Unmanned aerial vehicle operation risk mitigation effectiveness evaluation method based on Bayesian network

Abstract

The invention discloses an unmanned aerial vehicle operation risk mitigation efficiency evaluation method based on a Bayesian network, and belongs to the technical field of unmanned aerial vehicle risk evaluation, and the method comprises the following steps: obtaining an unmanned aerial vehicle operation accident report, and carrying out the recognition of risk factors in the unmanned aerial vehicle operation accident report; according to risk factors in the unmanned aerial vehicle operation accident report, constructing an accident causal influence graph; constructing a Bayesian belief network according to the accident causal influence graph; and mitigation measures are correspondingly added to the risk factors in the Bayesian belief network, and the risk mitigation efficiency of all mitigation measures on the accident occurrence is obtained by estimating and updating the risk factors and the accident occurrence probability. According to the invention, the problem that the unmanned aerial vehicle operation risk mitigation efficiency evaluation is not accurate and comprehensive enough is solved.

Description

Technical Field

[0001] The present invention belongs to the technical field of unmanned aerial vehicle risk assessment, and in particular relates to a method for evaluating the effectiveness of unmanned aerial vehicle operation risk mitigation based on a Bayesian network. Background Art

[0002] The effectiveness evaluation of risk mitigation measures aims to verify whether the currently implemented measures have achieved the expected results. It can provide a basis for timely adjustment of strategies to avoid risk accumulation caused by inefficient measures when risk mitigation measures fail to reduce risks as expected.

[0003] Drone operation risk assessment can be divided into two methods: quantitative and qualitative. The qualitative risk assessment process allows drone operators to conduct initial risk assessments on their operating scenarios according to their processes, thereby ensuring the safety of drone operations. However, qualitative drone risk assessment relies on subjective judgment and expert opinions. Different assessors may have different interpretations of risks, which is prone to bias and limitations. Therefore, qualitative assessments cannot accurately quantify risks, make it difficult to accurately compare risks or prioritize risks, and, due to different airspace operation conditions, qualitative risk assessments may affect the accuracy of the final assessment results, which may ultimately hinder drone decision-making and airspace resource allocation. Quantitative drone operation risk assessment is achieved through probabilistic estimation, while allowing the system to be analyzed in a corrected manner. Most safety assessments are achieved through qualitative analysis, but such assessment methods are incomplete and inaccurate, and quantitative supplementary analysis is still needed. Summary of the invention

[0004] In view of the above-mentioned deficiencies in the prior art, the present invention provides a method for evaluating the risk mitigation effectiveness of UAV operations based on a Bayesian network, which solves the problem that the risk mitigation effectiveness evaluation of UAV operations is not accurate and comprehensive enough.

[0005] In order to achieve the above-mentioned object of the invention, the technical solution adopted by the present invention is:

[0006] The present invention provides a method for evaluating the effectiveness of UAV operation risk mitigation based on a Bayesian network, comprising the following steps:

[0007] S1. Obtain UAV operation accident reports and identify the risk factors in the UAV operation accident reports;

[0008] S2. Construct an accident cause-effect impact diagram based on the risk factors in the UAV operation accident report;

[0009] S3. Construct a Bayesian belief network based on the accident causal impact diagram;

[0010] S4. Add mitigation measures to the corresponding risk factors in the Bayesian belief network, and obtain the risk mitigation effectiveness of all mitigation measures for accidents by estimating and updating the risk factors and the probability of accidents.

[0011] Furthermore, the risk factors in the UAV operation accident report include operating environment factors, operation factors and facility and equipment factors.

[0012] Furthermore, S3 comprises the following steps:

[0013] S31. Obtaining an accident causal impact diagram;

[0014] S32. Construct a single factor influence conditional probability model to calculate the conditional probability that a single risk factor in the accident causal influence diagram leads to a UAV operation accident;

[0015] S33. Based on the single-factor influence conditional probability model, a multi-factor influence conditional probability model is constructed to calculate the conditional probability of multiple dangerous factors leading to UAV operation accidents;

[0016] S34, constructing shape parameter optimization constraints according to a multi-factor influence conditional probability model, and calculating a target first shape parameter and a target second shape parameter;

[0017] S35. Based on the conditional probability of a single hazardous factor leading to a UAV operation accident, the conditional probability of multiple hazardous factors leading to a UAV operation accident, the target first shape parameter and the target second shape parameter, the construction of the Bayesian belief network is completed.

[0018] Furthermore, the calculation expression of the single factor influence conditional probability model in S32 is as follows:

[0019] ,

[0020] in, Indicates the presence of risk factors Accidents The conditional probability of Indicates an accident and risk factors The joint probability of simultaneous occurrence, Indicates risk factors Probability of occurrence.

[0021] Furthermore, the calculation expression of the multi-factor influence conditional probability model in S33 is as follows:

[0022] ,

[0023] , ,

[0024] in, Indicates that an accident occurs when n risk factors exist at the same time The conditional probability of Indicates the first risk factor, Indicates the second risk factor, represents the nth risk factor, represents the conditional probability ratio of multi-factor accidents, represents the first shape parameter, represents the second shape parameter, represents the differentiation with respect to time t, represents the beta function with respect to the first shape parameter and the second shape parameter, Indicates the accidents occurred under the first n risk factors The sum of the conditional probabilities of Indicates the total number of m risk factors that may cause an accident The sum of the conditional probabilities of represents the gamma function, represents the value of the gamma function at the first shape parameter, represents the value of the gamma function at the second shape parameter, represents the value of the gamma function at the sum of the first shape parameter and the second shape parameter, where and Both are greater than 0, m>n, and m and n are positive integers.

[0025] Furthermore, the calculation expression of the shape parameter optimization constraint in S34 is as follows:

[0026] ,

[0027] ,

[0028] Among them, min means taking the minimum value, represents the estimated probability of multi-factor accidents, Represents the true probability of a multi-factor accident, where k is a positive integer.

[0029] Furthermore, the Bayesian belief network is a directed acyclic graph consisting of a set of links and probability distributions on a number of nodes; each of the nodes corresponds to the probability distribution of the accident occurrence state and the probability distribution of the occurrence state of each risk factor.

[0030] Furthermore, the S4 comprises the following steps:

[0031] S41, corresponding the probability distribution of the hazardous factor occurrence state and the probability distribution of the accident occurrence state corresponding to each node in the Bayesian belief network as the prior probability of the hazardous factor occurrence state and the prior probability of the accident occurrence state;

[0032] S42, setting mitigation measures to nodes corresponding to risk factors in the Bayesian belief network to obtain a mitigation effectiveness evaluation Bayesian network;

[0033] S43. Obtain other drone accident cases caused by the risk factors corresponding to the mitigation measures, and based on the prior probability of the risk factor occurrence state, use the mitigation effectiveness evaluation Bayesian belief network to estimate and update the risk factor occurrence state probability distribution and accident occurrence state probability distribution along the link for the nodes corresponding to the risk factor, as the risk factor occurrence state update probability and accident occurrence state update probability;

[0034] S44. Based on the updated probability of the accident occurrence state and the prior probability of the accident occurrence state, the risk mitigation effectiveness of all mitigation measures for the accident occurrence is calculated.

[0035] Furthermore, the calculation expression of the probability of updating the state of the risk factor in S43 is as follows:

[0036] ,

[0037] in, It represents the probability of updating the state of the risk factor corresponding to the jth node after the kth mitigation measure is added, represents the likelihood of adding the kth mitigation measure when the risk factor corresponding to the jth node exists, represents the prior probability of the occurrence state of the risk factor corresponding to the jth node, It represents the probability of other drone accident cases caused by the hazardous factors corresponding to the kth mitigation measure, where k is a positive integer.

[0038] Furthermore, the calculation expression of the risk mitigation effectiveness of all the mitigation measures for the occurrence of accidents is as follows:

[0039] ,

[0040] in, It indicates the risk mitigation effectiveness of all mitigation measures for accidents. represents the probability of updating the accident status. Represents the prior probability of the accident occurrence state.

[0041] The beneficial effects of the present invention are as follows: a method for evaluating the effectiveness of risk mitigation for UAV operation based on a Bayesian network is provided by the present invention, which realizes a comprehensive analysis of accidents caused by hazardous factors during UAV operation through a qualitative evaluation of functional safety combined with a quantitative analysis; the present invention reflects the correlation mechanism between the probability of occurrence of hazardous factors and between the probability of occurrence of hazardous factors and the probability of occurrence of accidents by constructing a Bayesian belief network, and quantifies the probability of the impact of each risk factor on the risk of accidents in UAV operation. At the same time, the degree of reduction of the risk of occurrence of hazardous factors and the risk of occurrence of accidents after adding mitigation measures based on the Bayesian belief network is analyzed, and the risk mitigation effectiveness of mitigation measures adopted for the operation of UAVs in airport terminal areas can be accurately and comprehensively evaluated.

[0042] Other advantages of the present invention will be analyzed in more detail in subsequent embodiments. BRIEF DESCRIPTION OF THE DRAWINGS

[0043] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the following briefly introduces the drawings required for use in the embodiments. It should be understood that the following drawings only show certain embodiments of the present invention and therefore should not be regarded as limiting the scope. For ordinary technicians in this field, other related drawings can be obtained based on these drawings without creative work.

[0044] Figure 1 The present invention is a flowchart of a method for evaluating the effectiveness of UAV operation risk mitigation based on a Bayesian network in an embodiment of the present invention.

[0045] Figure 2 Schematic diagram of nodes of a Bayesian belief network in an embodiment of the present invention.

[0046] Figure 3 Schematic diagram of the probability distribution of risk factors and accidents corresponding to each node in the Bayesian belief network in an embodiment of the present invention.

[0047] Figure 4 Schematic diagram of nodes of a Bayesian network for mitigation effectiveness evaluation in an embodiment of the present invention.

[0048] Figure 5 It is a schematic diagram of the updated probability distribution of the risk factors and accidents corresponding to each node in the mitigation effectiveness evaluation Bayesian network after adding mitigation measures in an embodiment of the present invention. DETAILED DESCRIPTION

[0049] The technical solutions in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all the embodiments. The components of the embodiments of the present invention generally described and shown in the drawings here can be arranged and designed in various different configurations. Therefore, the following detailed description of the embodiments of the present invention provided in the drawings is not intended to limit the scope of the claimed invention, but merely represents the selected embodiments of the present invention. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without making creative work belong to the scope of protection of the present invention.

[0050] like Figure 1 As shown, in one embodiment of the present invention, the present invention provides a method for evaluating the effectiveness of UAV operation risk mitigation based on a Bayesian network, comprising the following steps:

[0051] S1. Obtain UAV operation accident reports and identify the risk factors in the UAV operation accident reports;

[0052] In this scheme, drone operation accident reports include accident reports of drones operating near airport terminal areas, as well as accident reports in the database provided by the drone accident website as a supplement; drone accident reports include scenarios of single accident cases or complex scenarios containing common risk factors in multiple accident cases.

[0053] The risk factors in the UAV operation accident report include operating environment factors, operation factors and facility and equipment factors.

[0054] In this plan, operating environment factors include weather conditions, airspace factors, obstacle factors, bird strike factors, etc.; human factors include operational errors, violations of airspace regulations, technical deficiencies, etc.; facility and equipment factors include electromagnetic interference factors, remote communication failure factors, braking system failure factors, flight control system failure factors, navigation capability failure factors, insufficient power factors, etc.

[0055] S2. Construct an accident cause-effect impact diagram based on the risk factors in the UAV operation accident report;

[0056] In this scheme, the accident causal influence diagram is used to depict the interaction between each risk factor and the corresponding accident, providing a basis for constructing a Bayesian belief network.

[0057] S3. Construct a Bayesian belief network based on the accident causal impact diagram;

[0058] In this scheme, the Bayesian belief network is used to simulate the complex interaction of risk factors that lead to accidents in drone aviation systems. The Bayesian belief network is suitable for decomposing the occurrence of accidents into a series of relatively simple risk factors through probabilistic reasoning to solve the complex joint probability distribution problem, thereby solving the problem of lack of drone operation data and uncertainty.

[0059] The S3 comprises the following steps:

[0060] S31. Obtaining an accident causal impact diagram;

[0061] In this embodiment, the accident report of drone aviation safety and the ASN database are combined to estimate the prior probability and conditional probability of the risk factors. Since the number of accidents reported in the ASN database is limited, the data cannot cover the probability of all different types of events or all accidents. Therefore, the conditional probability derived only by the data in the database may make the result inaccurate. Therefore, this solution provides a conditional probability calculation model to calculate the conditional probability affected by multiple risk factors. ASN, Aviation Safety Network is a professional website in the field of aviation safety and an important source of information in the field of aviation safety. The ASN database contains a lot of detailed information on aviation accidents, aircraft failures, hijacking incidents, etc., aiming to provide a comprehensive and reliable information platform for the public, airlines, regulators and researchers.

[0062] S32. Construct a single factor influence conditional probability model to calculate the conditional probability that a single risk factor in the accident causal influence diagram leads to a UAV operation accident;

[0063] The calculation expression of the single factor influence conditional probability model in S32 is as follows:

[0064] ,

[0065] in, Indicates the presence of risk factors Accidents The conditional probability of Indicates an accident and risk factors The joint probability of simultaneous occurrence, Indicates risk factors Probability of occurrence.

[0066] In this scenario, accidents and risk factors Joint probability of simultaneous occurrence and risk factors Probability of occurrence All of them can be obtained through the data in the ASN accident database.

[0067] In this embodiment, the single-factor influence conditional probability model can quantify the impact of a single risk factor on the occurrence of UAV operation accidents. For example, taking the UAV loss of control operation accident as an example, this accident may be caused by multiple risk factors, such as brake system failure factors, electromagnetic interference factors, flight control system failure factors, remote communication failure factors, etc. Using the single-factor influence conditional probability model, we focus on the conditional probability of the brake system failure factor on the UAV loss of control operation accident, and the corresponding calculation expression is as follows:

[0068] ,

[0069] in, It indicates the probability of a loss of control operation accident due to brake system failure. represents the joint probability of the loss of control operation accident and brake system failure factors occurring simultaneously, Indicates the probability of occurrence of brake system failure factors.

[0070] S33. Based on the single-factor influence conditional probability model, a multi-factor influence conditional probability model is constructed to calculate the conditional probability of multiple dangerous factors leading to UAV operation accidents;

[0071] The calculation expression of the multi-factor influence conditional probability model in S33 is as follows:

[0072] ,

[0073] , ,

[0074] in, Indicates that an accident occurs when n risk factors exist at the same time The conditional probability of Indicates the first risk factor, Indicates the second risk factor, represents the nth risk factor, represents the conditional probability ratio of multi-factor accidents, represents the first shape parameter, represents the second shape parameter, represents the differentiation with respect to time t, represents the beta function with respect to the first shape parameter and the second shape parameter, Indicates the accidents occurred under the first n risk factors The sum of the conditional probabilities of Indicates the total number of m risk factors that may cause an accident The sum of the conditional probabilities of represents the gamma function, represents the value of the gamma function at the first shape parameter, represents the value of the gamma function at the second shape parameter, represents the value of the gamma function at the sum of the first shape parameter and the second shape parameter, where and Both are greater than 0, m>n, and m and n are positive integers.

[0075] In this embodiment, since each risk factor is independent of each other, each risk factor may cause an accident to different degrees, and the risk factors are not mutually exclusive, when an accident is caused by multiple risk factors, it is necessary to meet the following requirements: (1) when the number of risk factors for an accident increases, the probability of the accident increases; (2) the impact of each risk factor on the accident should be considered within the conditional probability; and (3) the situation in which the accident occurs needs to be compatible with special circumstances. This solution selects the Beta distribution to meet the above requirements. The cumulative distribution function of the Beta distribution is used as a nonlinear function, defined on the interval [0, 1], and the probability density is controlled by the first shape function and the second shape function.

[0076] S34, constructing shape parameter optimization constraints according to a multi-factor influence conditional probability model, and calculating a target first shape parameter and a target second shape parameter;

[0077] The calculation expression of the shape parameter optimization constraint in S34 is as follows:

[0078] ,

[0079] ,

[0080] Among them, min means taking the minimum value, represents the estimated probability of multi-factor accidents, Represents the true probability of a multi-factor accident, where k is a positive integer.

[0081] In this embodiment, the values ​​of the target first shape parameter and the target second shape parameter obtained by shape parameter optimization constraint optimization are 0.92797 and 0.14121 respectively, and the error variance between the estimated probability of multi-factor accident and the true probability of multi-factor accident obtained by fitting the target first shape parameter and the target second shape parameter is 3.56608×10 -7 , which shows that the Beta function fits the observed conditional probability very well. The simulation results also show that the value of the cumulative distribution function of the Beta function almost perfectly matches the data points of the estimated probability of multi-factor accidents.

[0082] S35. Based on the conditional probability of a single hazardous factor leading to a UAV operation accident, the conditional probability of multiple hazardous factors leading to a UAV operation accident, the target first shape parameter and the target second shape parameter, the construction of the Bayesian belief network is completed.

[0083] In this scheme, the Bayesian belief network represents drone operation accidents and risk factors as a series of random variables and the dependencies between them through probability modeling.

[0084] The Bayesian belief network is a directed acyclic graph consisting of a set of links and probability distributions on a number of nodes; each of the nodes corresponds to the probability distribution of the accident occurrence state and the probability distribution of the occurrence state of each risk factor.

[0085] like Figure 2 As shown, as an example, the Bayesian belief network constructed in this embodiment includes: accident occurrence node, other aircraft in the airspace node, system component failure node, pilot error node, insufficient operation skill level node, obstacle node, extreme weather node, bird strike node, insufficient power node, remote communication failure node, flight control system failure node, electromagnetic interference node, brake system failure node, and illegal intrusion node;

[0086] The other aircraft in the airspace nodes, illegal intrusion nodes and insufficient operating technology level nodes all point to the accident occurrence node; the obstacle node points to the system component failure node; the extreme weather nodes point to the system component failure node and the operating system failure node respectively; the bird strike node points to the system component failure node; the low power node points to the operating system failure node and the flight control system failure node respectively; the remote communication failure node, flight control system failure node, electromagnetic interference node and brake system failure node all point to the operating system failure node; the system component failure node and the operating system failure node all point to the accident occurrence node; the pilot error node points to the operating system failure node and the accident occurrence node respectively.

[0087] S4. Add mitigation measures to the corresponding risk factors in the Bayesian belief network, and obtain the risk mitigation effectiveness of all mitigation measures for accidents by estimating and updating the risk factors and the probability of accidents.

[0088] The S4 comprises the following steps:

[0089] S41, corresponding the probability distribution of the hazardous factor occurrence state and the probability distribution of the accident occurrence state corresponding to each node in the Bayesian belief network as the prior probability of the hazardous factor occurrence state and the prior probability of the accident occurrence state;

[0090] like Figure 3As shown, in this embodiment, as an example, the YES state probability of an accident occurring in the Bayesian belief network is 38%, the NO state probability of an accident occurring is 62%, the YES state probability of dangerous factors of other aircraft in the airspace is 32%, and the NO state probability of dangerous factors of other aircraft in the airspace is 68%, among which the YES state is a state that may occur, and the NO state is a state that cannot occur.

[0091] S42, setting mitigation measures to nodes corresponding to risk factors in the Bayesian belief network to obtain a mitigation effectiveness evaluation Bayesian network;

[0092] like Figure 4 As shown, in this embodiment, five corresponding risk mitigation measures are formulated for the dangerous factors that lead to drone accidents: the first measure M1 is to use a four-dimensional meteorological data system to provide weather information for ATM decision-making, helping decision makers to determine whether the operation is safe; the second measure M2 is to use a detection and collision avoidance system, which can effectively replace the environmental perception capability of manned aircraft; the third measure M3 is to set up a geographic fence system to prevent the terminal area from being invaded by drones; the fourth measure M4 is to set up an electromagnetic protection system to prevent electromagnetic interference from causing the drone to lose control; the fifth measure M5 is to set up an automatic emergency braking system to enable the drone to perform emergency parking or emergency hovering when it loses control, so as to avoid accidents.

[0093] S43. Obtain other drone accident cases caused by the risk factors corresponding to the mitigation measures, and based on the prior probability of the risk factor occurrence state, use the mitigation effectiveness evaluation Bayesian belief network to estimate and update the risk factor occurrence state probability distribution and accident occurrence state probability distribution along the link for the nodes corresponding to the risk factor, as the risk factor occurrence state update probability and accident occurrence state update probability;

[0094] like Figure 5 As shown in the figure, as an example, this scheme adds five mitigation measures to the Bayesian belief network and uses this as an example as evidence for causal inference to predict the change in the probability of an accident. For example, after the drone is equipped with a detection and collision avoidance device, it can effectively avoid the approach of other aircraft and obstacles in the airspace, which means that the probability of the occurrence of the YES state of the dangerous factors corresponding to the nodes and obstacle nodes of other aircraft in the airspace will be estimated and updated to 0%, and the probability distribution of the occurrence state of the dangerous factors corresponding to the backward nodes and the probability distribution of the accident state will be affected, and finally the probability of the YES state of the accident in this example will be reduced to 20%. By using the mitigation effectiveness evaluation Bayesian belief network to estimate and update the probability distribution of the occurrence state of the dangerous factors after adding mitigation measures and the probability distribution of the accident state, the degree of influence of the drone accident case on the backward verification of the target node is effectively revealed.

[0095] The calculation expression of the probability of updating the state of the risk factor in S43 is as follows:

[0096] ,

[0097] in, It represents the probability of updating the state of the risk factor corresponding to the jth node after the kth mitigation measure is added, represents the likelihood of adding the kth mitigation measure when the risk factor corresponding to the jth node exists, represents the prior probability of the occurrence state of the risk factor corresponding to the jth node, It represents the probability of other drone accident cases caused by the dangerous factors corresponding to the kth mitigation measure, where k is a positive integer. In this scheme, the probability of accident state update can be estimated and updated backwards according to the probability of dangerous factor state update.

[0098] S44. Based on the updated probability of the accident occurrence state and the prior probability of the accident occurrence state, the risk mitigation effectiveness of all mitigation measures for the accident occurrence is calculated.

[0099] The calculation expression of the risk mitigation effectiveness of all the mitigation measures for accidents is as follows:

[0100] ,

[0101] in, It indicates the risk mitigation effectiveness of all mitigation measures for accidents. represents the probability of updating the accident status. Represents the prior probability of the accident occurrence state.

[0102] Based on the UAV operation accident reports and the risk factors in the reports, this scheme constructed an accident causal influence diagram and a Bayesian belief network. By adding mitigation measures analysis, it achieved a comprehensive qualitative and quantitative analysis of the mitigation effectiveness of UAV operation risks, and was able to accurately and comprehensively evaluate the risk mitigation effectiveness of mitigation measures adopted for UAV operation in airport terminal areas.

[0103] The above description is only a specific implementation mode of the present invention, but the protection scope of the present invention is not limited thereto. Any technician familiar with the technical field can easily think of changes or substitutions within the technical scope disclosed by the present invention, which should be covered by the protection scope of the present invention.

Claims

1. A method for evaluating the effectiveness of UAV operation risk mitigation based on Bayesian network, characterized in that: The steps include: S1. Obtain UAV operation accident reports and identify the risk factors in the UAV operation accident reports; S2. Construct an accident cause-effect impact diagram based on the risk factors in the UAV operation accident report; S3. Construct a Bayesian belief network based on the accident causal impact diagram; S4. Add mitigation measures to the corresponding risk factors in the Bayesian belief network, and obtain the risk mitigation effectiveness of all mitigation measures for accidents by estimating and updating the risk factors and the probability of accidents.

2. The method for evaluating the effectiveness of UAV operation risk mitigation based on Bayesian network according to claim 1 is characterized in that: The risk factors in the UAV operation accident report include operating environment factors, operation factors and facility and equipment factors.

3. The method for evaluating the effectiveness of UAV operation risk mitigation based on Bayesian network according to claim 1 is characterized in that: The S3 comprises the following steps: S31. Obtaining an accident causal impact diagram; S32. Construct a single factor influence conditional probability model to calculate the conditional probability that a single risk factor in the accident causal influence diagram leads to a UAV operation accident; S33. Based on the single-factor influence conditional probability model, a multi-factor influence conditional probability model is constructed to calculate the conditional probability of multiple dangerous factors leading to UAV operation accidents; S34, constructing shape parameter optimization constraints according to a multi-factor influence conditional probability model, and calculating a target first shape parameter and a target second shape parameter; S35. Based on the conditional probability of a single hazardous factor leading to a UAV operation accident, the conditional probability of multiple hazardous factors leading to a UAV operation accident, the target first shape parameter and the target second shape parameter, the construction of the Bayesian belief network is completed.

4. The method for evaluating the effectiveness of UAV operation risk mitigation based on Bayesian network according to claim 3 is characterized in that: The calculation expression of the single factor influence conditional probability model in S32 is as follows: ， in, Indicates the presence of risk factors Accidents The conditional probability of Indicates an accident and risk factors The joint probability of simultaneous occurrence, Indicates risk factors Probability of occurrence.

5. The method for evaluating the effectiveness of UAV operation risk mitigation based on Bayesian network according to claim 3 is characterized in that: The calculation expression of the multi-factor influence conditional probability model in S33 is as follows: ， ， ， in, Indicates that an accident occurs when n risk factors exist at the same time The conditional probability of Indicates the first risk factor, Indicates the second risk factor, represents the nth risk factor, represents the conditional probability ratio of multi-factor accidents, represents the first shape parameter, represents the second shape parameter, represents the differentiation with respect to time t, represents the beta function with respect to the first shape parameter and the second shape parameter, Indicates the accidents occurred under the first n risk factors The sum of the conditional probabilities of Indicates the total number of m risk factors that may cause an accident The sum of the conditional probabilities of represents the gamma function, represents the value of the gamma function at the first shape parameter, represents the value of the gamma function at the second shape parameter, represents the value of the gamma function at the sum of the first shape parameter and the second shape parameter, where and Both are greater than 0, m>n, and m and n are positive integers.

6. The method for evaluating the effectiveness of UAV operation risk mitigation based on Bayesian network according to claim 3 is characterized in that: The calculation expression of the shape parameter optimization constraint in S34 is as follows: ， ， Among them, min means taking the minimum value, represents the estimated probability of multi-factor accidents, Represents the true probability of a multi-factor accident, where k is a positive integer.

7. The method for evaluating the effectiveness of UAV operation risk mitigation based on Bayesian network according to claim 3 is characterized in that: The Bayesian belief network is a directed acyclic graph consisting of a set of links and probability distributions on a number of nodes; each of the nodes corresponds to the probability distribution of the accident occurrence state and the probability distribution of the occurrence state of each risk factor.

8. The method for evaluating the effectiveness of UAV operation risk mitigation based on Bayesian network according to claim 3 is characterized in that: The S4 comprises the following steps: S41, corresponding the probability distribution of the hazardous factor occurrence state and the probability distribution of the accident occurrence state corresponding to each node in the Bayesian belief network as the prior probability of the hazardous factor occurrence state and the prior probability of the accident occurrence state; S42, setting mitigation measures to nodes corresponding to risk factors in the Bayesian belief network to obtain a mitigation effectiveness evaluation Bayesian network; S43. Obtain other drone accident cases caused by the risk factors corresponding to the mitigation measures, and based on the prior probability of the risk factor occurrence state, use the mitigation effectiveness evaluation Bayesian belief network to estimate and update the risk factor occurrence state probability distribution and accident occurrence state probability distribution along the link for the nodes corresponding to the risk factor, as the risk factor occurrence state update probability and accident occurrence state update probability; S44. Based on the updated probability of the accident occurrence state and the prior probability of the accident occurrence state, the risk mitigation effectiveness of all mitigation measures for the accident occurrence is calculated.

9. The method for evaluating the effectiveness of UAV operation risk mitigation based on Bayesian network according to claim 8 is characterized in that: The calculation expression of the probability of updating the state of the risk factor in S43 is as follows: ， in, It represents the probability of updating the state of the risk factor corresponding to the jth node after the kth mitigation measure is added, represents the likelihood of adding the kth mitigation measure when the risk factor corresponding to the jth node exists, represents the prior probability of the occurrence state of the risk factor corresponding to the jth node, It represents the probability of other drone accident cases caused by the hazardous factors corresponding to the kth mitigation measure, where k is a positive integer.

10. The method for evaluating the effectiveness of UAV operation risk mitigation based on Bayesian network according to claim 8, characterized in that: The calculation expression of the risk mitigation effectiveness of all the mitigation measures for accidents is as follows: ， in, It indicates the risk mitigation effectiveness of all mitigation measures for accidents. represents the probability of updating the accident status. Represents the prior probability of the accident occurrence state.

Images