Unmanned aerial vehicle risk assessment method, device and equipment based on unmanned aerial vehicle multi-source parameters

Abstract

The invention discloses an unmanned aerial vehicle risk assessment method based on unmanned aerial vehicle multi-source parameters, and relates to the technical field of unmanned aerial vehicle risk assessment, and the method comprises the steps: constructing an expected label for sample unmanned aerial vehicle multi-source parameters through unmanned aerial vehicle accident data, and obtaining the sample unmanned aerial vehicle multi-source parameters and corresponding labels; training the AI model according to the sample unmanned aerial vehicle multi-source parameters and the corresponding labels; and finally, the AI model after scheduling training identifies the multi-source parameters of the to-be-identified unmanned aerial vehicle, and determines an unmanned aerial vehicle risk assessment result, so that the loss risk and the corresponding loss degree of the unmanned aerial vehicle can be rapidly and accurately assessed, and workers are assisted to perform rapid insurance pricing.

Description

Technical Field

[0001] This application relates to the field of unmanned aerial vehicle (UAV) risk assessment technology, specifically to a method, apparatus, and equipment for UAV risk assessment based on multi-source parameters of UAVs. Background Technology

[0002] With the maturity and widespread adoption of drone technology, its applications are increasingly seen in agricultural plant protection, logistics and distribution, geographic surveying, and security patrols. The surge in drone ownership has created enormous insurance demand. However, traditional insurance pricing models have revealed numerous limitations and shortcomings when dealing with this emerging and dynamic insured object. Currently, drone insurance in the industry often employs a crude pricing strategy, typically pricing based solely on the purchase value of the drone or a single brand and model, failing to reflect the true risk profile of individual drones. This leads to overly homogenized drone insurance pricing. Traditional actuarial models heavily rely on ample historical claims data for a particular drone model or group of workers, resulting in insurance companies making unreasonable quotes or refusing coverage when new models emerge. Furthermore, the existing pricing process heavily relies on manual risk assessment and verification by underwriters, leading to inefficiency and hindering large-scale underwriting. Therefore, the drone insurance sector urgently needs a technological solution that overcomes these shortcomings, fully utilizing quantifiable drone data and leveraging advanced artificial intelligence technology to achieve efficient and accurate individual risk assessment and insurance pricing. Summary of the Invention

[0003] The purpose of this application is to provide a method, apparatus, and equipment for drone risk assessment based on multi-source parameters of drones, which solves the problem that the existing technology cannot conduct accurate and rapid risk assessment of drones, resulting in low pricing efficiency and inaccuracy.

[0004] This application is achieved through the following technical solution:

[0005] The first aspect of this application provides a method for assessing the risk of unmanned aerial vehicles (UAVs) based on multi-source parameters, including:

[0006] Collect multi-source parameters of the sample drone and drone accident data corresponding to the multi-source parameters of the sample drone;

[0007] Based on the drone accident data, expected labels are constructed for the multi-source parameters of the sample drone, thus obtaining the multi-source parameters of the sample drone and their corresponding labels.

[0008] The AI ​​model is trained based on the multi-source parameters of the sample drone and their corresponding labels to obtain the trained AI model.

[0009] Collect multi-source parameters of the drone to be identified, and schedule the trained AI model to identify the multi-source parameters of the drone to be identified, and determine the drone risk assessment result;

[0010] The drone risk assessment results are presented to staff to assist them in conducting risk assessments and insurance pricing for drones.

[0011] In one possible implementation, it also includes:

[0012] Based on the drone risk assessment results and drone multi-source parameters, target drone accident data is matched from all drone accident data, and the target drone accident data is displayed to staff to assist them in conducting drone risk assessments and insurance pricing.

[0013] In one possible implementation, the multi-source parameters of the sample UAV include: UAV static basic parameters, UAV flight performance parameters, UAV power performance parameters, UAV battery capability parameters, and UAV obstacle avoidance related parameters.

[0014] In one possible implementation, the drone obstacle avoidance parameters include the obstacle avoidance coverage space;

[0015] The method for obtaining the obstacle avoidance coverage space is as follows: In the formula, For maximum detection range, The obstacle avoidance sensor's horizontal field of view. This is the vertical view of the obstacle avoidance sensor.

[0016] In one possible implementation, the drone accident data includes the claim amount, accident type, extent of loss, drone model, and number of insurance policies issued.

[0017] In one possible implementation, the AI ​​model is trained based on the multi-source parameters of the sample UAV and their corresponding labels to obtain the trained AI model, including:

[0018] Using the multi-source parameters of the sample drone as input data and the labels corresponding to the multi-source parameters of the sample drone as expected output data, the AI ​​model is trained to obtain the trained AI model.

[0019] In one possible implementation, it also includes:

[0020] The Yeo-Johnson transform is used to preprocess the multi-source parameters of the sample UAVs and their corresponding labels to obtain the preprocessed multi-source parameters of the sample UAVs and their corresponding labels. The AI ​​model is then trained using the preprocessed multi-source parameters of the sample UAVs and their corresponding labels.

[0021] In one possible implementation, the AI ​​model employs the Tweedie regression model.

[0022] Based on the same inventive concept, this application also provides a drone risk assessment device based on multi-source parameters of a drone, comprising:

[0023] The sample data acquisition module is used to collect multi-source parameters of the sample UAV and UAV accident data corresponding to the multi-source parameters of the sample UAV.

[0024] The sample data processing module is used to construct expected labels for the multi-source parameters of the sample drone based on the drone accident data, and obtain the multi-source parameters of the sample drone and their corresponding labels.

[0025] The AI ​​model training module is used to train the AI ​​model based on the multi-source parameters of the sample UAV and their corresponding labels, and obtain the trained AI model.

[0026] The drone risk assessment module is used to collect multi-source parameters of the drone to be identified, and to schedule the trained AI model to identify the multi-source parameters of the drone to be identified, and determine the drone risk assessment result.

[0027] The drone risk feedback module is used to display the drone risk assessment results to staff to assist them in conducting risk assessments and insurance pricing for drones.

[0028] Based on the same inventive concept, this application also provides an electronic device, including a processor and a memory;

[0029] The memory stores computer-executed instructions;

[0030] The processor executes computer execution instructions stored in the memory, causing the processor to perform the UAV risk assessment method based on UAV multi-source parameters as described in any one of claims 1 to 8.

[0031] Compared with the prior art, this application has the following advantages and beneficial effects:

[0032] This application provides a drone risk assessment method based on multi-source parameters of drones. It constructs expected labels for the multi-source parameters of sample drones using drone accident data, obtaining the multi-source parameters and their corresponding labels for the sample drones. An AI model is then trained based on these parameters and labels. Finally, the trained AI model is used to identify the multi-source parameters of the drone to be identified, determining the drone risk assessment result. This method can quickly and accurately assess the loss risk and corresponding degree of loss of drones, thereby assisting staff in rapid insurance pricing. Attached Figure Description

[0033] To more clearly illustrate the technical solutions of the exemplary embodiments of this application, the accompanying drawings used in the embodiments will be briefly described below. It should be understood that the following drawings only show some embodiments of this application and should not be considered as a limitation of the scope. For those skilled in the art, other related drawings can be obtained based on these drawings without creative effort. In the drawings:

[0034] Figure 1 A flowchart illustrating a UAV risk assessment method based on multi-source parameters of a UAV, provided for embodiments of this application;

[0035] Figure 2 A schematic diagram of the structure of a UAV risk assessment device based on multi-source parameters of a UAV provided in this application embodiment;

[0036] Figure 3 This is a schematic diagram of the structure of an electronic device provided in an embodiment of this application;

[0037] The attached diagram shows the markings and corresponding component names:

[0038] 201-Sample data acquisition module, 202-Sample data processing module, 203-AI model training module, 204-UAV risk assessment module, 205-UAV risk feedback module, 301-Memory, 302-Processor, 303-Bus. Detailed Implementation

[0039] To make the objectives, technical solutions, and advantages of this application clearer, the following detailed description is provided in conjunction with the embodiments and accompanying drawings. The illustrative embodiments and descriptions of this application are only for explaining this application and are not intended to limit this application.

[0040] like Figure 1 As shown in the figure, this application provides a method for assessing the risk of a drone based on multi-source parameters of the drone, including:

[0041] S101. Collect multi-source parameters of the sample drone and drone accident data corresponding to the multi-source parameters of the sample drone;

[0042] S102. Construct expected labels for the multi-source parameters of the sample drone based on the drone accident data, and obtain the multi-source parameters of the sample drone and their corresponding labels.

[0043] S103. Train the AI ​​model based on the multi-source parameters of the sample UAV and their corresponding labels to obtain the trained AI model.

[0044] S104. Collect multi-source parameters of the drone to be identified, and schedule the trained AI model to identify the multi-source parameters of the drone to be identified, and determine the drone risk assessment result.

[0045] The drone risk assessment results are used to characterize the predicted loss level of the drone in an accident state. Therefore, when constructing expected labels for the multi-source parameters of the sample drone based on the drone accident data, the loss level of the drone in an accident state corresponding to the multi-source parameters of the sample drone should be used as the expected label.

[0046] Optionally, if the same sample UAV multi-source parameters correspond to multiple loss levels, the mean of the loss levels can also be used as the expected label.

[0047] S105. Present the drone risk assessment results to the staff to assist them in conducting risk assessments and insurance pricing for the drones.

[0048] It is worth noting that if the staff has a pre-set insurance pricing calculation formula, the predicted loss level from the drone risk assessment results can be substituted into the insurance pricing calculation formula set by the staff, and the result can be directly output to the staff, thereby further improving the staff's work efficiency.

[0049] The above technical solution can provide staff with a prediction of the extent of damage to the new drone in the event of an accident when the new drone leaves the factory. This provides staff with a basis for insurance pricing, assists them in quickly pricing insurance, and improves work efficiency.

[0050] In one possible implementation, it also includes:

[0051] Based on the drone risk assessment results and drone multi-source parameters, target drone accident data is matched from all drone accident data, and the target drone accident data is displayed to staff to assist them in conducting drone risk assessments and insurance pricing.

[0052] In one possible implementation, the multi-source parameters of the sample UAV include: UAV static basic parameters, UAV flight performance parameters, UAV power performance parameters, UAV battery capability parameters, and UAV obstacle avoidance related parameters.

[0053] For example, the static basic parameters of a drone include empty weight, maximum takeoff weight, dimensions, waterproof and dustproof rating, and sensor parameters and types; the drone's flight performance parameters include maximum flight speed, maximum endurance, wind resistance rating, communication and navigation system configuration, maximum power consumption, and maximum takeoff altitude; the drone's power performance parameters include motor performance; and the drone's battery capability parameters may include parameters related to battery output capacity and battery storage capacity. It is worth noting that the above parameters are merely examples of embodiments of this application. Other drone-related parameters can also be used to construct multi-source parameters for the drone. The multi-source parameters of the sample drone can all be represented numerically or using numerical encoding, thereby facilitating data learning and data recognition.

[0054] In one possible implementation, the drone obstacle avoidance parameters include the obstacle avoidance coverage space;

[0055] The method for obtaining the obstacle avoidance coverage space is as follows: In the formula, For maximum detection range, The obstacle avoidance sensor's horizontal field of view. This is the vertical view of the obstacle avoidance sensor.

[0056] Optionally, the obstacle avoidance parameters for the drone may also include the azimuth angle, which is calculated using the following formula: .

[0057] Optionally, by analyzing and statistically analyzing crash cases, the accident rate for obstacle avoidance in each direction can be obtained, and a weighted calculation coefficient can be provided:

[0058] Forward 0.30: The main risk direction for high-speed flight, accounting for 42% of all collision accidents;

[0059] 0.15 behind: Requirements for inverted flight shooting mode;

[0060] Lateral 0.20: Requirements for lateral obstacle avoidance and passage through narrow spaces (weighted by left and right sides);

[0061] Above 0.10: Indoor flight ceiling collision risk, but the incidence rate is <5%;

[0062] 0.25 below: Core risks during takeoff and landing, with terrain / building collisions accounting for 33% of accidents.

[0063] The obstacle avoidance performance score is obtained by weighted calculation of the obstacle avoidance coverage space in each direction:

[0064] Obstacle avoidance performance score = 0.30 Forward obstacle avoidance coverage area +0.15 Rear obstacle avoidance coverage space + Left-side obstacle avoidance coverage space +0.10 Right-direction obstacle avoidance coverage space +0.10 Obstacle avoidance coverage space above +0.25 The obstacle avoidance coverage area below.

[0065] Obstacle avoidance performance scores can also be used as parameters related to drone obstacle avoidance, further improving the accuracy of drone risk assessment.

[0066] In one possible implementation, the drone accident data includes the claim amount, accident type, extent of loss, drone model, and number of insurance policies issued.

[0067] In one possible implementation, the AI ​​model is trained based on the multi-source parameters of the sample UAV and their corresponding labels to obtain the trained AI model, including:

[0068] Using the multi-source parameters of the sample drone as input data and the labels corresponding to the multi-source parameters of the sample drone as expected output data, the AI ​​model is trained to obtain the trained AI model.

[0069] In one possible implementation, it also includes:

[0070] The Yeo-Johnson transform is used to preprocess the multi-source parameters and their corresponding labels of the sample UAVs, resulting in preprocessed multi-source parameters and their corresponding labels. These preprocessed parameters and labels are then used to train the AI ​​model. Furthermore, the AI ​​model can employ a Tweedie regression model.

[0071] Optionally, in addition to the Yeo-Johnson transform, data cleaning can be performed by handling missing and outlier values. Missing data for data-type features can be filled with KNN, and the inability parameter of data-type features can be set to 0. Feature encoding can be performed, and target encoding can be used for categorical features to reduce dimensionality while retaining the relationship information between features and target variables.

[0072] In this application, the Yeo-Johnson transform is used for data preprocessing. This transform effectively reduces the skewness of the data distribution, making it closer to a normal distribution and improving the stability of the model. The Tweedie regression model is specifically designed to handle the mixed distribution of numerous zero values ​​and right-skewed continuous values ​​in the data. This model can directly and accurately predict the extent of loss due to drone-related risks, effectively assisting staff in determining the accuracy of the final premium pricing. However, it is worth noting that other deep learning models, such as convolutional neural networks and long short-term memory networks, can also be used as AI models.

[0073] Optionally, the mutual information score method can be used to filter the multi-source parameters of the UAV. This method can capture complex nonlinear relationships without requiring assumptions about data distribution and is applicable to various data types with clear guidance. This method can eliminate redundant and irrelevant parameters, improve model training efficiency, enhance the model's generalization ability, and prevent overfitting. Cross-validation coding can also be used to perform unbiased evaluation of the risk assessment AI model's performance. The advantage of this method is that it smooths the data, ensuring that the model's performance evaluation on the validation set truly reflects its ability to generalize to unknown data, thus providing a reliable guarantee for model selection and tuning, and ultimately improving the accuracy of the results.

[0074] In this embodiment, three different strategies can be used to optimize the data model: Automated Machine Learning (AutoML), Grid Search, and Manual Parameter Tuning. The AutoML strategy efficiently and automatically finds the global optimum in the parameter space using advanced search strategies such as Bayesian optimization and evolutionary algorithms. In the Grid Search strategy, the administrator can predefine the candidate value range of key parameters, and the system iterates and evaluates all parameter combinations through cross-validation, selecting the best-performing set. A model diagnostic interface is also provided, allowing manual setting and iterative adjustment of parameters based on domain knowledge and insights into model behavior, enabling deep customization and fine-tuning of the risk assessment model for specific business scenarios.

[0075] like Figure 2 As shown, based on the same inventive concept, this application also provides a drone risk assessment device based on multi-source parameters of a drone, comprising:

[0076] The sample data acquisition module 201 is used to acquire multi-source parameters of the sample UAV and UAV accident data corresponding to the multi-source parameters of the sample UAV.

[0077] The sample data processing module 202 is used to construct expected labels for the multi-source parameters of the sample drone based on the drone accident data, and obtain the multi-source parameters of the sample drone and their corresponding labels.

[0078] AI model training module 203 is used to train the AI ​​model based on the multi-source parameters of the sample UAV and their corresponding labels, and obtain the trained AI model.

[0079] The drone risk assessment module 204 is used to collect multi-source parameters of the drone to be identified, and to schedule the trained AI model to identify the multi-source parameters of the drone to be identified, and determine the drone risk assessment result.

[0080] The drone risk feedback module 205 is used to display the drone risk assessment results to staff to assist staff in conducting drone risk assessments and insurance pricing.

[0081] The apparatus provided in this application embodiment can execute the above-described method and technical solution, and its principle and beneficial effects are similar, so they will not be described again here.

[0082] like Figure 3 As shown, based on the same inventive concept, this application also provides an electronic device, including a processor 302 and a memory 301; the memory 301 and the processor 302 are interconnected via a bus 303.

[0083] The memory 301 stores computer-executed instructions;

[0084] The processor 302 executes the computer execution instructions stored in the memory 301, causing the processor 302 to execute a UAV risk assessment method based on UAV multi-source parameters as described in any embodiment of this application.

[0085] For specific examples, memory may include, but is not limited to, random access memory (RAM), read-only memory (ROM), flash memory, first-in-first-out (FIFO) memory, and / or first-in-last-out (FILO) memory, etc.; specifically, processor may include one or more processing cores, such as a 4-core processor, an 8-core processor, etc. The processor may be implemented using at least one hardware form of DSP (Digital Signal Processing), FPGA (Field-Programmable Gate Array), PLA (Programmable Logic Array). Furthermore, the processor may include a main processor and coprocessors. The main processor, also known as the CPU (Central Processing Unit), is used to process data in the wake-up state; the coprocessor is a low-power processor used to process data in the standby state.

[0086] This application also provides a computer-readable storage medium storing computer-executable instructions, which, when executed by a processor, are used to implement the UAV risk assessment method based on multi-source parameters of UAVs as described in any of the above embodiments.

[0087] This application also provides a computer program product, including a computer program that, when executed by a processor, implements the UAV risk assessment method based on multi-source parameters of UAVs as described in any of the above embodiments.

[0088] All or part of the steps in the above method embodiments can be implemented by hardware related to program instructions. The aforementioned program can be stored in a readable memory. When the program is executed, it performs the steps of the above method embodiments; and the aforementioned memory (storage medium) includes: read-only memory (ROM), RAM, flash memory, hard disk, solid-state drive, magnetic tape, floppy disk, optical disk, and any combination thereof.

[0089] This application describes embodiments with reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of this application. It should be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processing unit of a general-purpose computer, special-purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processing unit of the computer or other programmable data processing apparatus, generate instructions for implementing the flowchart illustrations. Figure 1 One or more processes and / or boxes Figure 1 A device that provides the functions specified in one or more boxes.

[0090] These computer program instructions may also be stored in a computer-readable storage medium that can direct a computer or other programmable data processing device to function in a particular manner, such that the instructions stored in the computer-readable storage medium produce an article of manufacture including instruction means, which are implemented in a process Figure 1 One or more processes and / or boxes Figure 1 The function specified in one or more boxes.

[0091] These computer program instructions may also be loaded onto a computer or other programmable data processing equipment to cause a series of operational steps to be performed on the computer or other programmable equipment to produce a computer-implemented process, thereby providing instructions that execute on the computer or other programmable equipment for implementing the process. Figure 1 One or more processes and / or boxes Figure 1 The steps of the function specified in one or more boxes.

[0092] Other embodiments of this application will readily occur to those skilled in the art upon consideration of the specification and practice of the invention disclosed herein. This application is intended to cover any variations, uses, or adaptations of this application that follow the general principles of this application and include common knowledge or customary techniques in the art not disclosed herein. The specification and examples are to be considered exemplary only, and the true scope and spirit of this application are indicated by the claims.

[0093] It should be understood that this application is not limited to the precise structure described above and shown in the accompanying drawings, and various modifications and changes can be made without departing from its scope. The scope of this application is limited only by the appended claims.

Claims

1. A method for assessing the risk of unmanned aerial vehicles (UAVs) based on multi-source parameters, characterized in that, include: Collect multi-source parameters of the sample drone and drone accident data corresponding to the multi-source parameters of the sample drone; Based on the drone accident data, expected labels are constructed for the multi-source parameters of the sample drone, thus obtaining the multi-source parameters of the sample drone and their corresponding labels. The AI ​​model is trained based on the multi-source parameters of the sample drone and their corresponding labels to obtain the trained AI model. Collect multi-source parameters of the drone to be identified, and schedule the trained AI model to identify the multi-source parameters of the drone to be identified, and determine the drone risk assessment result; The drone risk assessment results are presented to staff to assist them in conducting risk assessments and insurance pricing for drones.

2. The UAV risk assessment method based on multi-source parameters of UAVs according to claim 1, characterized in that, Also includes: Based on the drone risk assessment results and drone multi-source parameters, target drone accident data is matched from all drone accident data, and the target drone accident data is displayed to staff to assist them in conducting drone risk assessments and insurance pricing.

3. The UAV risk assessment method based on multi-source parameters of UAVs according to claim 1, characterized in that, The multi-source parameters of the sample UAV include: UAV static basic parameters, UAV flight performance parameters, UAV power performance parameters, UAV battery capability parameters, and UAV obstacle avoidance related parameters.

4. The UAV risk assessment method based on multi-source parameters of UAVs according to claim 1, characterized in that, The obstacle avoidance parameters for the UAV include the obstacle avoidance coverage space; The method for obtaining the obstacle avoidance coverage space is as follows: In the formula, For maximum detection range, The obstacle avoidance sensor's horizontal field of view. This is the vertical view of the obstacle avoidance sensor.

5. The UAV risk assessment method based on multi-source parameters of UAVs according to claim 1, characterized in that, The drone accident data includes the claim amount, accident type, extent of loss, drone model, and number of insurance policies issued.

6. The UAV risk assessment method based on multi-source parameters of UAVs according to claim 1, characterized in that, The AI ​​model is trained based on the multi-source parameters of the sample drones and their corresponding labels to obtain the trained AI model, including: Using the multi-source parameters of the sample drone as input data and the labels corresponding to the multi-source parameters of the sample drone as expected output data, the AI ​​model is trained to obtain the trained AI model.

7. The UAV risk assessment method based on multi-source parameters of UAVs according to claim 1, characterized in that, Also includes: The Yeo-Johnson transform is used to preprocess the multi-source parameters of the sample UAVs and their corresponding labels to obtain the preprocessed multi-source parameters of the sample UAVs and their corresponding labels. The AI ​​model is then trained using the preprocessed multi-source parameters of the sample UAVs and their corresponding labels.

8. The UAV risk assessment method based on multi-source parameters of UAVs according to claim 1, characterized in that, The AI ​​model uses the Tweedie regression model.

9. A drone risk assessment device based on multi-source parameters of a drone, characterized in that, include: The sample data acquisition module is used to collect multi-source parameters of the sample UAV and UAV accident data corresponding to the multi-source parameters of the sample UAV. The sample data processing module is used to construct expected labels for the multi-source parameters of the sample drone based on the drone accident data, and obtain the multi-source parameters of the sample drone and their corresponding labels. The AI ​​model training module is used to train the AI ​​model based on the multi-source parameters of the sample UAV and their corresponding labels, and obtain the trained AI model. The drone risk assessment module is used to collect multi-source parameters of the drone to be identified, and to schedule the trained AI model to identify the multi-source parameters of the drone to be identified, and determine the drone risk assessment result. The drone risk feedback module is used to display the drone risk assessment results to staff to assist them in conducting risk assessments and insurance pricing for drones.

10. An electronic device, characterized in that, Including processor and memory; The memory stores computer-executed instructions; The processor executes computer execution instructions stored in the memory, causing the processor to perform the UAV risk assessment method based on UAV multi-source parameters as described in any one of claims 1 to 8.

Images