good performance when there are no negative-weighted edges, with a time complexity typically of O((V+E)logV), where V is the number of nodes and E is the number of edges.

[0070] 2. Path optimization

[0071] Based on the initially generated path, further optimization is performed for different navigation objectives (optimal time, shortest distance, and lowest risk). First, the specific objectives of the route are determined. Based on these objectives, preliminary feature engineering is conducted to select different models and set feature parameters that vary depending on the objective for training (see...). Figure 6 ).

[0072] Using machine learning methods, a preliminary objective function is derived, and its performance is evaluated by constructing a loss function and using test data from a test dataset. Figure 7 As shown.

[0073] The following distance loss functions, L1 and L2 loss functions can be used to optimize the objective function during machine learning model training:

[0074]

[0075] When finding the minimum value of the loss function, gradient descent (GD) can be used to iteratively solve the problem step by step, obtaining the minimized loss function and model parameter values. The weight values ​​of each influencing factor are continuously adjusted to gradually approach the optimal solution and finally determine the optimal path.

[0076] Gradient descent minimizes the objective function by iteratively updating parameters along the negative direction of the gradient. It is an optimization algorithm used to find the minimum value of a function. Its core idea is to use the gradient information of the objective function to guide the parameter updates, thereby gradually approaching the minimum value of the function.

[0077] Suppose we have a differentiable objective function f(θ), where θ is a parameter vector. The goal of gradient descent is to find a set of parameters θ that minimizes f(θ). Gradient The direction in which the function grows fastest at θ is indicated, so the function value can be gradually decreased by updating θ in the opposite direction of the gradient.

[0078] The update formula for gradient descent is:

[0079]

[0080] Here, α is the learning rate, used to control the step size of the update. Gradient It can be obtained by taking the partial derivative:

[0081]

[0082] The gradient descent method is calculated as follows:

[0083] for i = 0 to the number of training data:

[0084] 1. Calculate the gradient of the weights and biases b of the i-th training data with respect to the loss function. This will ultimately yield the gradient values ​​of the weights and biases for each training data point.

[0085] 2. Calculate the sum of the gradients of all training data weights.

[0086] 3. Calculate the sum of the gradients of all training data biases.

[0087] After completing the above calculations, proceed with the following calculations:

[0088] 1. Using the results obtained in steps 2 and 3 above, calculate the average gradient of the weights and biases for all samples.

[0089] 2. Use the following formula to update the weight and bias values ​​for each sample.

[0090]

[0091] 3. Repeat the above process until the loss function converges and remains constant.

[0092] Example

[0093] Example 1: Time-optimal route planning

[0094] In a drone delivery system, the goal is to minimize delivery time, with the system configuration aiming for optimal time management. First, based on the user-defined objective and the system's configured weights, the weight value W for each flight segment is calculated. Then, the Dixtro algorithm is used to calculate a preliminary path. Finally, to further optimize the path for optimal time management, machine learning methods are employed to arrive at the final optimal flight route.

[0095] The characteristic parameters that are relevant to time comparison include the horizontal distance, the true altitude of the flight path, the congestion level of the flight path, and the speed of the UAV. Therefore, the weight W of each flight path segment is related to these characteristic parameters.

[0096] Example 2: Route planning based on the shortest distance

[0097] In another application scenario, drones are used for long-distance inspections, requiring the shortest possible flight distance. The system configuration objective is to minimize the distance traveled. First, the weight value W of each flight segment is calculated, and a preliminary path is calculated using the Dixtro algorithm. Subsequently, machine learning methods are applied to optimize the path to achieve the shortest distance objective, ultimately determining the optimal flight route.

[0098] The characteristic parameters related to the distance traveled include the horizontal distance and the true altitude of the route. Therefore, the weight W of each route segment is related to these characteristic parameters.

[0099] Example 3: Route planning based on minimum risk

[0100] For drone missions that require traversing complex environments, minimizing risk is the primary objective. The system sets initial weights based on environmental factors and calculates a preliminary path using the Dixtro algorithm. Building upon this, machine learning methods are employed to optimize the route for minimizing risk, resulting in the safest possible flight path.

[0101] The characteristic parameters related to safety comparison include weather conditions and route topographic complexity. Therefore, the weight W of each route segment is related to these characteristic parameters.

[0102] This invention integrates multiple flight path planning algorithms through a unified API interface, enabling dynamic loading and switching, scalability design, and multi-level algorithm application strategies for a multi-algorithm intelligent flight path engine for unmanned aerial vehicles (UAVs). This engine can select the optimal algorithm and generate the optimal flight path based on actual needs, improving the flexibility and optimization level of UAV flight path planning, and has broad application prospects and market value.

[0103] The performance of the multi-algorithm UAV intelligent route engine can be further optimized by combining more machine learning algorithms and data analysis techniques to continuously improve the intelligence and adaptability of route planning. By introducing real-time data and environmental perception technologies, dynamic adjustments to UAV route planning can be achieved, improving flight safety and mission completion quality.

[0104] This invention also provides a storage medium for storing a computer program, which, when executed, performs at least the methods described above.

[0105] This invention also provides a control device, including a processor and a storage medium for storing a computer program; wherein the processor executes the computer program by performing at least the method described above.

[0106] This invention also provides a processor that executes a computer program, at least performing the methods described above.

[0107] The storage medium can be implemented by any type of non-volatile storage device, or a combination thereof. The non-volatile memory can be read-only memory (ROM), programmable read-only memory (PROM), erasable programmable read-only memory (EPROM), electrically erasable programmable read-only memory (EEPROM), magnetic random access memory (FRAM), flash memory, magnetic surface memory, optical disc, or compact disc read-only memory (CD-ROM); the magnetic surface memory can be a disk drive or magnetic tape drive. The storage media described in the embodiments of this invention are intended to include, but are not limited to, these and any other suitable types of memory.

[0108] In the several embodiments provided by this invention, it should be understood that the disclosed systems and methods can be implemented in other ways. The device embodiments described above are merely illustrative. For example, the division of units is only a logical functional division, and in actual implementation, there may be other division methods, such as: multiple units or components can be combined, or integrated into another system, or some features can be ignored or not executed. In addition, the coupling, direct coupling, or communication connection between the various components shown or discussed can be through some interfaces, and the indirect coupling or communication connection between devices or units can be electrical, mechanical, or other forms.

[0109] The units described above as separate components may or may not be physically separate. The components shown as units may or may not be physical units, that is, they may be located in one place or distributed across multiple network units. Some or all of the units may be selected to achieve the purpose of this embodiment according to actual needs.

[0110] In addition, in the various embodiments of the present invention, each functional unit can be integrated into one processing unit, or each unit can be a separate unit, or two or more units can be integrated into one unit; the integrated unit can be implemented in hardware or in the form of hardware plus software functional units.

[0111] Those skilled in the art will understand that all or part of the steps of the above method embodiments can be implemented by hardware related to program instructions. The aforementioned program can be stored in a computer-readable storage medium. When the program is executed, it performs the steps of the above method embodiments. The aforementioned storage medium includes various media capable of storing program code, such as mobile storage devices, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.

[0112] Alternatively, if the integrated units of this invention are implemented as software functional modules and sold or used as independent products, they can also be stored in a computer-readable storage medium. Based on this understanding, the technical solutions of the embodiments of this invention, or the parts that contribute to the prior art, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the methods described in the various embodiments of this invention. The aforementioned storage medium includes various media capable of storing program code, such as mobile storage devices, ROM, RAM, magnetic disks, or optical disks.

[0113] The methods disclosed in the several method embodiments provided by this invention can be arbitrarily combined without conflict to obtain new method embodiments.

[0114] The features disclosed in the several product embodiments provided by this invention can be arbitrarily combined without conflict to obtain new product embodiments.

[0115] The features disclosed in the several method or device embodiments provided by the present invention can be arbitrarily combined without conflict to obtain new method or device embodiments.

[0116] The above description, in conjunction with specific preferred embodiments, provides a further detailed explanation of the present invention. It should not be construed that the specific implementation of the present invention is limited to these descriptions. For those skilled in the art, various equivalent substitutions or obvious modifications can be made without departing from the concept of the present invention, and all such modifications, achieving the same performance or application, should be considered within the scope of protection of the present invention.

Claims

1. A multi-algorithm intelligent route planning method for unmanned aerial vehicles, characterized in that, The method comprises the following steps: S1: receiving a route planning task, determining the flight target and the constraint condition of the UAV; S2: analyzing the required route planning algorithm characteristics according to the flight target and the constraint condition of the UAV; S3: dynamically selecting one or more route planning algorithms suitable for the current task requirement from a preset algorithm library according to the analysis result of step S2; S4: dynamically loading the route planning algorithm selected in step S3; S5: generating a preliminary path of the route by using the algorithm loaded in step S4; S6: defining the objective function and the loss function for different flight targets; S7: applying a machine learning method to optimize the objective function according to the loss function defined in step S6, and adjusting the weight value of the influencing factor of the flight target; S8: determining the final optimal route based on the optimization result in step S7; Wherein, the corresponding relationship between different flight targets and related influencing factors is established in advance, and in step S2, the flight target and the related influencing factors are input during the operation of analyzing the task requirement, and the influencing factors and the weight value are returned; The path weight is established in advance, and in step S7, the path weight optimization layer is called to optimize the path weight during the operation of path optimization execution, and the path weight optimization layer is set to adjust and optimize the weight of each influencing factor in the route path.

2. The multi-algorithm UAV intelligent path planning method of claim 1, wherein, Further comprising: By pre-constructing the target influencing factor library, the corresponding relationship between different flight targets and related influencing factors is stored; In step S2, the flight target and the related influencing factors are input through a unified API interface, and the influencing factors and the weight value are returned; By pre-constructing the target influencing factor weight library, the path weight is stored and optimized; In step S7, the path weight optimization layer is called through a unified API interface. 3.The multi-algorithm UAV intelligent path planning method of claim 1 or 2, wherein, In step S5, the generation of the preliminary path adopts the Dijkstra algorithm as the default basic algorithm, calculates the weight value of each route according to the target set by the user and the weight factor configured by the system, and generates the preliminary path.

4. The multi-algorithm UAV intelligent path planning method of claim 3, wherein, In step S5, the preliminary path generation specifically comprises: Determine the flight target, select one of the time-optimal, shortest distance or minimum risk targets; Obtain the influencing factors affecting the flight target, including various parameters that may affect the route planning; Assign an initial weight to each influencing factor, which represents the importance of each influencing factor in route planning; According to the flight target, the influencing factor and the weight, the weight function is used to calculate the weight value of each route; wherein the weight of the influencing factor is evolved to the optimal value through an iteration process; Apply the Dijkstra algorithm to generate the preliminary path of the route from the starting point to the ending point by using the calculated weight value of each route.

5. The multi-algorithm UAV intelligent path planning method of claim 4, wherein, During the execution of the Dijkstra algorithm, the node closest to the starting node is selected from the priority queue for expansion each time, and the shortest distance of other nodes is dynamically updated, and the final returned result includes the shortest distance of each node and the predecessor node of each node, which can be used to backtrack the shortest path.

6. The multi-algorithm UAV intelligent path planning method of any one of claims 1 to 2, wherein, In step S7, the path optimization execution specifically comprises: Define the objective function, and preliminarily obtain the objective function based on the flight target and the related influencing factors; A loss function is constructed to evaluate the performance of the objective function and determine the deviation between the objective function and the actual navigation target. A gradient descent method is executed to iteratively optimize the objective function and gradually adjust the weight values of each influencing factor. The optimization results after each iteration are evaluated using the loss function to determine whether the optimal solution is approaching. The optimal weight values of each influencing factor are determined through the gradient descent method to form the optimal solution function. An optimized path is generated by adjusting the preliminary path using the optimal weight values to meet the navigation target.

7. The multi-algorithm UAV intelligent path planning method of claim 6, wherein, In step S7, the gradient descent method specifically includes the following processes: Calculate the gradient of the weight and bias with respect to the loss function for each training data. Calculate the total gradient of the weight of all training data. Calculate the total gradient of the bias of all training data. Calculate the average gradient of the weight and bias of all samples. Update the weight and bias according to the average gradient. Repeat the gradient calculation and weight update process until the loss function converges.

8. The multi-algorithm UAV intelligent path planning method of any one of claims 1 to 2, wherein, The navigation target includes time optimization, shortest distance, or minimum risk. The characteristic parameters related to time optimization include the distance of horizontal route, true altitude of route, congestion degree of route, and speed of UAV, and the weight of each route is related to the characteristic parameters related to time optimization. The characteristic parameters related to shortest distance include the distance of horizontal route and true altitude of route, and the weight of each route is related to the characteristic parameters related to shortest distance. The characteristic parameters related to minimum risk include weather conditions and terrain complexity of route, and the weight of each route is related to the characteristic parameters related to minimum risk.

9. The multi-algorithm UAV intelligent path planning method of any one of claims 1 to 2, wherein, The method is implemented based on a data-driven optimization framework, including the following data-driven implementation processes: Split the data set, divide the collected data into training data set and test data set for model training and verification. Preliminary analysis, preliminary analysis of training data to identify key features affecting route planning. Feature engineering, based on the results of preliminary analysis, feature selection and construction; Select a model, select a suitable machine learning model based on the navigation target and the results of feature engineering; Set parameters, set the necessary parameters for the selected model to adapt to the specific route planning target; Model training, train the model using the training data set; Model testing, test the trained model using the test data set; Model prediction, apply the trained and tested model to predict and optimize the initially generated path to generate an optimized path that better meets the navigation target; Model evaluation and iteration, evaluate the performance of the model based on the test results, if necessary, reselect the model or adjust the parameters, and repeat the training and testing process. The computer program, when executed by a processor, implements the multi-algorithm intelligent route planning method for unmanned aerial vehicles according to any one of claims 1 to 9.

10. A computer program product comprising a computer program, characterized in that, ​

Images