Method, system and device for verifying reliability of unmanned aerial vehicle and medium

Abstract

The invention provides an unmanned aerial vehicle reliability verification method, system and device and a medium. The method comprises the steps of obtaining environmental parameter data of a to-be-verified unmanned aerial vehicle under a non-standard working condition; calculating the battery efficiency of the unmanned aerial vehicle according to the environmental parameter data; according to the environmental parameter data and the battery efficiency, performing flight state estimation on the unmanned aerial vehicle by using a Kalman filtering algorithm to obtain predicted flight state data; performing fault detection on the unmanned aerial vehicle by using a deep reinforcement learning algorithm according to the predicted flight state data to obtain a fault risk index of the unmanned aerial vehicle, and performing reliability verification on the unmanned aerial vehicle according to the fault risk index; according to the method, battery efficiency compensation and air density correction are introduced into the Kalman filtering algorithm, so that the flight state of the unmanned aerial vehicle under the non-standard working condition can be predicted more accurately; the state features, the environmental parameters and the battery efficiency are fused through the deep reinforcement learning algorithm, the fault risk index can be accurately calculated, and the accuracy of reliability verification of the unmanned aerial vehicle is improved.

Description

Technical Field

[0001] The present invention relates to the technical field of unmanned aerial vehicle reliability verification, and in particular to a method, system, device and medium for verifying the reliability of an unmanned aerial vehicle. Background Art

[0002] At present, with the rapid development of UAV (Unmanned Aerial Vehicle) technology, UAVs are widely used in agriculture, logistics, surveying and mapping, military and other fields. The reliability verification of UAVs is considered to be a key factor in ensuring their safe and efficient operation. Reliability verification can not only improve the stability and safety of UAVs in performing complex tasks and reduce operational risks, but also enhance users' trust in UAV technology, thereby promoting the further development of the UAV market.

[0003] However, the reliability verification of UAVs is mainly based on standard operating conditions (usually refers to controllable and relatively stable environmental conditions in the laboratory, including parameters such as temperature, humidity and air pressure within the normal range). Reliability verification based on standard operating conditions also has significant limitations. First, standard operating conditions cannot fully reproduce the various extreme and changeable environments that may be encountered in actual operations. For example, in plateau areas, low temperatures, thin air and rapidly changing meteorological conditions will significantly affect the power output and flight stability of UAVs, resulting in performance degradation, which may lead to safety hazards; secondly, tests under standard operating conditions often do not take into account the problem of battery efficiency degradation in extreme environments, which may lead to an incorrect assessment of battery life and thus affect the operating period of the entire UAV; finally, the research results of standard operating conditions may not be applicable to actual application scenarios, making the evaluation of UAV reliability in specific environments inaccurate, thereby affecting the success rate of the mission and flight safety.

[0004] Therefore, the existing UAV reliability verification method based on standard operating conditions cannot fully consider the impact of extreme conditions under non-standard operating conditions such as low temperature in plateaus and thin air on UAV performance and battery life, resulting in inaccurate evaluation results, which may lead to safety hazards and mission failures. Summary of the invention

[0005] In order to solve the problem that the existing UAV reliability verification method based on standard working conditions cannot fully consider the impact of extreme conditions under non-standard working conditions on the performance and battery life of the UAV, resulting in inaccurate reliability verification results, the present invention proposes a UAV reliability verification method, comprising: Obtain environmental parameter data of the drone to be verified under non-standard working conditions; Calculating the battery efficiency of the UAV according to the environmental parameter data; According to the environmental parameter data of the UAV and the battery efficiency, a Kalman filter algorithm is used to estimate the flight state of the UAV to obtain predicted flight state data of the UAV; According to the predicted flight status data of the UAV, a deep reinforcement learning algorithm is used to perform fault detection on the UAV to obtain a fault risk index of the UAV, and reliability verification of the UAV is performed according to the fault risk index; Wherein, battery efficiency compensation and air density correction are introduced into the Kalman filter algorithm.

[0006] Optionally, estimating the flight state of the drone using a Kalman filter algorithm according to the environmental parameter data of the drone and the battery efficiency to obtain predicted flight state data of the drone includes: Initializing the flight state of the UAV according to the environmental parameter data of the UAV and the battery efficiency to obtain an initial flight state of the UAV; According to the environmental parameter data of the UAV, the battery efficiency and the initial flight state, the state of the UAV is predicted by a state prediction equation to obtain predicted flight state data of the UAV.

[0007] Optionally, the expression corresponding to the state prediction equation is as follows: ; In the formula, ; ; in, Indicates drone Predicted flight status data at all times; Indicates drone The state transfer matrix at time; Indicates drone Predicted flight status data at all times; represents the battery efficiency compensation factor; Indicates drone The control input matrix at each moment; Indicates drone The control vector at the moment; Indicates the battery efficiency of the drone; Indicates drone The process noise at each moment; represents the air density correction factor; Indicates drone Wind speed influence coefficient matrix at the moment; represents the wind speed vector; Indicates the air density; Indicates standard air density.

[0008] Optionally, the performing fault detection on the drone using a deep reinforcement learning algorithm according to the predicted flight state data of the drone to obtain a fault risk index of the drone includes: Perform feature extraction based on the predicted flight state data of the UAV to obtain state features of the UAV; According to the state characteristics, calculating the corresponding characteristic function value; Perform risk value mapping according to the characteristic function value to obtain a failure risk index of the drone; The state characteristics include one or more of the following: position deviation, speed fluctuation and attitude angle deviation.

[0009] Optionally, the expression corresponding to the deep reinforcement learning algorithm is as follows: ; in, represents a failure risk index of the UAV; represents the risk value mapping function; Indicates The feature weight of each state feature; ; Indicates the total number of state features; Indicates the battery efficiency of the drone; Indicates The feature vector of the state features The state characteristic function value of ; represents the temperature influence coefficient; Indicates standard temperature; Indicates the difference between the current temperature and the standard temperature; Indicates the air density influence coefficient; Indicates the current air density; Indicates standard air density; Represents the battery efficiency influence coefficient.

[0010] Optionally, the calculation formula corresponding to the battery efficiency of the drone is as follows: ; in, Indicates the battery efficiency of the drone; Indicates the benchmark value of drone battery efficiency under standard working conditions; represents the exponential function; Indicates standard temperature; Indicates the difference between the current temperature and the standard temperature; Indicates the air density influence coefficient; Indicates the current air density; Indicates standard air density; represents the load factor; Indicates the current battery current data of the drone; Indicates the maximum allowable current of the battery; represents the humidity influence coefficient; Represents the current humidity data of the drone; Indicates the maximum allowable humidity of the drone.

[0011] Optionally, the environmental parameter data includes: environmental data and drone status data; The environmental data includes one or more of the following: environmental temperature, air density, humidity data, wind speed data and location data; The drone status data includes one or more of the following: battery voltage data, battery current data, and sensor data.

[0012] Based on the same inventive concept, the present invention also provides a UAV reliability verification system, comprising: A data acquisition module, used to obtain environmental parameter data of the UAV to be verified under non-standard working conditions; An efficiency calculation module, used to calculate the battery efficiency of the drone according to the environmental parameter data; A state estimation module, used to estimate the flight state of the UAV using a Kalman filter algorithm according to the environmental parameter data of the UAV and the battery efficiency, so as to obtain predicted flight state data of the UAV; A fault detection module, configured to perform fault detection on the drone using a deep reinforcement learning algorithm according to the predicted flight state data of the drone, obtain a fault risk index of the drone, and perform reliability verification on the drone according to the fault risk index; Wherein, battery efficiency compensation and air density correction are introduced into the Kalman filter algorithm.

[0013] Optionally, the state estimation module includes: A state initialization submodule, used to initialize the flight state of the UAV according to the environmental parameter data of the UAV and the battery efficiency, so as to obtain the initial flight state of the UAV; The state prediction submodule is used to predict the state of the UAV through a state prediction equation according to the environmental parameter data of the UAV, the battery efficiency and the initial flight state, so as to obtain the predicted flight state data of the UAV.

[0014] Optionally, the expression corresponding to the state prediction equation is as follows: ; In the formula, ; ; in, Indicates drone Predicted flight status data at all times; Indicates drone The state transfer matrix at time; Indicates drone Predicted flight status data at all times; represents the battery efficiency compensation factor; Indicates drone The control input matrix at each moment; Indicates drone The control vector at the moment; Indicates the battery efficiency of the drone; Indicates drone The process noise at each moment; represents the air density correction factor; Indicates drone Wind speed influence coefficient matrix at the moment; represents the wind speed vector; Indicates the air density; Indicates standard air density.

[0015] Optionally, the fault detection module includes: A feature extraction submodule, used to extract features according to the predicted flight state data of the UAV to obtain the state features of the UAV; A function value calculation submodule, used to calculate the corresponding characteristic function value according to the state characteristics; The risk mapping submodule is used to perform risk value mapping according to the characteristic function value to obtain the failure risk index of the drone.

[0016] The state characteristics include one or more of the following: position deviation, speed fluctuation and attitude angle deviation.

[0017] Optionally, the expression corresponding to the deep reinforcement learning algorithm is as follows: ; in, represents a failure risk index of the UAV; represents the risk value mapping function; Indicates The feature weight of each state feature; ; Indicates the total number of state features; Indicates the battery efficiency of the drone; Indicates The feature vector of the state features The state characteristic function value of ; represents the temperature influence coefficient; Indicates standard temperature; Indicates the difference between the current temperature and the standard temperature; Indicates the air density influence coefficient; Indicates the current air density; Indicates standard air density; Represents the battery efficiency influence coefficient.

[0018] Optionally, the calculation formula corresponding to the battery efficiency of the drone is as follows: ; in, Indicates the battery efficiency of the drone; Indicates the benchmark value of drone battery efficiency under standard working conditions; represents the exponential function; Indicates standard temperature; Indicates the difference between the current temperature and the standard temperature; Indicates the air density influence coefficient; Indicates the current air density; Indicates standard air density; represents the load factor; Indicates the current battery current data of the drone; Indicates the maximum allowable current of the battery; represents the humidity influence coefficient; Represents the current humidity data of the drone; Indicates the maximum allowable humidity of the drone.

[0019] Optionally, the environmental parameter data includes: environmental data and drone status data; The environmental data includes one or more of the following: environmental temperature, air density, humidity data, wind speed data and location data; The drone status data includes one or more of the following: battery voltage data, battery current data, and sensor data.

[0020] In another aspect, the present invention further provides an electronic device, comprising: at least one processor and a memory; the memory and the processor are connected via a bus; The memory is used to store one or more programs; When the one or more programs are executed by the at least one processor, a method for verifying the reliability of a drone as described above is implemented.

[0021] On the other hand, the present invention also provides a computer-readable storage medium having an execution program stored thereon, and when the execution program is executed, a method for verifying the reliability of a drone as described above is implemented.

[0022] Compared with the prior art, the present invention has the following beneficial effects: The present invention provides a method, system, device and medium for verifying the reliability of an unmanned aerial vehicle, comprising: obtaining environmental parameter data of an unmanned aerial vehicle to be verified under non-standard working conditions; calculating the battery efficiency of the unmanned aerial vehicle according to the environmental parameter data; estimating the flight state of the unmanned aerial vehicle using a Kalman filter algorithm according to the environmental parameter data and the battery efficiency of the unmanned aerial vehicle, and obtaining predicted flight state data of the unmanned aerial vehicle; performing fault detection on the unmanned aerial vehicle using a deep reinforcement learning algorithm according to the predicted flight state data of the unmanned aerial vehicle, obtaining a fault risk index of the unmanned aerial vehicle, and verifying the reliability of the unmanned aerial vehicle according to the fault risk index; wherein, battery efficiency compensation and air density correction are introduced into the Kalman filter algorithm; the present application can more accurately predict the flight state of the unmanned aerial vehicle under non-standard working conditions by introducing battery efficiency compensation and air density correction into the Kalman filter algorithm; the deep reinforcement learning algorithm can integrate state characteristics, environmental parameters and battery efficiency, which is conducive to accurately calculating the fault risk index, improving the accuracy of the reliability verification of the unmanned aerial vehicle, and thus improving the safety and reliability of the unmanned aerial vehicle when performing tasks. BRIEF DESCRIPTION OF THE DRAWINGS

[0023] Figure 1 A schematic diagram of a flow chart of a method for verifying the reliability of a drone provided by the present invention; Figure 2 A schematic diagram of a framework flow of flight state estimation in a method for verifying the reliability of a UAV provided by the present invention; Figure 3 A schematic diagram of the structure of a UAV reliability verification system provided by the present invention; Figure 4 The present invention provides a schematic structural diagram of an electronic device. DETAILED DESCRIPTION

[0024] The present invention proposes a method, system, device and medium for verifying the reliability of a drone. The specific implementation methods of the present invention are further described in detail below with reference to the accompanying drawings.

[0025] Embodiment 1: The present invention provides a method for verifying the reliability of a drone, the flow chart of which is as follows: Figure 1 As shown, including: Step 1: Obtain environmental parameter data of the UAV to be verified under non-standard working conditions; Step 2: Calculate the battery efficiency of the drone based on the environmental parameter data; Step 3: According to the environmental parameter data and battery efficiency of the UAV, the flight state of the UAV is estimated using the Kalman filter algorithm to obtain the predicted flight state data of the UAV; Step 4: Based on the predicted flight status data of the drone, use the deep reinforcement learning algorithm to detect the drone's faults, obtain the drone's fault risk index, and verify the reliability of the drone based on the fault risk index; Among them, battery efficiency compensation and air density correction are introduced into the Kalman filter algorithm.

[0026] Generally, the reliability verification of UAVs is mainly based on standard operating conditions (such as ideal environments such as normal temperature, normal pressure, and no wind). This verification method evaluates the performance indicators of UAVs (such as flight stability, battery life, etc.) through laboratory simulation or limited field testing. However, in non-standard operating conditions such as low temperature on plateaus (usually refers to conditions where the altitude is greater than 3,000 meters and the ambient temperature is less than 0°C), thin air (usually refers to conditions where the air density is lower than the standard sea level air density), and strong winds (usually refers to wind speeds that are significantly higher than the wind speed threshold under normal flight conditions of UAVs), the performance and battery life of UAVs may drop significantly, and existing methods cannot fully consider the impact of these extreme conditions, resulting in a large deviation between the reliability evaluation results and the actual situation. Therefore, the present invention conducts research by considering environmental parameter data under non-standard conditions (which can be collected by electronic devices embedded in the UAV); for example, the environmental parameter data in the above step 1 may include: environmental data and UAV status data; The above-mentioned environmental data may include one or more of the following: environmental temperature, air density, humidity data, wind speed data and location data; The above-mentioned drone status data may include one or more of the following: battery voltage data, battery current data and sensor data.

[0027] On the basis of obtaining environmental parameter data and drone status data, in order to accurately evaluate the battery performance changes of drones under non-standard working conditions, it is necessary to further analyze the factors affecting battery efficiency. These environmental data and status data together constitute the comprehensive data basis for drone operation, making the calculation of battery efficiency an important part of verifying the reliability of drones in the next step. For example, the calculation formula corresponding to the battery efficiency of the drone in step 2 above can be as follows: ; in, Indicates the battery efficiency of the drone; Indicates the benchmark value of drone battery efficiency under standard working conditions; represents the exponential function; Indicates standard temperature; Indicates the difference between the current temperature and the standard temperature; Indicates the air density influence coefficient; Indicates the current air density; Indicates standard air density; represents the load factor; Indicates the current battery current data of the drone; Indicates the maximum allowable current of the battery; represents the humidity influence coefficient; Represents the current humidity data of the drone; Represents the maximum allowable humidity of the drone; in this example, by introducing multimodal data fusion, the impact of multiple factors such as temperature, air density, battery load and humidity on the drone battery efficiency is comprehensively considered, which can significantly improve the comprehensiveness and accuracy of the battery efficiency calculation. Traditional battery efficiency calculations usually only consider a single factor (such as temperature or load) and cannot fully reflect the changes in battery performance under complex working conditions. This example dynamically adjusts the calculation results of battery efficiency by introducing exponential functions and linear correction factors. It can dynamically adapt to changes in different environmental conditions and make it closer to actual working conditions, especially in complex and extreme operating environments. For example, through the temperature change rate Corrected the effect of low temperature on battery chemical activity by the rate of change of air density Correct the effect of thin plateau air on battery heat dissipation through load factor Corrects the effect of high load on battery internal resistance and humidity change rate By correcting the impact of high humidity on battery performance, this multimodal data fusion method can not only improve the comprehensiveness of the model, but also significantly improve the accuracy of battery efficiency calculations, providing more reliable data support for UAV endurance assessment and flight status prediction, thereby enhancing the adaptability and mission execution capabilities of UAVs under non-standard working conditions.

[0028] The battery efficiency calculated by the above formula can provide relatively accurate data support for the performance evaluation of the UAV under different environmental conditions. Using these refined and quantified parameters, a detailed prediction of the flight dynamics of the UAV can be achieved. Based on this, the flight state of the UAV can be estimated by the Kalman filter method to obtain more accurate predicted flight state data. Specifically: In one implementation, Figure 2As shown, in the above step 3, the process of estimating the flight state of the UAV by using the Kalman filter algorithm according to the environmental parameter data and battery efficiency of the UAV to obtain the predicted flight state data of the UAV may include: Initialize the flight state of the UAV according to the environmental parameter data and battery efficiency of the UAV to obtain the initial flight state of the UAV; According to the environmental parameter data, battery efficiency and initial flight state of the UAV, the state of the UAV is predicted through the state prediction equation to obtain the predicted flight state data of the UAV; In this implementation, by combining the environmental parameter data of the UAV (such as fixed-wing UAV, multi-rotor UAV, hybrid UAV, etc.) with the battery efficiency for state initialization, it is beneficial to improve the uncertainty processing capability of the initial state, making the initial flight state more accurate; at this time, by utilizing the recursive characteristics of the Kalman filter algorithm, by integrating the dynamic environmental factors with the battery efficiency and applying them to the state initialization, the algorithm obtains a more accurate initial state at the beginning, and allows the flight path and energy consumption plan of the UAV to be adjusted in real time under the continuously updated environmental parameter conditions, which is beneficial to enhance the responsiveness to complex environments, thereby improving flight safety and efficiency, and also showing unique advantages in reducing energy consumption, extending the endurance time and improving the task completion rate. In particular, its recursive estimation characteristics enable the prediction model to respond to changing environmental conditions in real time during the flight, thereby showing stronger adaptability. Through this optimization method, not only the accuracy of flight state prediction can be directly improved, but also the overall operational efficiency of the UAV can be indirectly enhanced, significantly extending the flight time and reducing energy consumption.

[0029] For example, the expression corresponding to the above state prediction equation can be as follows: ; In the formula, ; ; in, Indicates drone Predicted flight status data at all times; Indicates drone The state transfer matrix at time; Indicates drone Predicted flight status data at all times; represents the battery efficiency compensation factor; Indicates drone The control input matrix at each moment; Indicates drone The control vector at the moment; Indicates the battery efficiency of the drone; Indicates drone The process noise at each moment; represents the air density correction factor; Indicates drone Wind speed influence coefficient matrix at the moment; represents the wind speed vector; Indicates the air density; represents the standard air density; in this example, the battery efficiency of the drone is taken as the key factor. By introducing the battery efficiency compensation factor and the battery efficiency, the state prediction can be dynamically adjusted in combination with the battery performance. This method is in contrast to the traditional prediction model that ignores the impact of battery efficiency on the flight state, which is conducive to greatly improving the accuracy of energy consumption modeling and provides basic support for optimizing endurance and energy management; secondly, the air density correction factor is introduced into the model, and the wind speed influence coefficient matrix C k and wind speed vector This method incorporates the dynamic change analysis of environmental parameters. Compared with the simple assumption that the environment is constant, it comprehensively considers the complex impact of the changing environment on the flight state, especially in scenarios where wind speed and air density change greatly. It has significant advantages, and by adjusting the ratio of air density to standard air density, it can effectively correct the influence of air resistance, thereby more accurately predicting the flight behavior of the UAV. In addition, in terms of process noise considerations, it can effectively reduce noise interference, thereby further ensuring the reliability and stability of flight predictions.

[0030] In the above steps, the flight status of the drone is predicted by the Kalman filter algorithm, and relatively accurate predicted flight status data is obtained. However, relying solely on the prediction of flight status is not enough to fully evaluate the health of the drone. In order to further ensure the safe flight of the drone in a complex environment, it is necessary to conduct in-depth analysis of the predicted flight status data to identify potential failure risks. Specifically: In one implementation, the process of performing fault detection on the drone using a deep reinforcement learning algorithm based on the predicted flight status data of the drone in step 4 to obtain the fault risk index of the drone may include: Feature extraction is performed based on the predicted flight status data of the UAV to obtain the state characteristics of the UAV; According to the state characteristics, calculate the corresponding characteristic function value; According to the characteristic function value, the risk value is mapped to obtain the failure risk index of the UAV; The state characteristics include one or more of the following: position deviation, speed fluctuation and attitude angle deviation; For example, suppose the state characteristics of the drone are ,in, Indicates position deviation; Indicates speed fluctuations; Represents attitude angle deviation; defines a comprehensive feature vector : ; According to the feature vector , construct the characteristic function , and use the neural network model to calculate the characteristic function value: ; in, Represents a neural network model; are model parameters; According to the characteristic function value Perform risk value mapping and define a risk mapping function , and convert it into a failure risk index ; In the risk mapping function, you can choose to use nonlinear mapping to increase the sensitivity to changes in state characteristics, such as using a sigmoid function for normalization: ; in, is the slope of the mapping; is the bias term of the mapping, which is used to adjust the sensitivity and range of the risk index; In this implementation, by accurately extracting the state features of the drone, it is possible to ensure that the fault detection algorithm can be analyzed based on reliable data input. For example, in the example, the state features may include position deviation , speed fluctuation and attitude angle deviation These important state indicators are integrated through the feature vector Being effectively integrated can greatly improve the accuracy of fault detection, and by using the neural network model to build the characteristic function This process means that the model can automatically optimize and identify complex failure modes by learning historical data. Compared with traditional methods, this implementation method can handle higher-dimensional and nonlinear risks, making the determination of failure risks more accurate. Furthermore, the risk mapping function The adoption of, especially the selection of nonlinear mapping, can significantly improve the sensitivity to changes in state characteristics, making the fault risk index It can more sensitively reflect environmental changes and abnormal states; therefore, this implementation method can improve the accuracy and sensitivity of fault detection through deep reinforcement learning, and improve the overall adaptability of the drone by comprehensively considering state characteristics.

[0031] For example, the expression corresponding to the above deep reinforcement learning algorithm can be as follows: ; in, represents the failure risk index of the drone; represents the risk value mapping function; Indicates The feature weight of each state feature; ; Indicates the total number of state features; Indicates the battery efficiency of the drone; Indicates The feature vector of the state features The state characteristic function value of ; represents the temperature influence coefficient; Indicates standard temperature; Indicates the difference between the current temperature and the standard temperature; Indicates the air density influence coefficient; Indicates the current air density; Indicates standard air density; Represents the battery efficiency impact coefficient; in this example, by introducing a weighted summation method, the influence of multiple state characteristics of the drone (such as position deviation, speed fluctuation, and attitude angle deviation) is integrated, and the weight parameter Dynamically adjust the influence of different features. This design enables the model to flexibly adjust the weight of features based on specific tasks or flight environments, significantly enhancing the adaptability and generalization ability of the model. In addition, a variety of key environmental parameters (such as temperature difference) are added to the expression. , air density , standard density ) and the drone’s system performance parameters (such as battery efficiency ), through the temperature influence coefficient γ, air density influence coefficient α and battery efficiency influence coefficient These variables are processed in a layered and weighted manner. This design significantly improves the model's ability to handle complex environmental variables and can dynamically capture the impact of temperature changes, air density fluctuations and battery performance degradation on potential UAV failures, especially in extreme environments or long-term flight scenarios. This method demonstrates stronger robustness and early warning capabilities. Finally, through the introduction of the risk value mapping function, the calculation results in the multivariate complex model can be mapped into an intuitive and quantitative risk index R, which not only facilitates risk transfer and understanding in engineering practice, but also provides a clear guiding logic for the automatic decision-making and fault handling of UAVs.

[0032] In summary, the present invention aims at the problem that the existing UAV reliability verification method based on standard working conditions cannot fully consider the impact of extreme conditions such as low temperature and thin air in non-standard working conditions on the performance and battery life of the UAV, resulting in inaccurate reliability verification results. A UAV reliability verification method is proposed. By acquiring environmental parameter data under non-standard working conditions and calculating the battery efficiency of the UAV, key environmental factors and the UAV's own performance are introduced into the reliability analysis framework, breaking through the defect of the traditional method being limited to the standard working condition assumption; the flight state of the UAV is estimated by using the Kalman filter algorithm, and at the same time, combined with the battery efficiency compensation factor and the air density correction, the predicted flight state data can more truly reflect the impact of non-standard working conditions on the UAV; further, the predicted flight state data is fault detected by a deep reinforcement learning algorithm, the risks that may arise in the UAV under extreme environments are dynamically captured, and the fault risk index is quantified and output; finally, through the analysis and processing of the risk index, a comprehensive, accurate and dynamic verification of the performance and reliability of the UAV is achieved. The method of the present invention effectively makes up for the shortcomings of existing technologies in complex environments, significantly improves the reliability verification accuracy of UAVs under various non-standard working conditions, provides important technical support for the safe operation of UAVs in extreme mission scenarios, and also provides a scientific basis for the design optimization and operation management of UAVs.

[0033] Embodiment 2: In practical applications, drones often need to perform tasks in complex and dynamic environments. Therefore, by combining fault detection with multi-machine collaborative tasks, it is possible to ensure that the entire mission can still proceed smoothly when a single aircraft fails, thereby reducing the risk of mission interruption. Specifically: First, a shared status information platform can be established to enable each drone to obtain the flight status and fault risk index of other drones. This information sharing can enhance the fault warning capability of the entire group and reduce the impact of single-point failures on the mission. In the implementation of multi-machine collaborative tasks, intelligent scheduling can be carried out in combination with the fault detection results of drones. For example, for drones with high risk levels, other low-risk drones can be mobilized to undertake their tasks to ensure the continuous execution of the overall task; and task planning can be dynamically adjusted based on real-time fault detection data. If a drone fails, it can be quickly identified and the task can be automatically re-planned to other drones to minimize task delays or risks; By using the data collected in multi-drone collaboration, the deep learning model is used to identify the common fault modes faced by multiple drones, thereby optimizing the fault detection algorithm. At the same time, the structure and mission characteristics of different drones are considered to make the model more adaptable; In multi-UAV missions, it is also possible to consider designing a real-time feedback mechanism to ensure that UAVs can recognize and respond to each other’s faults. For example, if a UAV detects a fault, it can immediately pass the information to other UAVs to initiate emergency measures. Therefore, associating the implementation method of this application with multi-machine collaborative tasks can not only improve the overall performance and safety of the drone system, but also be an effective response to the development direction of the drone industry.

[0034] Embodiment 3: The present invention based on the same inventive concept also provides a UAV reliability verification system, the structural composition diagram is as follows: Figure 3 As shown, including: A data acquisition module, used to obtain environmental parameter data of the UAV to be verified under non-standard working conditions; The efficiency calculation module is used to calculate the battery efficiency of the drone based on the environmental parameter data; The state estimation module is used to estimate the flight state of the UAV based on the environmental parameter data and battery efficiency of the UAV using the Kalman filter algorithm to obtain the predicted flight state data of the UAV; The fault detection module is used to detect the faults of the UAV based on the predicted flight status data of the UAV using a deep reinforcement learning algorithm, obtain the fault risk index of the UAV, and verify the reliability of the UAV based on the fault risk index; Among them, battery efficiency compensation and air density correction are introduced into the Kalman filter algorithm.

[0035] For example, the above-mentioned environmental parameter data may include: environmental data and drone status data; The above-mentioned environmental data may include one or more of the following: environmental temperature, air density, humidity data, wind speed data and location data; The above-mentioned drone status data may include one or more of the following: battery voltage data, battery current data and sensor data.

[0036] For example, the calculation formula corresponding to the battery efficiency of the above-mentioned drone can be as follows: ; in, Indicates the battery efficiency of the drone; Indicates the benchmark value of drone battery efficiency under standard working conditions; represents the exponential function; Indicates standard temperature; Indicates the difference between the current temperature and the standard temperature; Indicates the air density influence coefficient; Indicates the current air density; Indicates standard air density; represents the load factor; Indicates the current battery current data of the drone; Indicates the maximum allowable current of the battery; represents the humidity influence coefficient; Represents the current humidity data of the drone; Indicates the maximum allowable humidity of the drone.

[0037] In one implementation, the state estimation module may include: The state initialization submodule is used to initialize the flight state of the UAV according to the environmental parameter data and battery efficiency of the UAV to obtain the initial flight state of the UAV; The state prediction submodule is used to predict the state of the UAV through the state prediction equation according to the environmental parameter data, battery efficiency and initial flight state of the UAV, and obtain the predicted flight state data of the UAV.

[0038] For example, the expression corresponding to the above state prediction equation can be as follows: ; In the formula, ; ; in, Indicates drone Predicted flight status data at all times; Indicates drone The state transfer matrix at time; Indicates drone Predicted flight status data at all times; represents the battery efficiency compensation factor; Indicates drone The control input matrix at each moment; Indicates drone The control vector at the moment; Indicates the battery efficiency of the drone; Indicates drone The process noise at each moment; represents the air density correction factor; Indicates drone Wind speed influence coefficient matrix at the moment; represents the wind speed vector; Indicates the air density; Indicates standard air density.

[0039] In one implementation, the above-mentioned fault detection module may include: The feature extraction submodule is used to extract features based on the predicted flight status data of the UAV to obtain the state features of the UAV; The function value calculation submodule is used to calculate the corresponding characteristic function value according to the state characteristics; The risk mapping submodule is used to map the risk value according to the characteristic function value to obtain the failure risk index of the drone.

[0040] The state characteristics include one or more of the following: position deviation, speed fluctuation and attitude angle deviation.

[0041] For example, the expression corresponding to the above deep reinforcement learning algorithm can be as follows: ; in, represents the failure risk index of the drone; represents the risk value mapping function; Indicates The feature weight of each state feature; ; Indicates the total number of state features; Indicates the battery efficiency of the drone; Indicates The feature vector of the state features The state characteristic function value of ; represents the temperature influence coefficient; Indicates standard temperature; Indicates the difference between the current temperature and the standard temperature; Indicates the air density influence coefficient; Indicates the current air density; Indicates standard air density; Represents the battery efficiency influence coefficient.

[0042] Embodiment 4: like Figure 4 As shown, the present invention also provides an electronic device, which may be a computer device, a single-chip device, an intelligent mobile device, etc. The electronic device in this embodiment may include a processor, a memory, a transceiver component, etc. The memory, the processor, and the transceiver component are connected via a bus; the memory may be used to store an execution program, and an exemplary execution program may include instructions; the processor is used to execute the instructions stored in the memory. The memory may also be used to store data, which may be called and / or modified when the instructions are executed.

[0043] The processor may be a central processing unit (CPU), or other general-purpose processors, digital signal processors (DSP), application-specific integrated circuits (ASIC), field-programmable gate arrays (FPGA) or other programmable logic devices, discrete gates or transistor logic devices, discrete hardware components, etc. It is the computing core and control core of the terminal, which is suitable for implementing one or more instructions, specifically suitable for loading and executing one or more instructions in a storage medium to implement the corresponding method flow or corresponding functions, so as to implement the steps of a method for verifying the reliability of a drone in the above-mentioned embodiment.

[0044] Embodiment 5: Based on the same inventive concept, the present invention also provides a readable storage medium, specifically an electronic device readable storage medium (Memory), which is a memory device in an electronic device for storing programs and data. It can be understood that the storage medium here can include both built-in storage media in electronic devices and extended storage media supported by electronic devices. The storage medium provides a storage space, which stores the operating system of the terminal. In addition, one or more instructions suitable for being loaded and executed by a processor are also stored in the storage space, and these instructions can be one or more execution programs (including program codes). It should be noted that the storage medium here can be a high-speed RAM memory or a non-volatile memory, such as at least one disk memory. The processor loads and executes one or more instructions stored in the storage medium, which can implement the steps of a method for verifying the reliability of a drone in the above embodiment.

[0045] It will be appreciated by those skilled in the art that embodiments of the present invention may be provided as methods, systems, or computer program products. Therefore, the present invention may take the form of a complete hardware embodiment, a complete software embodiment, or an embodiment combining software and hardware. Furthermore, the present invention may take the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0046] The present invention is described with reference to flowcharts and / or block diagrams of methods, devices (systems), and computer program products according to embodiments of the present invention. It should be understood that each process and / or block in the flowchart and / or block diagram, as well as the combination of processes and / or blocks 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 processes in the flowchart and / or block diagram. 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.

[0047] These computer program instructions may also be stored in a computer-readable memory capable of directing a computer or other programmable data processing device to operate 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 A function specified in one or more boxes.

[0048] These computer program instructions can also be loaded onto a computer or other programmable data processing device so that a series of operating steps are executed on the computer or other programmable device to produce a computer-implemented process, thereby providing instructions for implementing the process. Figure 1 A process or multiple processes and / or boxes Figure 1 The steps for the functions specified in one or more boxes.

[0049] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention rather than to limit its protection scope. Although the present invention has been described in detail with reference to the above embodiments, ordinary technicians in the field should understand that after reading the present invention, those skilled in the art can still make various changes, modifications or equivalent substitutions to the specific implementation methods of the application, but these changes, modifications or equivalent substitutions are all within the protection scope of the claims to be approved.

Claims

1. A method for verifying the reliability of an unmanned aerial vehicle, characterized in that: include: Obtain environmental parameter data of the drone to be verified under non-standard working conditions; Calculating the battery efficiency of the UAV according to the environmental parameter data; According to the environmental parameter data of the UAV and the battery efficiency, a Kalman filter algorithm is used to estimate the flight state of the UAV to obtain predicted flight state data of the UAV; According to the predicted flight status data of the UAV, a deep reinforcement learning algorithm is used to perform fault detection on the UAV to obtain a fault risk index of the UAV, and reliability verification of the UAV is performed according to the fault risk index; Wherein, battery efficiency compensation and air density correction are introduced into the Kalman filter algorithm.

2. The method according to claim 1, characterized in that The method of estimating the flight state of the UAV by using a Kalman filter algorithm according to the environmental parameter data of the UAV and the battery efficiency to obtain the predicted flight state data of the UAV includes: Initializing the flight state of the UAV according to the environmental parameter data of the UAV and the battery efficiency to obtain an initial flight state of the UAV; According to the environmental parameter data of the UAV, the battery efficiency and the initial flight state, the state of the UAV is predicted by a state prediction equation to obtain predicted flight state data of the UAV.

3. The method according to claim 2, characterized in that The expression corresponding to the state prediction equation is as follows: ； In the formula, ； ； in, Indicates drone Predicted flight status data at all times; Indicates drone The state transfer matrix at time; Indicates drone Predicted flight status data at all times; represents the battery efficiency compensation factor; Indicates drone The control input matrix at each moment; Indicates drone The control vector at the moment; Indicates the battery efficiency of the drone; Indicates drone The process noise at each moment; represents the air density correction factor; Indicates drone Wind speed influence coefficient matrix at the moment; represents the wind speed vector; Indicates the air density; Indicates standard air density.

4. The method according to claim 1, characterized in that The method of performing fault detection on the drone using a deep reinforcement learning algorithm according to the predicted flight state data of the drone to obtain a fault risk index of the drone includes: Perform feature extraction based on the predicted flight state data of the UAV to obtain state features of the UAV; According to the state characteristics, calculating the corresponding characteristic function value; Perform risk value mapping according to the characteristic function value to obtain a failure risk index of the drone; The state characteristics include one or more of the following: position deviation, speed fluctuation and attitude angle deviation.

5. The method according to claim 1, characterized in that The expression corresponding to the deep reinforcement learning algorithm is as follows: ； in, represents a failure risk index of the drone; represents the risk value mapping function; Indicates The feature weight of each state feature; ; Indicates the total number of state features; Indicates the battery efficiency of the drone; Indicates The feature vector of the state features The state characteristic function value of ; represents the temperature influence coefficient; Indicates standard temperature; Indicates the difference between the current temperature and the standard temperature; Indicates the air density influence coefficient; Indicates the current air density; Indicates standard air density; Represents the battery efficiency influence coefficient.

6. The method according to claim 1, characterized in that The calculation formula corresponding to the battery efficiency of the drone is as follows: ； in, Indicates the battery efficiency of the drone; Indicates the benchmark value of drone battery efficiency under standard working conditions; represents the exponential function; Indicates standard temperature; Indicates the difference between the current temperature and the standard temperature; Indicates the air density influence coefficient; Indicates the current air density; Indicates standard air density; represents the load factor; Indicates the current battery current data of the drone; Indicates the maximum allowable current of the battery; represents the humidity influence coefficient; Represents the current humidity data of the drone; Indicates the maximum allowable humidity of the drone.

7. The method according to claim 1, characterized in that The environmental parameter data includes: environmental data and drone status data; The environmental data includes one or more of the following: environmental temperature, air density, humidity data, wind speed data and location data; The drone status data includes one or more of the following: battery voltage data, battery current data, and sensor data.

8. A UAV reliability verification system, characterized in that: include: A data acquisition module, used to obtain environmental parameter data of the UAV to be verified under non-standard working conditions; An efficiency calculation module, used to calculate the battery efficiency of the drone according to the environmental parameter data; A state estimation module, used to estimate the flight state of the UAV using a Kalman filter algorithm according to the environmental parameter data of the UAV and the battery efficiency, so as to obtain predicted flight state data of the UAV; A fault detection module, configured to perform fault detection on the drone using a deep reinforcement learning algorithm according to the predicted flight state data of the drone, obtain a fault risk index of the drone, and perform reliability verification on the drone according to the fault risk index; Wherein, battery efficiency compensation and air density correction are introduced into the Kalman filter algorithm.

9. An electronic device, characterized in that: include: at least one processor and memory; The memory and the processor are connected via a bus; The memory is used to store one or more programs; When the one or more programs are executed by the at least one processor, a method for verifying the reliability of a drone as described in any one of claims 1 to 7 is implemented.

10. A computing device readable storage medium, characterized in that: An execution program is stored thereon, and when the execution program is executed, a method for verifying the reliability of a drone as described in any one of claims 1 to 7 is implemented.

Images