Unmanned aerial vehicle autonomous obstacle avoidance path planning method and system based on AI

Abstract

The invention discloses an AI-based unmanned aerial vehicle autonomous obstacle avoidance path planning method and system, and the method comprises the steps: collecting path data and related data of a target region, and carrying out the preprocessing of the path data and related data; performing exception screening on the path data according to a passing threshold to obtain exceptional point information, and performing obstacle passing ability analysis on the exceptional point information to obtain a first parameter; performing variable influence analysis on the related data to obtain a second parameter, and constructing a joint function according to the first parameter and the second parameter; and constructing a target path optimization model according to the joint function, and outputting an optimization result from the to-be-optimized data to the target path optimization model. The method improves the precision of joint optimization of the search mode and the flight path, has good interpretability, and can be directly applied to a joint optimization system of the loading mode and the flight path.

Description

Technical Field

[0001] The present invention relates to the field of path optimization, and in particular to an AI-based autonomous obstacle avoidance path planning method and system for unmanned aerial vehicles. Background Art

[0002] In flight search management, efficient search methods and reasonable flight path planning are crucial for reducing costs, minimizing losses, and improving flight efficiency. Traditionally, search methods and flight paths are optimized independently, ignoring their mutual influence, making it difficult to achieve optimal overall flight efficiency. Furthermore, the flight process is affected by a variety of factors, such as weather conditions, road conditions, and airspace traffic. Changes in these factors can lead to anomalies during flight, such as loss of flight signals and damage, further increasing flight costs and management difficulties.

[0003] In recent years, the development of big data technology and intelligent optimization algorithms has provided new insights into the joint optimization of search methods and flight paths. Genetic algorithms, as a search algorithm that simulates the natural evolutionary process, have been widely used in complex system optimization due to their strong global search capabilities and robustness. However, there is currently no mature AI-based autonomous obstacle avoidance path planning method for UAVs that can comprehensively consider multiple influencing factors and achieve coordinated optimization of loading methods and flight paths. Therefore, it is necessary to improve the accuracy, precision, and real-time performance of the joint optimization of AI-based autonomous obstacle avoidance path planning methods for UAVs. Summary of the Invention

[0004] The purpose of this invention is to provide an AI-based autonomous obstacle avoidance path planning method and system for unmanned aerial vehicles.

[0005] To achieve the above object, the present invention is implemented according to the following technical solutions:

[0006] The present invention comprises the following steps:

[0007] Collecting path data and related data of the target area, and preprocessing the path data and the related data;

[0008] Performing anomaly screening on the path data according to a pass threshold to obtain abnormal point information, and performing obstacle passability analysis on the abnormal point information to obtain a first parameter variable;

[0009] Performing a variable influence analysis on the relevant data to obtain a second parameter variable, and constructing a joint function based on the first parameter variable and the second parameter variable;

[0010] A target path optimization model is constructed according to the joint function, the target path optimization model is constructed with the data to be optimized, and the optimization result is output.

[0011] Furthermore, the method of performing abnormal screening on the path data according to the pass threshold to obtain abnormal point information includes:

[0012] The time points of the path data where the stopover and flight speed are less than 30 km / h are taken as problem data points. The information weight of the problem data points is calculated based on the data characteristics and importance:

[0013]

[0014] The i-th path data of the u-th problem data point is u i , the yth path data of the uth problem data point is u y , path data u i The probability of occurrence is Path data u y The probability of occurrence is Path data u i and path data u y The probability of simultaneous occurrence is Observed path data u y Time path data u i The posterior probability is The number of path data is m, and the path data u i The information weight is

[0015] The problem data points with information weight greater than 0.412 are taken as key data points, and the key data points are optimized:

[0016] Calculate particle fitness:

[0017]

[0018] The probability parameter is sp, and the position of particle x is D x , position D x The information weight is The fitness function of particle x is

[0019] Use roulette wheel to select excellent particles as the parent generation of crossover mutation and calculate the selection probability:

[0020]

[0021] The zth particle is x z , particle x z The fitness of The number of particles is Particle x z The probability of selection is Particle x z The position of D(xz );

[0022] Perform crossover mutation on particles and calculate the crossover rate and mutation rate:

[0023]

[0024] The crossover rate The mutation rate is The parameters between 0 and 1 are β1, β2, β3, and β4, and the average fitness of the particle population is The maximum fitness of the particle population is The fitness of the crossover individual is The fitness of the mutant individual is

[0025] Reverse sort the gene segments of the particles after crossover mutation to obtain new particles, calculate the fitness of the new particles, and calculate the pass threshold based on the fitness:

[0026]

[0027] The pass threshold is η, and the new particle x new The fitness of The flying speed of the u-th problem data point is b u , the minimum flight speed of the u-th problem data point is b min ;

[0028] Iterate continuously until the maximum number of iterations is reached, and output the key data points with a pass threshold greater than 0.374 as outlier information.

[0029] Furthermore, the method of performing obstacle capability analysis on the abnormal point information to obtain a first parameter variable includes:

[0030] Input the outlier information into the obstacle passing capacity analysis model and calculate the airspace flow based on the outlier information:

[0031] p=a·f

[0032] a=(k) -1

[0033] The airspace flow is p, the blocking density is a, and the spacing between drones is k;

[0034] Calculate the relationship between drone flow and speed:

[0035]

[0036] The relationship between drone flow and speed is ρ, the function curvature control parameter is μ, and the maximum density that causes airspace congestion is a. max, the maximum flight speed of the UAV is f v ;

[0037] Calculate the distance between drones:

[0038]

[0039] The length of the drone is l, and the minimum safe distance of the drone is s o , the flight speed of the drone is f, and the free flow speed is f o The safe head-to-head distance between the human-driven drone and the vehicle is H. o ;

[0040] According to the spacing between UAVs and airspace flow, the airspace flow basic graph model analysis function is given:

[0041]

[0042] The abnormal point information is analyzed through the analytical function to obtain the airspace flow and airspace congestion density, and the airspace flow and airspace congestion density are weighted and integrated to obtain the first parameter variable.

[0043] Furthermore, the method of performing a variable influence analysis on the relevant data to obtain a second parameter variable includes:

[0044] Construct a state matrix based on the relevant data, use the time of the relevant data as the node, and calculate the contribution of the node:

[0045]

[0046] The contribution of the a-th related data of the t-th node is ψ a (t), the flight score of the a-th related data of the t-th node is The flight score of the a-th related data of the t-1th node is The amount of relevant data for the tth node is The control coefficient of the a-th related data is γ a , the ath related data of the tth node is h a (t), the a-th related data of the t-1-th node is h a (t-1), the influence degree is σ, and the variable of the a-th related data of the t-th node is Δη a,t , the wth flight score index value of the ath related data of the tth node is V a,w (t), the number of flight score indicators is N, and the weight of the w-th flight score indicator is λ w ;

[0047] Calculate the average probability of a node:

[0048]

[0049] The average probability of the a-th related data of the t-th node is The number of nodes with the a-th related data is M, and the probability of the a-th related data appearing at the t-th node is ζ a (t);

[0050] Construct a weighted directed graph based on the directed weighted network and calculate the importance index of the node:

[0051]

[0052] The importance index of the tth node is ε u , the natural constant is e, and the weight of the outgoing connection of the tth node is C ot (u), the weight of the connection received by the t-th node is C in (u), the influence coefficient is υ;

[0053] Calculate the weighted value of the node:

[0054]

[0055] The weighted value of the tth node is The probability that the a-th related data of the t-th node is in an influential state is ζ a,yx (t), the weighting factor is λ;

[0056] Calculate the second argument:

[0057]

[0058] The flight score of the tth node is The standard value of the a-th related data is h ao The standard value of flight score is The second parameter is

[0059] Furthermore, the method for constructing a joint function according to the first parameter and the second parameter includes:

[0060] Obtain key data based on importance screening of search methods and flight paths, and obtain flight evaluation indicators based on key data;

[0061] According to the constructed joint function, the expression is:

[0062]

[0063] The second parameter is The first parameter is The optimization coefficient is The actual arrival time is T st , agreed arrival time The number of the nth type of flying drones is B n , the energy consumption per kilometer of the nth type of flying UAV is g n , the flight speed of the nth flying drone is S n , the sign function is sgn(·), and the fixed cost of flight is R gd , the joint function is

[0064] Furthermore, the method for constructing a target path optimization model according to the joint function includes:

[0065] The objective function is constructed based on the joint function, and the expression is:

[0066]

[0067] The control factor is The loss function is The objective function is

[0068] Target path optimization models include particle swarm optimization algorithm, random forest algorithm, and recurrent neural network;

[0069] The particle swarm optimization algorithm adjusts the search direction and position through information sharing and accumulation of individual experience among individual particles in the swarm, adapts to the joint function requirements of the flight, realizes efficient search and optimization in the multi-dimensional feature space, and obtains an improved flight plan;

[0070] The random forest algorithm constructs multiple decision trees and obtains prediction results based on the improved flight plan, capturing the nonlinear relationship of the prediction results and extracting effective features based on the nonlinear relationship;

[0071] The recurrent neural network updates the internal state at the sequence time step through a recursive loop structure, captures the dynamic features of the data to be optimized, and predicts the dynamic features based on the effective features to obtain the joint function value.

[0072] In a second aspect, an embodiment of the present application further provides an electronic device, including:

[0073] A processor; and a memory arranged to store computer executable instructions, which when executed cause the processor to perform the method steps described in the first aspect.

[0074] In a third aspect, an embodiment of the present application further provides a computer-readable storage medium, which stores one or more programs. When the one or more programs are executed by an electronic device including multiple applications, the electronic device executes the method steps described in the first aspect.

[0075] The beneficial effects of the present invention are:

[0076] The present invention is an AI-based autonomous obstacle avoidance path planning method and system for unmanned aerial vehicles. Compared with the existing technology, the present invention has the following technical effects:

[0077] The present invention can improve the accuracy of user intention recognition in multi-dimensional human-computer interaction scenarios through preprocessing, anomaly screening, obstacle passability analysis, variable impact analysis, construction of joint functions, model building and model optimization steps, thereby improving the accuracy of joint optimization of search methods and flight paths. Jointly optimizing the search method and flight path can greatly save resources and improve work efficiency. It can realize intelligent joint optimization of the search method and flight path, and perform obstacle passability analysis and multi-data fusion on the joint optimization of the search method and flight path in real time, which is of great significance to obstacle passability analysis, can adapt to obstacle passability analysis of different standards and different obstacle passability analysis requirements, and has a certain universality. BRIEF DESCRIPTION OF THE DRAWINGS

[0078] Figure 1 This is a flowchart of the steps of the AI-based autonomous obstacle avoidance path planning method and system for drones of the present invention;

[0079] Figure 2 This is a schematic diagram of the structure of an electronic device in an embodiment of this specification. DETAILED DESCRIPTION

[0080] The present invention will be further described below through specific examples. The illustrative examples and descriptions of the present invention are used to explain the present invention but are not intended to limit the present invention.

[0081] The AI-based autonomous obstacle avoidance path planning method for a UAV of the present invention comprises the following steps:

[0082] like Figure 1 As shown, in this embodiment, the following steps are included:

[0083] Collecting path data and related data of the target area, and preprocessing the path data and the related data;

[0084] In the actual evaluation, a batch of electronic products in a logistics park needed to be transported autonomously from a warehouse to a distribution center 20 km away, avoiding obstacles.

[0085] Path data includes flight speed, altitude, route trajectory, temperature, humidity, wind speed, and traffic control information; related data includes the weight, volume, and fragility of electronic products;

[0086] Performing anomaly screening on the path data according to a pass threshold to obtain abnormal point information, and performing obstacle passability analysis on the abnormal point information to obtain a first parameter variable;

[0087] In the actual evaluation, there are 15 outlier points, and the first parameter is 0.6918;

[0088] Performing a variable influence analysis on the relevant data to obtain a second parameter variable, and constructing a joint function based on the first parameter variable and the second parameter variable;

[0089] In the actual evaluation, the second parameter is 0.4471;

[0090] Construct a target path optimization model according to the joint function, convert the target path optimization model into the data to be optimized, and output the optimization result;

[0091] In the actual evaluation, the optimization results were to change the flight altitude and route to avoid traffic control areas and high-risk airspace. The flight distance was changed to 22km, but the flight time was shortened by 15 minutes and energy consumption was reduced by 10%.

[0092] In this embodiment, the method for performing abnormality screening on the path data according to the pass threshold to obtain abnormal point information includes:

[0093] The time points of the path data where the stopover and flight speed are less than 30 km / h are taken as problem data points. The information weight of the problem data points is calculated based on the data characteristics and importance:

[0094]

[0095] The i-th path data of the u-th problem data point is u i , the yth path data of the uth problem data point is u y , path data u i The probability of occurrence is Path data u y The probability of occurrence is Path data u i and path data u y The probability of simultaneous occurrence is Observed path data u y Time path data u i The posterior probability is The number of path data is Path data u i The information weight is

[0096] The problem data points with information weight greater than 0.412 are taken as key data points, and the key data points are optimized:

[0097] Calculate particle fitness:

[0098]

[0099] The probability parameter is sp, and the position of particle x is D x , position D x The information weight is The fitness function of particle x is

[0100] Use roulette wheel to select excellent particles as the parent generation of crossover mutation and calculate the selection probability:

[0101]

[0102] The zth particle is x z , particle x z The fitness of The number of particles is Particle x z The probability of selection is Particle x z The position of D(x z );

[0103] Perform crossover mutation on particles and calculate the crossover rate and mutation rate:

[0104]

[0105] The crossover rate The mutation rate is The parameters between 0 and 1 are β1, β2, β3, and β4, and the average fitness of the particle population is The maximum fitness of the particle population is The fitness of the crossover individual is The fitness of the mutant individual is

[0106] Reverse sort the gene segments of the particles after crossover mutation to obtain new particles, calculate the fitness of the new particles, and calculate the pass threshold based on the fitness:

[0107]

[0108] The pass threshold is η, and the new particle x new The fitness of The flying speed of the u-th problem data point is bu , the minimum flight speed of the u-th problem data point is b min ;

[0109] Iterate continuously until the maximum number of iterations is reached, and output the key data points with a pass threshold greater than 0.374 as outlier information.

[0110] In this embodiment, the method for performing barrier pass capability analysis on the outlier information to obtain a first parameter variable includes:

[0111] Input the outlier information into the obstacle passing capacity analysis model and calculate the airspace flow based on the outlier information:

[0112] p=a·f

[0113] a=(k) -1

[0114] The airspace flow is p, the blocking density is a, and the spacing between drones is k;

[0115] Calculate the relationship between drone flow and speed:

[0116]

[0117] The relationship between drone flow and speed is ρ, the function curvature control parameter is μ, and the maximum density that causes airspace congestion is a. max , the maximum flight speed of the UAV is f v ;

[0118] Calculate the distance between drones:

[0119]

[0120] The length of the drone is l, and the minimum safe distance of the drone is s o , the flight speed of the drone is f, and the free flow speed is f o The safe head-to-head distance between the human-driven drone and the vehicle is H. o ;

[0121] According to the spacing between UAVs and airspace flow, the airspace flow basic graph model analysis function is given:

[0122]

[0123] The abnormal point information is analyzed through the analytical function to obtain the airspace flow and airspace congestion density, and the airspace flow and airspace congestion density are weighted and integrated to obtain the first parameter variable.

[0124] In this embodiment, the method of performing variable influence analysis on the relevant data to obtain the second parameter variable includes:

[0125] Construct a state matrix based on the relevant data, use the time of the relevant data as the node, and calculate the contribution of the node:

[0126]

[0127] The contribution of the a-th related data of the t-th node is ψ a (t), the flight score of the a-th related data of the t-th node is The flight score of the a-th related data of the t-1th node is The amount of relevant data for the tth node is The control coefficient of the a-th related data is γ a , the ath related data of the tth node is h a (t), the a-th related data of the t-1-th node is h a (t-1), the influence degree is σ, and the variable of the a-th related data of the t-th node is Δη a,t , the wth flight score index value of the ath related data of the tth node is V a,w (t), the number of flight score indicators is N, and the weight of the w-th flight score indicator is λ w ;

[0128] Calculate the average probability of a node:

[0129]

[0130] The average probability of the a-th related data of the t-th node is The number of nodes with the a-th related data is M, and the probability of the a-th related data appearing at the t-th node is ζ a (t);

[0131] Construct a weighted directed graph based on the directed weighted network and calculate the importance index of the node:

[0132]

[0133] The importance index of the tth node is ε u , the natural constant is e, and the weight of the outgoing connection of the tth node is C ot (u), the weight of the connection received by the t-th node is C in (u), the influence coefficient is υ;

[0134] Calculate the weighted value of the node:

[0135]

[0136] The weighted value of the tth node is The probability that the a-th related data of the t-th node is in an influential state is ζ a,yx (t), the weighting factor is λ;

[0137] Calculate the second argument:

[0138]

[0139] The flight score of the tth node is The standard value of the a-th related data is h ao The standard value of flight score is The second parameter is

[0140] In this embodiment, the method for constructing a joint function according to the first parameter and the second parameter includes:

[0141] Obtain key data based on importance screening of search methods and flight paths, and obtain flight evaluation indicators based on key data;

[0142] According to the constructed joint function, the expression is:

[0143]

[0144] The second parameter is The first parameter is The optimization coefficient is The actual arrival time is T st , agreed arrival time The number of the nth type of flying drones is B n , the energy consumption per kilometer of the nth type of flying UAV is g n , the flight speed of the nth flying drone is S n , the sign function is sgn(·), and the fixed cost of flight is R gd , the joint function is

[0145] In this embodiment, the method for constructing a target path optimization model according to the joint function includes:

[0146] The objective function is constructed based on the joint function, and the expression is:

[0147]

[0148] The control factor is χ, and the loss function is The objective function is

[0149] Target path optimization models include particle swarm optimization algorithm, random forest algorithm, and recurrent neural network;

[0150] The particle swarm optimization algorithm adjusts the search direction and position through information sharing and accumulation of individual experience among individual particles in the swarm, adapts to the joint function requirements of the flight, realizes efficient search and optimization in the multi-dimensional feature space, and obtains an improved flight plan;

[0151] The random forest algorithm constructs multiple decision trees and obtains prediction results based on the improved flight plan, capturing the nonlinear relationship of the prediction results and extracting effective features based on the nonlinear relationship;

[0152] The recurrent neural network updates the internal state at the sequence time step through a recursive loop structure, captures the dynamic features of the data to be optimized, and predicts the dynamic features based on the effective features to obtain the joint function value. Figure 2 This is a schematic diagram of the structure of an electronic device according to an embodiment of the present application. Figure 2 At the hardware level, the electronic device includes a processor and, optionally, an internal bus, a network interface, and memory. The memory may include internal memory, such as high-speed random-access memory (RAM), or non-volatile memory, such as at least one disk drive. Of course, the electronic device may also include other hardware required for its services.

[0153] The processor, network interface, and memory can be interconnected via an internal bus, which can be an ISA (Industry Standard Architecture) bus, a PCI (Peripheral Component Interconnect) bus, or an EISA (Extended Industry Standard Architecture) bus. The bus can be divided into an address bus, a data bus, a control bus, etc. For ease of representation, Figure 2 Only one bidirectional arrow is used in the diagram, but this does not mean that there is only one bus or one type of bus.

[0154] The memory is used to store programs. Specifically, the program may include program code, which includes computer operating instructions. The memory may include internal memory and non-volatile memory, and provides instructions and data to the processor.

[0155] The processor reads the corresponding computer program from the non-volatile memory into the internal memory and then runs it, forming an AI-based autonomous obstacle avoidance path planning device for drones at the logical level. The processor executes the program stored in the memory and is specifically configured to execute any of the aforementioned AI-based autonomous obstacle avoidance path planning methods for drones.

[0156] The above application Figure 1 The AI-based autonomous obstacle avoidance path planning method for a drone disclosed in the illustrated embodiments can be applied to or implemented by a processor. The processor may be an integrated circuit chip with signal processing capabilities. During implementation, each step of the above method can be completed by hardware integrated logic circuits in the processor or by software instructions. The above processor can be a general-purpose processor, including a central processing unit (CPU), a network processor (NP), etc.; it can also be a digital signal processor (DSP), an application-specific integrated circuit (ASIC), a field-programmable gate array (FPGA), or other programmable logic devices, discrete gate or transistor logic devices, or discrete hardware components. It can implement or execute the various methods, steps, and logic block diagrams disclosed in the embodiments of this application. The general-purpose processor can be a microprocessor or any conventional processor. The steps of the method disclosed in conjunction with the embodiments of this application can be directly implemented and executed by a hardware decoding processor, or by a combination of hardware and software modules in the decoding processor. The software module can be located in a storage medium well-known in the art, such as random access memory, flash memory, read-only memory, programmable read-only memory, electrically erasable programmable memory, registers, etc. The storage medium is located in the memory, and the processor reads the information in the memory and, in conjunction with its hardware, completes the steps of the above method.

[0157] The electronic device may also perform Figure 1 AI-based autonomous obstacle avoidance path planning method for UAVs, and realize Figure 1 The functions of the illustrated embodiment will not be described in detail in the embodiments of the present application.

[0158] An embodiment of the present application also proposes a computer-readable storage medium that stores one or more programs, which include instructions that, when executed by an electronic device including multiple applications, execute any of the aforementioned AI-based drone autonomous obstacle avoidance path planning methods.

[0159] Those skilled in the art will appreciate that the embodiments of the present application can be provided as methods, systems, or computer program products. Therefore, the present application can adopt the form of a complete hardware embodiment, a complete software embodiment, or an embodiment in combination with software and hardware. Moreover, the present application can adopt the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to magnetic disk storage, CD-ROM, optical storage, etc.) that contain computer-usable program code.

[0160] The present application is described with reference to the flowcharts and / or block diagrams of the methods, devices (systems), and computer program products according to the embodiments of the present application. It should be understood that each process and / or box in the flowchart and / or block diagram, as well as the combination of the processes and / or boxes in the flowchart and / or block diagram, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing device to produce a machine, so that the instructions executed by the processor of the computer or other programmable data processing device generate instructions for implementing the steps in the process. Figure 1 a process or multiple processes and / or boxes Figure 1 A device that provides the functions specified in a block or multiple blocks.

[0161] These computer program instructions may 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 produce an article of manufacture comprising an instruction device, which implements the process Figure 1 a process or multiple processes and / or boxes Figure 1 The function specified in one or more boxes.

[0162] These computer program instructions can also be loaded onto a computer or other programmable data processing device so that a series of operational steps are executed on the computer or other programmable device to produce a computer-implemented process, thereby providing the instructions executed on the computer or other programmable device for implementing the process. Figure 1 a process or multiple processes and / or boxes Figure 1 A step that specifies a function in one or more boxes.

[0163] In a typical configuration, a computing device includes one or more processors (CPUs), input / output interfaces, network interfaces, and memory.

[0164] Memory may include non-permanent storage in a computer-readable medium, random access memory (RAM) and / or non-volatile memory in the form of read-only memory (ROM) or flash RAM. Memory is an example of a computer-readable medium.

[0165] Computer-readable media includes permanent and non-permanent, removable and non-removable media that can be implemented by any method or technology to store information. The information can be computer-readable instructions, data structures, program modules or other data. Examples of computer storage media include, but are not limited to, phase change memory (PRAM), static random access memory (SRAM), dynamic random access memory (DRAM), other types of random access memory (RAM), read-only memory (ROM), electrically erasable programmable read-only memory (EEPROM), flash memory or other memory technology, compact disc read-only memory (CD-ROM), digital versatile disc (DVD) or other optical storage, magnetic cassettes, magnetic tape, magnetic disk storage or other magnetic storage devices or any other non-transmission media that can be used to store information that can be accessed by a computing device. As defined herein, computer-readable media does not include transitory computer-readable media (transitory media), such as modulated data signals and carrier waves.

[0166] It should also be noted that the terms "comprises," "includes," or any other variations thereof are intended to encompass non-exclusive inclusion, such that a process, method, commodity, or apparatus that includes a series of elements includes not only those elements but also other elements not explicitly listed, or includes elements inherent to such process, method, commodity, or apparatus. In the absence of further limitations, an element defined by the phrase "comprises a ..." does not exclude the presence of other identical elements in the process, method, commodity, or apparatus that includes the element.

[0167] Those skilled in the art will appreciate that the embodiments of the present application may be provided as methods, systems, or computer program products. Therefore, the present application may take the form of a complete hardware embodiment, a complete software embodiment, or an embodiment combining software and hardware. Furthermore, the present application may take the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to magnetic disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0168] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principles of the present invention should be included in the scope of protection of the present invention.

Claims

1. The AI-based autonomous obstacle avoidance path planning method for UAVs is characterized by: The following steps are involved: Collecting path data and related data of the target area, and preprocessing the path data and the related data; Performing anomaly screening on the path data according to a pass threshold to obtain abnormal point information, and performing obstacle passability analysis on the abnormal point information to obtain a first parameter variable; Performing a variable influence analysis on the relevant data to obtain a second parameter variable, and constructing a joint function based on the first parameter variable and the second parameter variable; A target path optimization model is constructed according to the joint function, the target path optimization model is constructed with the data to be optimized, and the optimization result is output.

2. The AI-based autonomous obstacle avoidance path planning method for unmanned aerial vehicles according to claim 1 is characterized in that: The method for obtaining abnormal point information by performing abnormal screening on the path data according to the pass threshold includes: The time points of the path data where the stopover and flight speed are less than 30 km / h are taken as problem data points. The information weight of the problem data points is calculated based on the data characteristics and importance: The i-th path data of the u-th problem data point is u i , the yth path data of the uth problem data point is u y , path data u i The probability of occurrence is Path data u y The probability of occurrence is Path data u i and path data u y The probability of simultaneous occurrence is Observed path data u y Time path data u i The posterior probability is The number of path data is Path data u i The information weight is The problem data points with information weight greater than 0.412 are taken as key data points, and the key data points are optimized: Calculate particle fitness: The probability parameter is sp, and the position of particle x is D x , position D x The information weight is The fitness function of particle x is Use roulette wheel to select excellent particles as the parent generation of crossover mutation and calculate the selection probability: The zth particle is x z , particle x z The fitness of The number of particles is Particle x z The probability of selection is Particle x z The position of D(x z ); Perform crossover mutation on particles and calculate the crossover rate and mutation rate: The crossover rate The mutation rate is The parameters between 0 and 1 are β1, β2, β3, and β4, and the average fitness of the particle population is The maximum fitness of the particle population is The fitness of the crossover individual is The fitness of the mutant individual is Reverse sort the gene segments of the particles after crossover mutation to obtain new particles, calculate the fitness of the new particles, and calculate the pass threshold based on the fitness: The pass threshold is η, and the new particle x new The fitness of The flying speed of the u-th problem data point is b u , the minimum flight speed of the u-th problem data point is b min ; Iterate continuously until the maximum number of iterations is reached, and output the key data points with a pass threshold greater than 0.374 as outlier information.

3. The AI-based autonomous obstacle avoidance path planning method for unmanned aerial vehicles according to claim 1 is characterized in that: The method for performing an obstacle capability analysis on the outlier information to obtain a first parameter variable includes: Input the outlier information into the obstacle passing capacity analysis model and calculate the airspace flow based on the outlier information: p=a·f a=(k) -1 The airspace flow is p, the blocking density is a, and the spacing between drones is k; Calculate the relationship between drone flow and speed: The relationship between drone flow and speed is ρ, the function curvature control parameter is μ, and the maximum density that causes airspace congestion is a. max , the maximum flight speed of the UAV is f v ; Calculate the distance between drones: The length of the drone body is The minimum safe distance of the drone is s o , the flight speed of the drone is f, and the free flow speed is f o , the safe distance of the drone during manual intervention is H o ; According to the spacing between UAVs and airspace flow, the airspace flow basic graph model analysis function is given: The abnormal point information is analyzed through the analytical function to obtain the airspace flow and airspace congestion density, and the airspace flow and airspace congestion density are weighted and integrated to obtain the first parameter variable.

4. The AI-based autonomous obstacle avoidance path planning method for unmanned aerial vehicles according to claim 1, characterized in that: The method of performing variable influence analysis on the relevant data to obtain a second parameter variable includes: Construct a state matrix based on the relevant data, use the time of the relevant data as the node, and calculate the contribution of the node: The contribution of the a-th related data of the t-th node is ψ a (t), the flight score of the a-th related data of the t-th node is The flight score of the a-th related data of the t-1th node is The amount of relevant data for the tth node is The control coefficient of the a-th related data is γ a , the coefficient for adjusting the degree of flight impact based on the relevant data, the a-th relevant data of the t-th node is h a (t), the a-th related data of the t-1-th node is h a (t-1), the influence degree is σ, and the variable of the a-th related data of the t-th node is Δη a,t , the wth flight score index value of the ath related data of the tth node is V a,w (t), the number of flight score indicators is N, and the weight of the w-th flight score indicator is λ w ; Calculate the average probability of a node: The average probability of the a-th related data of the t-th node is The number of nodes with the a-th related data is M, and the probability of the a-th related data appearing at the t-th node is ζ a (t); Construct a weighted directed graph based on the directed weighted network and calculate the importance index of the node: The importance index of the tth node is ε u , the natural constant is e, and the weight of the outgoing connection of the tth node is C ot (u), the weight of the connection received by the t-th node is C in (u), the influence coefficient is υ; Calculate the weighted value of the node: The weighted value of the tth node is The probability that the a-th related data of the t-th node is in an influential state is ζ a,yx (t), the weighting factor is λ; Calculate the second argument: The flight score of the tth node is The standard value of the a-th related data is h ao The standard value of flight score is The second parameter is 5. The AI-based autonomous obstacle avoidance path planning method for unmanned aerial vehicles according to claim 1, characterized in that: The method for constructing a joint function according to a first parameter and a second parameter includes: Obtain key data based on importance screening of search methods and flight paths, and obtain flight evaluation indicators based on key data; According to the constructed joint function, the expression is: The second parameter is The first parameter is The optimization coefficient is The actual arrival time is T st , agreed arrival time The number of the nth type of flying drones is B n , the energy consumption per kilometer of the nth type of flying UAV is g n , the flight speed of the nth flying drone is S n , the sign function is sgn(·), and the fixed cost of flight is R gd , the joint function is 6. The AI-based autonomous obstacle avoidance path planning method for unmanned aerial vehicles according to claim 1, characterized in that: The method for constructing a target path optimization model according to the joint function includes: The objective function is constructed based on the joint function, and the expression is: The control factor is χ, and the loss function is The objective function is Target path optimization models include particle swarm optimization algorithm, random forest algorithm, and recurrent neural network; The particle swarm optimization algorithm adjusts the search direction and position through information sharing and accumulation of individual experience among individual particles in the swarm, adapts to the joint function requirements of the flight, realizes efficient search and optimization in the multi-dimensional feature space, and obtains an improved flight plan; The random forest algorithm constructs multiple decision trees and obtains prediction results based on the improved flight plan, capturing the nonlinear relationship of the prediction results and extracting effective features based on the nonlinear relationship; The recurrent neural network updates the internal state at the sequence time step through a recursive loop structure, captures the dynamic features of the data to be optimized, and predicts the dynamic features based on the effective features to obtain the joint function value.

7. An electronic device comprising: processor; as well as A memory arranged to store computer executable instructions, which, when executed, cause the processor to perform the method according to any one of claims 1 to 6.

8. A computer-readable storage medium storing one or more programs, wherein when the one or more programs are executed by an electronic device including a plurality of application programs, the electronic device executes the method according to any one of claims 1 to 6.

Images