Multi-unmanned aerial vehicle autonomous cooperation task allocation method based on GWCA algorithm

Abstract

The invention discloses a multi-unmanned aerial vehicle autonomous cooperative task allocation method based on a GWCA algorithm, and particularly relates to the technical field of task allocation, which comprises the steps of training unmanned aerial vehicles in a fixed scene based on the GWCA algorithm, determining a task execution sequence and route planning of the unmanned aerial vehicles at different positions of the fixed scene, and determining the task execution sequence of the unmanned aerial vehicles; according to the current flight process data of each unmanned aerial vehicle and the historical training data of the unmanned aerial vehicle, determining the flight path information and flight risk information of a single unmanned aerial vehicle, obtaining the unmanned aerial vehicle with a problem, and carrying out task allocation on the other unmanned aerial vehicles by using the GWCA algorithm again. According to the multi-unmanned-aerial-vehicle task scheduling method, the matching degree information and the risk information of the cooperative work of the multiple unmanned aerial vehicles are collected through the overall analysis of the multiple unmanned aerial vehicles, the performance of the cooperative work of the multiple unmanned aerial vehicles is quantified, the task scheduling of the multiple unmanned aerial vehicles becomes more flexible and intelligent, and complex environments and task changes can be coped with.

Description

Technical Field

[0001] The present invention relates to the technical field of task allocation, and more specifically, to a multi-UAV autonomous collaborative task allocation method based on a GWCA algorithm. Background Art

[0002] With the rapid development of drone technology, multi-UAV systems (UAVs) are increasingly used in various application scenarios, especially in logistics, environmental monitoring, agricultural inspection, military reconnaissance and other fields. The collaborative work of multiple UAV systems can improve task execution efficiency, reduce costs and increase flexibility. In the existing technology, the actual flight track distance is replaced by the straight-line distance through the GWCA algorithm, so that drones can stably obtain better task allocation results. However, it is not easy to find the problematic drones among multiple UAVs, which may lead to a decrease in the success rate of task execution. In addition, it is not easy to analyze the reliability of the collaborative work of drones during the collaborative work of multiple UAVs.

[0003] In order to solve the above two defects, a technical solution is now provided. Summary of the invention

[0004] In order to overcome the above-mentioned defects of the prior art, an embodiment of the present invention provides a safety chip based on unmanned driving and an information interaction method thereof to solve the problems raised in the above-mentioned background technology.

[0005] To achieve the above object, the present invention provides the following technical solutions:

[0006] The multi-UAV autonomous collaborative task allocation method based on the GWCA algorithm specifically includes the following steps:

[0007] S1: Train the UAV in a fixed scene based on the GWCA algorithm to determine the task execution sequence and route planning of the UAV at different locations in the fixed scene;

[0008] S2: Determine the flight path information and flight risk information of a single UAV based on the current flight process data of each UAV and the historical training data of the UAV;

[0009] S3: Comprehensively analyze the flight path information and flight risk information of individual drones, identify the drones with problems, and reassign tasks to the remaining drones using the GWCA algorithm;

[0010] S4: In the process of multi-UAV collaborative work, through the overall analysis of multiple UAVs, the matching information and risk information of the multi-UAV collaborative work are collected, the performance of the multi-UAV collaborative work is quantified, and an alarm signal is generated.

[0011] In a preferred embodiment, determining the flight path information of a single UAV includes:

[0012] The flight path information of a single UAV is represented by the actual time deviation coefficient;

[0013] The logic for obtaining the actual time deviation coefficient is as follows: determine the preset route path when the drone performs the task through the GWCA algorithm, set special marking points in the preset route path, set a monitoring time period, obtain the special marking points passed by the drone during the monitoring time period, determine the preset time point when the drone arrives at the special marking point, and mark the preset time point when the drone arrives at the special marking point as: YS n , set the monitoring time period, obtain the special marking points that the drone passes during the monitoring time period, and determine the time point when the drone actually arrives at the special marking point, and mark the time point when the drone actually arrives at the special marking point as: SJ n ;

[0014] By comparing the preset time point with the actual time point, the actual time deviation coefficient is determined. The calculation formula of the actual time deviation coefficient is: Among them, PC sj is the actual time deviation coefficient.

[0015] In a preferred embodiment, determining the flight risk information of a single drone includes:

[0016] The flight risk information of a single drone is represented by the Bayesian probability coefficient;

[0017] The logic for obtaining the Bayesian probability coefficient is: determine the flight characteristics of the drone during flight, including power, speed, altitude, and wind speed, match the flight characteristics in the drone's historical training data in the scene, and determine the historical similar flight data through cosine similarity. The calculation formula is: Among them, A is the flight feature vector of the current UAV, B i is the flight feature vector in the historical training data, and XS is the flight feature vector of historical similar flight data;

[0018] Based on the historical similar flight data, the prior probability of the drone mission completion is determined, and the prior probability of the drone mission completion is marked as: P(WC), and the probability of the mission completion in the historical similar flight data is determined, and the probability of the mission completion in the historical similar flight data is marked as: P(B xs |WC);

[0019] The probability of historical similar flight data is determined using kernel density estimation based on historical training data, and the probability of historical similar flight data is marked as: P(B xs), the probability that the UAV can complete the task under the current flight characteristics is calculated by the Bayesian formula, and the probability that the UAV can complete the task is expressed by the Bayesian probability coefficient. The calculation formula is: Among them, GL kn is the Bayesian likelihood coefficient.

[0020] In a preferred embodiment, the flight path information and flight risk information of a single UAV are comprehensively analyzed, including:

[0021] The flight path information and flight risk information of a single UAV are comprehensively analyzed, and weighted calculation is performed through the actual time deviation coefficient and the Bayesian possibility probability coefficient to construct a single UAV evaluation model and generate a single UAV evaluation coefficient. The calculation formula of the single UAV evaluation coefficient is: Among them, pg q is the single UAV evaluation coefficient, q = 1, 2, 3, ..., Q, Q is a positive integer, q is the number of the UAV when multiple UAVs work together, α1 and α2 are the proportional coefficients of the actual time deviation coefficient and the Bayesian possibility probability coefficient, respectively, and α1 and α2 are both greater than 0;

[0022] Set the single UAV assessment coefficient threshold, compare the single UAV assessment coefficient with the single UAV assessment coefficient threshold, if the single UAV assessment coefficient is greater than the single UAV assessment coefficient threshold, no warning signal is generated, if the single UAV assessment coefficient is less than the single UAV assessment coefficient threshold, then the UAVs less than the single UAV assessment coefficient threshold will generate a warning signal, and the remaining UAVs greater than the single UAV assessment coefficient threshold will be automatically reassigned tasks using the GWCA algorithm.

[0023] In a preferred embodiment, collecting matching information of multiple UAVs working in collaboration includes:

[0024] The matching degree information of multiple UAVs working together is expressed by the collaborative matching degree coefficient;

[0025] The logic for obtaining the collaborative matching coefficient is as follows: obtaining the work progress of each UAV working in collaboration within the monitoring interval, the work progress of the UAV is determined based on the actual route obtained by the GWCA algorithm, and is determined according to the ratio of the navigation distance of the UAV in the actual route to the total distance of the actual route, and marking the work progress of each UAV working in collaboration within the monitoring interval as: JD q ,in, LC q is the actual flight distance of the qth UAV during the monitoring period, ZLC q is the total distance of the actual route of the qth UAV during the monitoring period;

[0026] The Pearson correlation coefficient is used to determine the correlation matrix of multiple UAVs working together in the monitoring interval, and the correlation matrix of multiple UAVs working together in the monitoring interval is marked as: R, Where j, k∈{1, 2, 3, …, Q}, R is a Q*Q matrix, and each element in the correlation matrix represents the Pearson correlation coefficient between drone j and drone k. M is a positive integer, m is the number of different moments in the monitoring interval, JD j,m is the work progress of the j-th UAV at the m-th moment, JD k,m is the work progress of the k-th UAV at the m-th moment, is the average working progress of the jth UAV in the monitoring interval, is the average working progress of the jth UAV in the monitoring interval;

[0027] Calculate the synergy matching coefficient, the calculation formula is: Among them, PP xt is the synergy matching coefficient.

[0028] In a preferred embodiment, collecting risk information of multiple drones working together includes:

[0029] The risk information of multi-UAV collaboration is expressed through the risk concealment coefficient;

[0030] The logic for obtaining the risk concealment coefficient is as follows: UAVs are divided into different types according to their usage performance in historical data, risk events and risk event results of different types of UAVs are determined, and different types of UAVs are scored based on the degree of impact of risk event results on UAV collaboration to obtain risk scores of UAVs;

[0031] According to the number of drone types in multi-drone collaborative work and the risk scores of different types of drones, a regression model is constructed to obtain the risk concealment coefficient. The expression of the risk concealment coefficient is: Among them, FX yn is the risk concealment coefficient, SL1, SL2, SL3, ..., SL W is the number of different drone types in multi-drone collaborative work, PF1, PF2, PF3, …, PF W is the risk score for different types of drones, and e is a natural number.

[0032] In a preferred embodiment, quantifying the performance of multiple UAVs working together includes:

[0033] Through comprehensive analysis of the matching information and risk information of multi-UAV collaborative work, the collaborative matching coefficient and risk concealment coefficient are weighted and calculated to build a collaborative work evaluation model and generate a collaborative work evaluation coefficient. The calculation formula of the collaborative work evaluation coefficient is: Among them, pg xt is the collaborative work evaluation coefficient, β1 and β2 are the proportional coefficients of the collaborative matching coefficient and the risk concealment coefficient, respectively, and β1 and β2 are both greater than 0.

[0034] In a preferred embodiment, generating an alarm signal includes:

[0035] A collaborative work evaluation coefficient threshold is set, and the collaborative work evaluation coefficient is compared with the collaborative work evaluation coefficient threshold. If the collaborative work evaluation coefficient is less than the collaborative work evaluation coefficient threshold, an alarm signal is generated to notify professional staff to stop the collaborative work of multiple drones to avoid risks during the work process. If the collaborative work evaluation coefficient is greater than the collaborative work evaluation coefficient threshold, no alarm signal is generated.

[0036] Technical effects and advantages of the present invention:

[0037] The present invention utilizes the GWCA algorithm to optimize the allocation of collaborative tasks for multiple UAVs, and combines real-time data and historical training data during flight to evaluate the path planning and flight risks of UAVs. By combining the GWCA algorithm with flight path and risk analysis, efficient collaborative work management of multiple UAVs can be achieved. By dynamically adjusting task allocation and quantifying collaborative work performance, potential risks can be discovered and responded to in a timely manner, which helps to make the task scheduling of multiple UAVs more flexible and intelligent, able to cope with complex environments and task changes, and improve the robustness of the system. BRIEF DESCRIPTION OF THE DRAWINGS

[0038] In order to facilitate understanding by those skilled in the art, the present invention is further described below in conjunction with the accompanying drawings;

[0039] Figure 1 The figure is a flow chart of the multi-UAV autonomous collaborative task allocation method based on the GWCA algorithm of the present invention. DETAILED DESCRIPTION

[0040] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0041] Example 1

[0042] Figure 1 The figure is a flow chart of the multi-UAV autonomous collaborative task allocation method based on the GWCA algorithm of the present invention, which specifically includes the following steps:

[0043] S1: Train the UAV in a fixed scene based on the GWCA algorithm to determine the task execution sequence and route planning of the UAV at different locations in the fixed scene;

[0044] S2: Determine the flight path information and flight risk information of a single UAV based on the current flight process data of each UAV and the historical training data of the UAV;

[0045] S3: Comprehensively analyze the flight path information and flight risk information of individual drones, identify the drones with problems, and reassign tasks to the remaining drones using the GWCA algorithm;

[0046] S4: In the process of multi-UAV collaborative work, through the overall analysis of multiple UAVs, the matching information and risk information of the multi-UAV collaborative work are collected, the performance of the multi-UAV collaborative work is quantified, and an alarm signal is generated.

[0047] The Great Wall Construction Algorithm (GWCA) is a heuristic optimization algorithm inspired by the competition and elimination mechanism among workers during the construction of the ancient Great Wall. It solves various constrained and unconstrained optimization problems by simulating the dynamic interaction between workers (or "individuals"). GWCA has the advantages of dynamicity, simple structure, high convergence performance, and good solution quality. It has demonstrated competitiveness and superiority in many optimization problems, especially for some complex and dynamically changing optimization problems.

[0048] In UAV applications, GWCA can be used for multi-UAV task allocation and trajectory planning. Specifically, it can solve the problems of how to allocate UAVs in multi-node tasks and how to dynamically adjust task allocation strategies, including:

[0049] In GWCA, each "worker" represents a solution, and during each iteration, these workers update their positions according to a specific motion model. Similarly, in multi-UAV task allocation, each "worker" can be regarded as a task allocation plan for a UAV, and the task allocation plan is updated through algorithm iteration;

[0050] Workers continuously improve their work efficiency through competition and elimination. In the drone task allocation, the task allocation plan will be continuously optimized through competition and elimination mechanisms, and the optimal task allocation plan will be finally selected;

[0051] Each drone updates its "position" based on the current mission status and target location in a manner similar to the worker movement in GWCA. In each iteration, the algorithm assigns a predefined "motion model" to each drone (worker), that is, dynamically adjusts the task allocation based on information such as flight status, mission priority, and remaining battery power;

[0052] The trajectory planning of UAVs can borrow the "labor movement model" in GWCA, and dynamically adjust the flight path by simulating different flight models (such as the shortest path model, efficient path model, etc.) to reduce energy consumption, avoid obstacles and improve task completion efficiency;

[0053] Considering that drones may be affected by the environment, obstacles or other emergencies during flight, GWCA assigns a random motion model to each worker, so that the drone can autonomously adjust the order of task execution and optimize the task execution path in the face of dynamic changes;

[0054] By simulating the dynamic process of workers, drones can automatically adjust their trajectories and task allocation under different environmental conditions. For example, when some drones cannot complete the task on time due to weather or battery power problems, other drones can take over their tasks.

[0055] Based on the GWCA algorithm, UAVs are trained in fixed scenes to determine the task execution sequence and route planning of UAVs at different locations in the fixed scenes. By using the actual route distance instead of the straight-line distance, the task allocation and trajectory planning are tightly coupled. That is, multiple UAVs are trained in a fixed scene, and the optimal solution for multiple UAVs to perform tasks in the fixed scene is gradually found through the optimization process. The flight path information and flight risk information of a single UAV are determined based on the current flight process data of each UAV and the historical training data of the UAV. The flight path information of a single UAV is represented by the actual time deviation coefficient, and the flight mode information of a single UAV is represented by the Bayesian possibility probability coefficient.

[0056] The logic for obtaining the actual time deviation coefficient is as follows: determine the preset route path when the drone performs the task through the GWCA algorithm, set special marking points in the preset route path, set a monitoring time period, obtain the special marking points passed by the drone during the monitoring time period, determine the preset time point when the drone arrives at the special marking point, and mark the preset time point when the drone arrives at the special marking point as: YS n , set the monitoring time period, obtain the special marking points that the drone passes during the monitoring time period, and determine the time point when the drone actually arrives at the special marking point, and mark the time point when the drone actually arrives at the special marking point as: SJ n ;

[0057] By comparing the preset time point with the actual time point, the actual time deviation coefficient is determined. The calculation formula of the actual time deviation coefficient is: Among them, PC sj is the actual time deviation coefficient.

[0058] It should be noted that the preset route path for drones when performing tasks is selected by the GWCA algorithm based on the different states of multiple drones after scene training. Different drones should travel according to the preset routes. Special marking points and monitoring time periods are set by professional staff. Usually there are multiple special marking points in the monitoring time period. Special marking points are usually marked in sections according to the length of the preset route.

[0059] The actual time deviation coefficient can obtain the actual arrival time of the drone in real time by monitoring the special marking points that the drone passes through on the preset route. By comparing this with the preset time points, deviations can be discovered and adjusted in a timely manner. This real-time feedback mechanism helps to quickly respond to the flight status of the drone and ensure the smooth progress of the mission.

[0060] The logic for obtaining the Bayesian probability coefficient is: determine the flight characteristics of the drone during flight, including power, speed, altitude, and wind speed, match the flight characteristics in the drone's historical training data in the scene, and determine the historical similar flight data through cosine similarity. The calculation formula is: Among them, A is the flight feature vector of the current UAV, B i is the flight feature vector in the historical training data, and XS is the flight feature vector of historical similar flight data;

[0061] It should be noted that the flight characteristics are determined by the characteristics of the UAV, and the flight characteristics determine the flight quality of the UAV. The flight quality is also affected by environmental factors such as wind speed. By matching the flight characteristics of the current UAV with the flight characteristics in the historical training data, the probability of the UAV successfully completing the current task can be determined.

[0062] Based on the historical similar flight data, the prior probability of the drone mission completion is determined, and the prior probability of the drone mission completion is marked as: P(WC), and the probability of the mission completion in the historical similar flight data is determined, and the probability of the mission completion in the historical similar flight data is marked as: P(B xs |WC);

[0063] The probability of historical similar flight data is determined using kernel density estimation based on historical training data, and the probability of historical similar flight data is marked as: P(B xs), the probability that the UAV can complete the task under the current flight characteristics is calculated by the Bayesian formula, and the probability that the UAV can complete the task is expressed by the Bayesian probability coefficient. The calculation formula is: Among them, GL kn is the Bayesian likelihood coefficient.

[0064] It should be noted that the standard for drones to complete tasks is set by professional staff. Usually, the degree of mission completion is used to determine whether the drone is completed on time. The Bayesian likelihood coefficient is updated through prior knowledge and observed data (such as current flight characteristics), and the probability estimate can be continuously adjusted as new data arrives. Therefore, when using the Bayesian formula to calculate the probability of mission completion under the current flight characteristics, the relationship between the current flight status and historical data can be dynamically considered to adjust the probability of mission completion in real time.

[0065] The flight path information and flight risk information of a single UAV are comprehensively analyzed, and weighted calculation is performed through the actual time deviation coefficient and the Bayesian possibility probability coefficient to construct a single UAV evaluation model and generate a single UAV evaluation coefficient. The calculation formula of the single UAV evaluation coefficient is: Among them, pg q is the single UAV evaluation coefficient, q=1, 2, 3, …, Q, Q is a positive integer, q is the number of the UAV when multiple UAVs work together, α1 and α2 are the proportional coefficients of the actual time deviation coefficient and the Bayesian possibility probability coefficient, respectively, and α1 and α2 are both greater than 0.

[0066] It can be seen from the formula that the smaller the actual time deviation coefficient and the larger the Bayesian possibility probability coefficient, the larger the single UAV evaluation coefficient is, which means that the UAV performs better when multiple UAVs work together, and the possibility that the UAV will have problems and fail to complete the task is low. Conversely, the larger the actual time deviation coefficient and the smaller the Bayesian possibility probability coefficient, the smaller the single UAV evaluation coefficient is, which means that the UAV performs poorly when multiple UAVs work together.

[0067] Set the single UAV assessment coefficient threshold, compare the single UAV assessment coefficient with the single UAV assessment coefficient threshold, if the single UAV assessment coefficient is greater than the single UAV assessment coefficient threshold, no warning signal is generated, if the single UAV assessment coefficient is less than the single UAV assessment coefficient threshold, then the UAVs less than the single UAV assessment coefficient threshold will generate a warning signal, and the remaining UAVs greater than the single UAV assessment coefficient threshold will be automatically reassigned tasks using the GWCA algorithm.

[0068] In the process of multi-UAV collaborative work, the matching information and risk information of multi-UAV collaborative work are collected through the overall analysis of multi-UAVs. The matching information of multi-UAV collaborative work is represented by the collaborative matching coefficient, and the risk information of multi-UAV collaborative work is represented by the risk concealment coefficient.

[0069] The logic for obtaining the collaborative matching coefficient is as follows: obtaining the work progress of each UAV working in collaboration within the monitoring interval, the work progress of the UAV is determined based on the actual route obtained by the GWCA algorithm, and is determined according to the ratio of the navigation distance of the UAV in the actual route to the total distance of the actual route, and marking the work progress of each UAV working in collaboration within the monitoring interval as: JD q ,in, LC q is the actual flight distance of the qth UAV during the monitoring period, ZLC q is the total distance of the actual route of the qth UAV during the monitoring period;

[0070] The Pearson correlation coefficient is used to determine the correlation matrix of multiple UAVs working together in the monitoring interval, and the correlation matrix of multiple UAVs working together in the monitoring interval is marked as: R, Where j, k∈{1, 2, 3, …, Q}, R is a Q*Q matrix, and each element in the correlation matrix represents the Pearson correlation coefficient between drone j and drone k. M, M is a positive integer, m is the number of different moments in the monitoring interval, JD j,m is the work progress of the j-th UAV at the m-th moment, JD k,m is the work progress of the k-th UAV at the m-th moment, is the average working progress of the jth UAV in the monitoring interval, is the average working progress of the jth UAV in the monitoring interval;

[0071] Calculate the synergy matching coefficient, the calculation formula is: Among them, PP xt is the synergy matching coefficient.

[0072] It should be noted that the collaborative matching coefficient can quickly evaluate the coordination between UAVs. By dynamically updating the correlation matrix, the collaborative performance of UAVs can be tracked in real time. If environmental conditions (such as wind speed, temperature, etc.) or mission requirements change, the correlation coefficient can quickly reflect these changes and help decision makers make adaptive adjustments to the current UAV mission execution status. If the collaborative matching coefficient of multiple UAVs is low when working together, it means that the collaborative performance of multiple UAVs is poor.

[0073] The logic for obtaining the risk concealment coefficient is as follows: UAVs are divided into different types according to their usage performance in historical data, risk events and risk event results of different types of UAVs are determined, and different types of UAVs are scored based on the degree of impact of risk event results on UAV collaboration to obtain risk scores of UAVs;

[0074] It should be noted that for different types of drones, the risk events that occurred in their historical missions and the possible consequences of these risk events are analyzed. Common risk events include failure to arrive on time, failure to complete the mission, etc. According to the usage performance in historical data, drones can be divided into different types. Generally, the types of drones can be classified according to their performance, usage frequency, environmental adaptability and other characteristics.

[0075] According to the number of drone types in multi-drone collaborative work and the risk scores of different types of drones, a regression model is constructed to obtain the risk concealment coefficient. The expression of the risk concealment coefficient is: Among them, FX yn is the risk concealment coefficient, SL1, SL2, SL3, ..., SL W is the number of different drone types in multi-drone collaborative work, PF1, PF2, PF3, …, PF W is the risk score for different types of drones, and e is a natural number.

[0076] It should be noted that by analyzing risk events of different types of drones through historical data, it is possible to more accurately identify the risks that each type of drone may face. For different risk events (such as failure to arrive on time, uncompleted missions, etc.), scores are given based on the drone's performance, environmental adaptability, frequency of use and other characteristics. The impact of each risk event can be quantified, which makes risk assessment more personalized and targeted, rather than just a unified standard.

[0077] In a scenario where multiple drones work together, the risk score of each drone not only reflects its own risk, but also its potential impact on the collaborative mission. By evaluating the risk scores of different types of drones, it is possible to understand the risk level of the overall mission and help optimize task allocation.

[0078] Through comprehensive analysis of the matching information and risk information of multi-UAV collaborative work, the collaborative matching coefficient and risk concealment coefficient are weighted and calculated to build a collaborative work evaluation model and generate a collaborative work evaluation coefficient. The calculation formula of the collaborative work evaluation coefficient is: Among them, pg xtis the collaborative work evaluation coefficient, β1 and β2 are the proportional coefficients of the collaborative matching coefficient and the risk concealment coefficient, respectively, and β1 and β2 are both greater than 0.

[0079] It can be seen from the formula that the larger the collaborative matching coefficient and the smaller the risk concealment coefficient, the larger the collaborative work evaluation coefficient, which means the better the performance of multi-UAV collaborative work. Conversely, the smaller the collaborative matching coefficient and the larger the risk concealment coefficient, the smaller the collaborative work evaluation coefficient, which means the worse the performance of multi-UAV collaborative work.

[0080] A collaborative work evaluation coefficient threshold is set, and the collaborative work evaluation coefficient is compared with the collaborative work evaluation coefficient threshold. If the collaborative work evaluation coefficient is less than the collaborative work evaluation coefficient threshold, an alarm signal is generated to notify professional staff to stop the collaborative work of multiple drones to avoid risks during the work process. If the collaborative work evaluation coefficient is greater than the collaborative work evaluation coefficient threshold, no alarm signal is generated.

[0081] The present invention utilizes the GWCA algorithm to optimize the allocation of collaborative tasks for multiple UAVs, and combines real-time data and historical training data during flight to evaluate the path planning and flight risks of UAVs. By combining the GWCA algorithm with flight path and risk analysis, efficient collaborative work management of multiple UAVs can be achieved. By dynamically adjusting task allocation and quantifying collaborative work performance, potential risks can be discovered and responded to in a timely manner, which helps to make the task scheduling of multiple UAVs more flexible and intelligent, able to cope with complex environments and task changes, and improve the robustness of the system.

[0082] The above formulas are all dimensionless and numerical calculations. The formula is a formula for the most recent real situation obtained by collecting a large amount of data and performing software simulation. The preset parameters in the formula are set by technicians in this field according to actual conditions.

[0083] The above embodiments can be implemented in whole or in part by software, hardware, firmware or any other combination. When implemented by software, the above embodiments can be implemented in whole or in part in the form of a computer program product. The computer program product includes one or more computer instructions or computer programs. When the computer instructions or computer programs are loaded or executed on a computer, the process or function described in the embodiment of the present application is generated in whole or in part. The computer can be a general-purpose computer, a special-purpose computer, a computer network, or other programmable device. The computer instructions can be stored in a computer-readable storage medium, or transmitted from one computer-readable storage medium to another computer-readable storage medium. For example, the computer instructions can be transmitted from one website site, computer, server or data center to another website site, computer, server or data center by wired or wireless (e.g., infrared, wireless, microwave, etc.). The computer-readable storage medium can be any available medium that a computer can access or a data storage device such as a server or data center that contains one or more available media sets. The available medium can be a magnetic medium (e.g., a floppy disk, a hard disk, a tape), an optical medium (e.g., a DVD), or a semiconductor medium. The semiconductor medium can be a solid-state hard disk.

[0084] It should be understood that in the various embodiments of the present application, the size of the serial numbers of the above-mentioned processes does not mean the order of execution. The execution order of each process should be determined by its function and internal logic, and should not constitute any limitation on the implementation process of the embodiments of the present application.

[0085] In the several embodiments provided in the present application, it should be understood that the disclosed systems, devices and methods can be implemented in other ways. For example, the device embodiments described above are only schematic. For example, the division of the units is only a logical function division. There may be other division methods in actual implementation, such as multiple units or components can be combined or integrated into another system, or some features can be ignored or not executed. Another point is that the mutual coupling or direct coupling or communication connection shown or discussed can be through some interfaces, indirect coupling or communication connection of devices or units, which can be electrical, mechanical or other forms.

[0086] If the functions are implemented in the form of software functional units and sold or used as independent products, they can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present application can be essentially or partly embodied in the form of a software product that contributes to the prior art. The computer software product is stored in a storage medium and includes several instructions for a computer device (which can be a personal computer, a server, or a network device, etc.) to perform all or part of the steps of the methods described in the various embodiments of the present application. The aforementioned storage media include: various media that can store program codes, such as USB flash drives, mobile hard disks, read-only memories (ROM), random access memories (RAM), magnetic disks or optical disks.

[0087] The above is only a specific implementation of the present application, but the protection scope of the present application is not limited thereto. Any technician familiar with the technical field can easily think of changes or substitutions within the technical scope disclosed in the present application, which should be included in the protection scope of the present application. Therefore, the protection scope of the present application should be based on the protection scope of the claims.

Claims

1. The multi-UAV autonomous collaborative task allocation method based on GWCA algorithm is characterized by: The specific steps include: S1: Train the UAV in a fixed scene based on the GWCA algorithm to determine the task execution sequence and route planning of the UAV at different locations in the fixed scene; S2: Determine the flight path information and flight risk information of a single UAV based on the current flight process data of each UAV and the historical training data of the UAV; S3: Comprehensively analyze the flight path information and flight risk information of individual drones, identify the drones with problems, and reassign tasks to the remaining drones using the GWCA algorithm; S4: In the process of multi-UAV collaborative work, through the overall analysis of multiple UAVs, the matching information and risk information of the multi-UAV collaborative work are collected, the performance of the multi-UAV collaborative work is quantified, and an alarm signal is generated.

2. The multi-UAV autonomous collaborative task allocation method based on the GWCA algorithm according to claim 1 is characterized in that: Determine the flight path information of a single drone, including: The flight path information of a single UAV is represented by the actual time deviation coefficient; The logic for obtaining the actual time deviation coefficient is as follows: determine the preset route path when the drone performs the task through the GWCA algorithm, set special marking points in the preset route path, set a monitoring time period, obtain the special marking points passed by the drone during the monitoring time period, determine the preset time point when the drone arrives at the special marking point, and mark the preset time point when the drone arrives at the special marking point as: YS n , set the monitoring time period, obtain the special marking points that the drone passes during the monitoring time period, and determine the time point when the drone actually arrives at the special marking point, and mark the time point when the drone actually arrives at the special marking point as: SJ n ; By comparing the preset time point with the actual time point, the actual time deviation coefficient is determined. The calculation formula of the actual time deviation coefficient is: Among them, PC sj is the actual time deviation coefficient.

3. The multi-UAV autonomous collaborative task allocation method based on the GWCA algorithm according to claim 2 is characterized in that: Determine the flight risk information of a single drone, including: The flight risk information of a single UAV is represented by the Bayesian probability coefficient; The logic for obtaining the Bayesian probability coefficient is: determine the flight characteristics of the drone during flight, including power, speed, altitude, and wind speed, match the flight characteristics in the drone's historical training data in the scene, and determine the historical similar flight data through cosine similarity. The calculation formula is: Among them, A is the flight feature vector of the current UAV, B i is the flight feature vector in the historical training data, and XS is the flight feature vector of historical similar flight data; Based on the historical similar flight data, the prior probability of the drone mission completion is determined, and the prior probability of the drone mission completion is marked as: P(WC), and the probability of the mission completion in the historical similar flight data is determined, and the probability of the mission completion in the historical similar flight data is marked as: P(B xs |WC); The probability of historical similar flight data is determined using kernel density estimation based on historical training data, and the probability of historical similar flight data is marked as: P(B xs ), the probability that the UAV can complete the task under the current flight characteristics is calculated by the Bayesian formula, and the probability that the UAV can complete the task is expressed by the Bayesian probability coefficient. The calculation formula is: Among them, GL kn is the Bayesian likelihood probability coefficient.

4. The multi-UAV autonomous collaborative task allocation method based on GWCA algorithm according to claim 3 is characterized in that: Comprehensively analyze the flight path information and flight risk information of a single drone, including: The flight path information and flight risk information of a single UAV are comprehensively analyzed, and weighted calculation is performed through the actual time deviation coefficient and the Bayesian possibility probability coefficient to construct a single UAV evaluation model and generate a single UAV evaluation coefficient. The calculation formula of the single UAV evaluation coefficient is: Among them, pg q is the single UAV evaluation coefficient, q = 1, 2, 3, ..., Q, Q is a positive integer, q is the number of the UAV when multiple UAVs work together, α1 and α2 are the proportional coefficients of the actual time deviation coefficient and the Bayesian possibility probability coefficient, respectively, and α1 and α2 are both greater than 0; Set the single UAV assessment coefficient threshold, compare the single UAV assessment coefficient with the single UAV assessment coefficient threshold, if the single UAV assessment coefficient is greater than the single UAV assessment coefficient threshold, no warning signal is generated, if the single UAV assessment coefficient is less than the single UAV assessment coefficient threshold, then the UAVs less than the single UAV assessment coefficient threshold will generate a warning signal, and the remaining UAVs greater than the single UAV assessment coefficient threshold will be automatically reassigned tasks using the GWCA algorithm.

5. The multi-UAV autonomous collaborative task allocation method based on the GWCA algorithm according to claim 4 is characterized in that: Collect matching information of multiple drones working together, including: The matching degree information of multiple UAVs working together is expressed by the collaborative matching degree coefficient; The logic for obtaining the collaborative matching coefficient is as follows: obtaining the work progress of each UAV working in collaboration within the monitoring interval, the work progress of the UAV is determined based on the actual route obtained by the GWCA algorithm, and is determined according to the ratio of the navigation distance of the UAV in the actual route to the total distance of the actual route, and marking the work progress of each UAV working in collaboration within the monitoring interval as: JD q ,in, LC q is the actual flight distance of the qth UAV during the monitoring period, ZLC q is the total distance of the actual route of the qth UAV during the monitoring period; The Pearson correlation coefficient is used to determine the correlation matrix of multiple UAVs working together in the monitoring interval, and the correlation matrix of multiple UAVs working together in the monitoring interval is marked as: R, Where j, k∈{1, 2, 3, …, Q}, R is a Q*Q matrix, and each element in the correlation matrix represents the Pearson correlation coefficient between drone j and drone k. m=1, 2, 3, ..., M, M is a positive integer, m is the number of different time in the monitoring interval, JD j,m is the work progress of the j-th UAV at the m-th moment, JD k,m is the work progress of the k-th UAV at the m-th moment, is the average working progress of the jth UAV in the monitoring interval, is the average working progress of the jth UAV in the monitoring interval; Calculate the synergy matching coefficient, the calculation formula is: Among them, PP xt is the synergy matching coefficient.

6. The multi-UAV autonomous collaborative task allocation method based on GWCA algorithm according to claim 5 is characterized in that: Collect risk information of multiple drones working together, including: The risk information of multi-UAV collaboration is expressed through the risk concealment coefficient; The logic for obtaining the risk concealment coefficient is as follows: UAVs are divided into different types according to their usage performance in historical data, risk events and risk event results of different types of UAVs are determined, and different types of UAVs are scored based on the degree of impact of risk event results on UAV collaboration to obtain risk scores of UAVs; According to the number of drone types in multi-drone collaborative work and the risk scores of different types of drones, a regression model is constructed to obtain the risk concealment coefficient. The expression of the risk concealment coefficient is: Among them, FX yn is the risk concealment coefficient, SL1, SL2, SL3, ..., SL W is the number of different drone types in multi-drone collaborative work, PF1, PF2, PF3, …, PF W is the risk score for different types of drones, and e is a natural number.

7. The multi-UAV autonomous collaborative task allocation method based on GWCA algorithm according to claim 6 is characterized in that: Quantify the performance of multi-UAV collaboration, including: Through the comprehensive analysis of the matching information and risk information of multi-UAV collaborative work, the collaborative matching coefficient and the risk concealment coefficient are weighted and calculated to build a collaborative work evaluation model and generate a collaborative work evaluation coefficient. The calculation formula of the collaborative work evaluation coefficient is: Among them, pg xt is the collaborative work evaluation coefficient, β1 and β2 are the proportional coefficients of the collaborative matching coefficient and the risk concealment coefficient, respectively, and β1 and β2 are both greater than 0.

8. The multi-UAV autonomous collaborative task allocation method based on GWCA algorithm according to claim 7 is characterized in that: Generate alarm signals, including: A collaborative work evaluation coefficient threshold is set, and the collaborative work evaluation coefficient is compared with the collaborative work evaluation coefficient threshold. If the collaborative work evaluation coefficient is less than the collaborative work evaluation coefficient threshold, an alarm signal is generated to notify professional staff to stop the collaborative work of multiple drones to avoid risks during the work process. If the collaborative work evaluation coefficient is greater than the collaborative work evaluation coefficient threshold, no alarm signal is generated.

Images