Unmanned aerial vehicle equipment system performance evaluation method

Abstract

The invention provides an unmanned aerial vehicle equipment system efficiency evaluation method. The method comprises the following steps: acquiring real-time operation data of each subsystem in an unmanned aerial vehicle equipment system; according to the obtained real-time operation data, constructing an efficiency evaluation model of each subsystem, and performing weighted calculation on the efficiency index of each subsystem by adopting a preset weight coefficient to obtain an efficiency score of each subsystem; the method comprises the following steps: acquiring real-time environment data by adopting an environment sensing module aiming at uncertainty factors of an external environment, and inputting the environment data into a pre-established environment influence model to obtain a correction coefficient of the environment to the efficiency of each subsystem; and performing dynamic adjustment on the efficiency score of each subsystem according to the environment correction coefficient to obtain the efficiency score of the subsystem under a specific environment condition, and inputting the adjusted score into the multi-subsystem collaborative efficiency model to perform efficiency evaluation of each subsystem.

Description

Technical Field

[0001] The present invention relates to the field of information technology, and in particular to a method for evaluating the effectiveness of an unmanned aerial vehicle equipment system. Background Art

[0002] Background: In the evaluation of the effectiveness of UAV equipment systems, there is a key technical problem, that is, how to accurately quantify the dynamic adaptability of UAV systems in multi-task collaboration in a complex and changing environment. Specifically, UAV equipment systems are usually composed of multiple subsystems, including reconnaissance, communication, navigation and other modules, and the performance of these modules in different mission scenarios varies significantly. However, due to the interdependence between the subsystems, the performance change of a certain module will directly affect the overall performance of the entire system.

[0003] In addition, the uncertainty of the operating environment further increases the difficulty of the evaluation. For example, factors such as weather conditions, electromagnetic interference, and terrain complexity will have varying degrees of impact on the performance of drones. In the evaluation process, how to take these external factors into consideration and establish a dynamic performance model that adapts to them is a technical problem that needs to be solved urgently. Summary of the invention

[0004] The present invention provides a method for evaluating the effectiveness of an unmanned aerial vehicle equipment system, which mainly includes: Obtain real-time operating data of each subsystem in the UAV equipment system, including the target recognition accuracy of the video module, the transmission rate of the communication module, and the positioning error of the navigation module; Based on the real-time operation data obtained, the performance evaluation model of each subsystem is constructed, and the performance indicators of each subsystem are weighted and calculated using the preset weight coefficient to obtain the performance score of each subsystem; In view of the uncertain factors of the external environment, including weather conditions, electromagnetic interference and terrain complexity, the environmental perception module is used to collect real-time environmental data, and the environmental data is input into the pre-established environmental impact model to obtain the correction coefficient of the environment on the performance of each subsystem; According to the environmental correction coefficient, the performance score of each subsystem is dynamically adjusted to obtain the subsystem performance score under specific environmental conditions, and the adjusted score is input into the multi-subsystem collaborative performance model to evaluate the performance of each subsystem.

[0005] The technical solution provided by the embodiment of the present invention may have the following beneficial effects: The present invention discloses a method for evaluating the effectiveness of an unmanned aerial vehicle equipment system. The method acquires the real-time operation data of each subsystem, constructs an effectiveness evaluation model, and considers the impact of external environmental factors on the effectiveness. The present invention uses an environmental perception module to collect real-time environmental data and dynamically adjusts the subsystem effectiveness score. The method realizes the dynamic adaptability evaluation of the unmanned aerial vehicle equipment system in complex environments and diversified tasks, and improves the accuracy and practicality of the effectiveness evaluation. BRIEF DESCRIPTION OF THE DRAWINGS

[0006] Figure 1 The present invention is a flowchart of a method for evaluating the effectiveness of an unmanned aerial vehicle equipment system. DETAILED DESCRIPTION

[0007] In order to further understand the content of the present invention, the present invention is described in detail in conjunction with the accompanying drawings and embodiments. The present application is further described in detail below in conjunction with the accompanying drawings and embodiments. It is understood that the specific embodiments described herein are only used to explain the relevant inventions, rather than to limit the invention. It is also necessary to explain that, for ease of description, only the parts related to the invention are shown in the accompanying drawings.

[0008] like Figure 1 The present embodiment provides a method for evaluating the effectiveness of a drone equipment system, which may specifically include: S101. Acquire real-time operating data of each subsystem in the UAV equipment system, including the target recognition accuracy of the video module, the transmission rate of the communication module, and the positioning error of the navigation module.

[0009] The target recognition accuracy data of the drone video module is obtained through the data acquisition interface, and the data preprocessing algorithm is used to remove outliers to generate standardized recognition accuracy data. The original transmission rate information is obtained from the communication module, and the noise interference is filtered out through the data cleaning algorithm to obtain stable transmission rate data. The positioning error information of the navigation module is extracted, and the missing values ​​are supplemented by the data interpolation algorithm to form a complete positioning error data set.

[0010] Specifically, in the UAV equipment system, obtaining real-time operating data of each subsystem is one of the key tasks. For the target recognition accuracy of the video module, real-time analysis can be performed through deep learning algorithms such as YOLOv5. YOLOv5 can achieve 95% target recognition accuracy while processing 60 frames of images per second. The transmission rate of the communication module can be monitored through the TCP / IP protocol, and the real-time data transmission rate can reach 100Mbps per second to ensure efficient transmission of information. The positioning error of the navigation module is optimized using the Kalman filter algorithm, which can reduce the positioning error from 10 meters to less than 2 meters, significantly improving navigation accuracy.

[0011] S102: constructing a performance evaluation model for each subsystem based on the acquired real-time operation data, performing weighted calculation on the performance indicators of each subsystem using a preset weight coefficient, and obtaining a performance score for each subsystem.

[0012] Specifically, the performance evaluation model S of the video module target recognition accuracy opt video It can be obtained by the following formula:

[0013] Among them, P is the precision, R is the recall, IoU is the intersection over union, S det is the recognition speed, F fp is the false positive rate, and its corresponding weight coefficient is w p 、w r 、w iou 、w s 、w f , w p +w r +w iou +w s +w f =1.

[0014] In practical applications, not only accurate recognition is required, but also the recognition results need to be given in a timely and fast manner, especially in the case of fast-moving drones. The recognition speed can be measured by the number of image frames processed per unit time or the time taken to recognize a single target. In addition, false detections will interfere with the judgment of operators and increase the complexity of subsequent processing. The higher the false detection rate, the lower the reliability of target recognition. The false detection rate can be calculated by the ratio of the number of falsely detected targets to the total number of detected targets.

[0015] The intersection over union (IoU) is an important indicator used in target detection tasks to measure the degree of match between the detection result and the true target position. It quantifies the degree of match between the bounding box and the ground truth box by calculating the ratio of the overlapping area between the bounding box and the ground truth box to their union area.

[0016] In one embodiment, the intersection over union (IoU) can be calculated by the following formula: Define the detection box as B det , the real box is B gt , their coordinates are: .

[0017] The area of ​​the overlapping part of the two detection boxes and the rectangular box of the real box is recorded as I; the total area of ​​the two rectangular boxes minus the area of ​​the overlapping part, that is, U=A det +A gt−I, so the calculation formula for the intersection over union (IoU) is: .

[0018] In one embodiment, the performance evaluation model S of the transmission rate of the communication module opt c satisfy:

[0019] Among them, R avg is the average transmission rate, R range is the transmission rate fluctuation range, R max_range is the maximum transmission rate fluctuation range allowed, R success is the percentage of data packets successfully transmitted within a specific time, L trans is the transmission delay index, F int is the link interruption frequency, and its corresponding weight coefficients are avg 、a ran 、a suc 、a l 、a f , and satisfies a avg +a ran +a suc +a l +a f =1; L max is the theoretical or maximum permissible propagation delay, expressed as 1−L max / L trans To reflect the impact of transmission delay on communication performance, that is, the smaller the delay, the higher the score of this part.

[0020] In one embodiment, the performance evaluation model S of the positioning error of the navigation module opt nav satisfy:

[0021] Among them, E mean is the mean positioning error, E std is the standard deviation of positioning error, E over is the ratio of color vision localization error, F update To consider the update frequency of positioning data, C consis is the multi-dimensional positioning error consistency index; and the corresponding weight coefficient b mean , b std , b over , b f , b cons , and satisfies b mean +b std +b over +bf +b cons =1,E max_mean and E max_std are the maximum allowable positioning error mean and the maximum allowable positioning error standard deviation respectively. max_mean As a divisor benchmark, it is used to convert the actual positioning error mean E mean Standardization processing. max_std It also plays a similar role as the normalization benchmark of the positioning error mean.

[0022] In a fast and dynamic environment, such as when a drone is performing high-speed maneuvering flight, timely updating of positioning data is crucial to navigation accuracy. The higher the update frequency, the better it can reflect the real-time position of the drone and improve navigation accuracy. The positioning data update frequency can be measured by the number of updates of the positioning data per unit time. The multi-dimensional positioning error consistency index can quantify the discrete degree of positioning errors in different directions and measure whether the distribution of positioning errors in each direction is uniform and consistent. If the positioning errors in each direction are consistent, it means that the positioning performance of the navigation module in each direction is relatively balanced and the overall positioning efficiency is higher.

[0023] S103. In view of the uncertain factors of the external environment, including weather conditions, electromagnetic interference and terrain complexity, an environmental perception module is used to collect real-time environmental data, and the environmental data is input into a pre-established environmental impact model to obtain the correction coefficient of the environment on the performance of each subsystem.

[0024] Specifically, the main factors affecting weather conditions are visibility correction factor C wea_d and the precipitation correction factor C wea_p .

[0025] Define the visibility correction factor C for the weather conditions described wea_d as follows:

[0026] Among them, D wea is the actual visibility, D wea_max is the maximum effective value of visibility (for example, in a specific application scenario, visibility of 10 km has little impact on the drone), D wea_min It is the minimum effective value of visibility (for example, when visibility is 0.5km, it has a great impact on the drone).

[0027] Define the precipitation correction factor C in the weather conditions described wea_p as follows:

[0028] Among them, P wea is the precipitation intensity, Pwea_max is the maximum effective value of precipitation intensity (such as the corresponding value of rainstorm intensity), P wea_min It is the minimum valid value of precipitation intensity (e.g. the value is 0 when there is no precipitation).

[0029] Combining the above visibility correction factor C wea_d and the precipitation correction factor C wea_p , obtain the comprehensive correction coefficient C of weather conditions wea for: .

[0030] It is understood that in other embodiments, the high wind correction factor C may be increased. wea_w Etc., no restriction is made here.

[0031] Define the electromagnetic interference correction factor C inter for:

[0032] Among them, E inter_max It is the maximum permissible value of electromagnetic interference intensity. If it exceeds this value, the communication and navigation of the drone may be seriously affected. inter_min It is the minimum value of electromagnetic interference intensity, usually the interference intensity of the background electromagnetic environment.

[0033] The terrain complexity correction factor mainly includes two aspects: undulation correction and obstacle density correction: Wherein, the fluctuation correction coefficient C rough satisfy:

[0034] Among them, T rough is the actual terrain relief, T rough_max is the maximum effective value of terrain relief (e.g., the relief of high mountain terrain is 100 meters), T rough_min It is the minimum effective value of terrain relief (e.g. the relief of plain terrain is 0 meters).

[0035] Define the obstacle correction factor C obs for:

[0036] Among them, T obs is the obstacle density, T obs_max is the maximum effective value of obstacle density (such as the corresponding value of obstacle density in dense urban environment), T obs_min is the minimum valid value of obstacle density (e.g., close to 0 in an open desert environment).

[0037] Therefore, the terrain complexity comprehensive correction coefficient Cterrain for: .

[0038] The real-time environmental data of weather conditions, electromagnetic interference and terrain complexity are collected through the environmental perception module. The collected real-time environmental data is input into the pre-established environmental impact model, and the model processes the data based on preset rules. The environmental impact model calculates the weights of weather conditions, electromagnetic interference and terrain complexity on the performance of the subsystem based on the input data. The weighted algorithm is used to combine the weight values ​​to calculate the comprehensive impact value of the environment on the performance of each subsystem. Based on the comprehensive impact value, the correction coefficient of the environment on the performance of each subsystem is generated. The correction coefficient is input into the subsystem control module to adjust the operating parameters of the subsystem. Based on the correction coefficient, the subsystem control module optimizes the operating status of the subsystem to adapt to environmental changes.

[0039] S104. Dynamically adjust the performance score of each subsystem according to the environmental correction coefficient to obtain the subsystem performance score under specific environmental conditions, and input the adjusted score into the multi-subsystem collaborative performance model.

[0040] Specifically, the video module is greatly affected by weather conditions and terrain complexity, and is relatively less affected by electromagnetic interference. Therefore, the performance evaluation model S of the target recognition accuracy of the video module is final video It can be corrected by the following formula: .

[0041] w video_wea 、w video_terrain They represent the weight coefficients of weather conditions and terrain complexity on the performance of the video module, satisfying w video_wea +w video_terrain =1.

[0042] However, the communication module is greatly affected by electromagnetic interference and terrain complexity, but less affected by weather conditions. Therefore, the performance evaluation model S of the communication module target recognition accuracy is c final It can be corrected by the following formula: .

[0043] w comm_inter 、w comm_terrain are weight coefficients representing the impact of electromagnetic interference and terrain complexity on the performance of the communication module, respectively, satisfying w comm_inter +w comm_terrain =1.

[0044] As a further improvement, in other embodiments, the precipitation correction coefficient C in the weather condition can also bewea_p As a performance evaluation model S final c Influencing factors.

[0045] However, the navigation module is greatly affected by terrain complexity and weather conditions, so the performance evaluation model S of the navigation module target recognition accuracy is final nav It can be corrected by the following formula:

[0046] w nav_terrain and w nav_wea Respectively represent the weight coefficients of the influence of terrain complexity and weather conditions on the effectiveness of the navigation module, satisfying w nav_terrain +w nav_wea =1.

[0047] Obtain the initial performance score of the subsystem, match the correction value of the current environment from the preset environmental correction coefficient library, multiply the initial performance score by the environmental correction coefficient, and obtain the adjusted subsystem performance score. Adopt the structure of the collaborative effectiveness model, and input the adjusted subsystem performance score as the input value into the multi-system collaborative effectiveness model. According to the output results of the collaborative effectiveness model, determine the overall performance score of the multi-system in a specific environment. If the overall performance score is lower than the preset threshold, readjust the environmental correction coefficient and calculate the subsystem performance score again. Use the regression analysis algorithm to predict the trend of changes in the collaborative effectiveness of multiple systems under different environmental correction coefficients. According to the prediction results, select the optimal environmental correction coefficient to complete the dynamic optimization of the multi-system collaborative effectiveness model.

[0048] The description of the above embodiments is only used to help understand the technical solutions and core ideas of the present application. Ordinary technical personnel in this field should understand that they can still modify the technical solutions recorded in the aforementioned embodiments, or replace some of the technical features therein with equivalents. However, these modifications or replacements do not make the essence of the corresponding technical solutions deviate from the scope of the technical solutions of the embodiments of the present application.

Claims

1. A method for evaluating the effectiveness of an unmanned aerial vehicle equipment system, characterized in that: The method comprises: The real-time operation data of each subsystem in the UAV equipment system is obtained, and the subsystems include the target recognition accuracy of the video module, the transmission rate of the communication module, and the positioning error of the navigation module; based on the real-time operation data obtained, a performance evaluation model of each subsystem is constructed, and the performance indicators of each subsystem are weighted and calculated using a preset weight coefficient to obtain the performance score of each subsystem; in view of the uncertainty factors of the external environment, including weather conditions, electromagnetic interference and terrain complexity, an environmental perception module is used to collect real-time environmental data, and the environmental data is input into a pre-established environmental impact model to obtain the correction coefficient of the environment on the performance of each subsystem; according to the environmental correction coefficient, the performance score of each subsystem is dynamically adjusted to obtain the subsystem performance score under specific environmental conditions, and the adjusted score is input into a multi-subsystem collaborative performance model to evaluate the performance of each subsystem.

2. The method for evaluating the effectiveness of an unmanned aerial vehicle equipment system according to claim 1, wherein: The step of obtaining the real-time operation data of each subsystem in the UAV equipment system, wherein each subsystem includes the target recognition accuracy of the video module, the transmission rate of the communication module, and the positioning error of the navigation module, specifically includes: The target recognition accuracy data of the drone video module is obtained through the data acquisition interface, and the data preprocessing algorithm is used to remove outliers to generate standardized recognition accuracy data; the original transmission rate information is obtained from the communication module, and the noise interference is filtered out through the data cleaning algorithm to obtain stable transmission rate data; the positioning error information of the navigation module is extracted, and the missing values ​​are supplemented by the data interpolation algorithm to form a complete positioning error data set.

3. The method for evaluating the effectiveness of an unmanned aerial vehicle equipment system according to claim 1, wherein: The performance evaluation model S of the target recognition accuracy of the video module opt video Obtained by the following formula: Among them, P is the precision, R is the recall, IoU is the intersection over union, S det is the recognition speed, F fp is the false positive rate, and its corresponding weight coefficient is w p 、w r 、w iou 、w s 、w f , w p +w r +w iou +w s +w f =1.

4. The method for evaluating the effectiveness of an unmanned aerial vehicle equipment system according to claim 3, wherein: The performance evaluation model S of the transmission rate of the communication module opt c satisfy: Among them, R avg is the average transmission rate, R range is the transmission rate fluctuation range, R max_range is the maximum transmission rate fluctuation range allowed, R success is the percentage of data packets successfully transmitted within a specific time, L trans is the transmission delay index, F int is the link interruption frequency, and its corresponding weight coefficients are avg 、a ran 、a suc 、a l 、a f , and satisfies a avg +a ran +a suc +a l +a f =1; L max is the theoretical or maximum permissible transmission delay.

5. The method for evaluating the effectiveness of an unmanned aerial vehicle equipment system according to claim 4, wherein: The effectiveness evaluation model S of the positioning error of the navigation module opt nav satisfy: Among them, E mean is the mean positioning error, E std is the standard deviation of positioning error, E over is the ratio of color vision localization error, F update To consider the update frequency of positioning data, C consis is the multi-dimensional positioning error consistency index; and the corresponding weight coefficient b mean 、b std 、b over 、b f 、b cons , and satisfies b mean +b std +b over +b f +b cons =1,E max_mean and E max_std They are the maximum allowable positioning error mean and the maximum allowable positioning error standard deviation respectively.

6. The method for evaluating the effectiveness of an unmanned aerial vehicle equipment system according to claim 5, wherein: The effectiveness evaluation model S of the communication module target recognition accuracy c final Corrected by the following formula: w comm_inter 、w comm_terrain are weight coefficients representing the impact of electromagnetic interference and terrain complexity on the performance of the communication module, respectively. inter is the electromagnetic interference correction factor, C terrain is the comprehensive correction coefficient of terrain complexity; satisfying w comm_inter +w comm_terrain =1.

7. The method for evaluating the effectiveness of an unmanned aerial vehicle equipment system according to claim 6, wherein: The performance evaluation model S of the navigation module target recognition accuracy final nav Corrected by the following formula: w nav_terrain and w nav_wea Respectively represent the weight coefficients of the influence of terrain complexity and weather conditions on the effectiveness of the navigation module, satisfying w nav_terrain +w nav_wea =1, C wea It is the comprehensive correction factor of weather conditions.

Images