Unmanned aerial vehicle performance testing system and method

Abstract

This invention discloses a UAV performance testing system and method, including a preprocessing module, a performance evaluation module, and an environment adaptation module. The advantages of this invention are: it constructs an intelligent testing system by employing three major modules—preprocessing, performance evaluation, and environment adaptation; the preprocessing module eliminates outliers using a three-standard-deviation principle to ensure data accuracy; the performance evaluation module decomposes core performance into three sub-performance categories, trains a model based on historical data, and dynamically adjusts the weights to make the evaluation more aligned with mission requirements; the environment adaptation module divides environmental dimensions into multiple scoring categories, identifies interference sources through the model, and matches corresponding strategies to improve adaptability to complex environments; the system achieves a closed loop of data processing, evaluation, and response, requiring minimal manual intervention, and possesses continuous learning and adaptive adjustment capabilities, significantly improving the accuracy, comprehensiveness, and efficiency of testing.

Description

Technical Field

[0001] This invention relates to the technical field of unmanned aerial vehicles (UAVs), and particularly to UAV performance testing systems and methods. Background Technology

[0002] With the rapid development of drone technology and the continuous expansion of its application scenarios, the scientific, accurate, and efficient testing and evaluation of its performance has become crucial. Currently, in the field of drone performance testing, existing technical solutions typically have the following shortcomings, making it difficult to meet the increasingly complex testing needs:

[0003] Existing systems typically lack real-time data cleaning and intelligent verification capabilities, failing to effectively identify and remove outliers caused by equipment malfunctions or environmental interference. They often rely on fixed thresholds for judgment, resulting in poor adaptability and low-quality data for evaluation. Current evaluation methods mostly use pre-set fixed indicators and weights, unable to dynamically adjust based on different task types or historical performance. This weakens the correlation between evaluation results and actual task effectiveness, making it difficult to provide targeted optimization guidance. Existing tests often consider environmental factors in isolation and statically, failing to comprehensively reflect the combined impact of complex environments. Especially when facing dynamic threats such as communication interference, systems mostly only record performance degradation, unable to proactively identify interference types and trigger adaptive strategies such as channel switching and safe return-to-home. This results in the unfulfilled verification and improvement of UAVs' combat capabilities in complex environments. The testing process still relies heavily on manual operation and judgment, failing to form an automated closed loop from data collection and analysis to response, leading to low efficiency and poor consistency. Furthermore, the system lacks continuous learning capabilities based on historical data, preventing it from self-optimizing the evaluation model. Summary of the Invention

[0004] To solve the above-mentioned technical problems, the technical solution of the present invention is: a UAV performance testing system and method, including a preprocessing module, a performance evaluation module and an environment adaptation module;

[0005] The preprocessing module preprocesses the collected data, determines the fluctuation range, identifies outliers based on the fluctuation range, removes outliers, and improves the accuracy of the collected data.

[0006] The performance evaluation module categorizes core performance based on sub-performance categories that affect core performance, calculates scores for each sub-performance category, and obtains the total core performance score by weighted summation of the sub-performance scores. It also acquires historical data, analyzes the sub-performance score data within the historical data, and determines the target weight coefficients for the corresponding sub-performance scores. Finally, it builds a model, trains it on historical data based on the target weight coefficients, obtains multiple training models, and dynamically adjusts the corresponding weight coefficients for the sub-performance scores based on the final trained model.

[0007] The environment adaptation module categorizes environment adaptability based on the environmental categories that affect it, calculates scores for each environmental category, and obtains the total environment adaptability score by weighted summation of the environmental category scores. It acquires test data, analyzes the environmental category scores in the test data, identifies the type of interference source by combining interference signal characteristics, and correlates them with environment adaptability performance. Based on the type of interference source, it matches preset response strategies and dynamically adjusts the UAV's working mode or communication strategy under corresponding environmental conditions to improve its performance maintenance capability in complex environments.

[0008] Preferably, the data preprocessing module performs the following steps:

[0009] E1: Sort the collected data according to the collection time, and sort the corresponding items collected at the same time. averaging the data and standard deviation The calculation, and the mean obtained from the calculation. and standard deviation Collect data fluctuation range for corresponding items The settings compare the collected data of the corresponding item with the fluctuation range of the corresponding item.

[0010] E2: Mark the corresponding data items that are outside the fluctuation range as outliers and record the number of outliers. ,like If the collected data is abnormal, the data will be re-tested; if If outliers are removed, the mean of the remaining corresponding test data after outlier removal is calculated. The calculation, and the mean obtained from the calculation. This refers to the corresponding data detected at the corresponding time.

[0011] Preferably, the analysis steps for the core performance total score by the performance evaluation module are as follows:

[0012] G1: Overall Core Performance Score , , , These are the weighting coefficients;

[0013] G2: Flight Performance Score Hovering accuracy score , and These are vertical hovering deviation and horizontal hovering deviation; maneuverability score. , and These are the actual maximum horizontal velocity and the actual maximum acceleration, respectively. and These are the reference velocity and reference acceleration defined by the type of drone. Acceleration due to gravity; battery life score , and These are the actual battery life and the target battery life, respectively; navigation accuracy score. , For dynamic positioning deviation;

[0014] G3: Powertrain Performance Score Power output score , and These are the actual maximum thrust and the target thrust, respectively. For dynamic response delay; energy system score , and These are the actual battery capacity and the designed capacity, respectively. Low-temperature discharge capacity decay rate; power redundancy score , This refers to the failover time.

[0015] G4: Load Performance Score Payload score , and These are the actual maximum load and the target load, respectively; load performance score. , This represents the total number of load performance indicators. For the first The weighting coefficient of the load performance index, and The first The actual maximum load and target load of the load performance index.

[0016] Preferably, the performance evaluation module performs the target weight analysis steps as follows:

[0017] H1: Obtain historical data and score flight performance from the historical data. Powertrain performance score Load performance score The data is acquired and categorized according to test time to obtain several test group data sets. These test group data sets are then further categorized according to the task type tested, resulting in several task group data sets. Test group data sets within the same task group are retrieved. Flight performance scores are then calculated from the retrieved test group data. Powertrain performance score Load performance score For comparison, data with the same two types of data in the test group are grouped together, and the change in the other type of data in the same group is compared with the total core performance score. The corresponding changes are recorded to obtain the total core performance score caused by another type of data unit change. Change;

[0018] H2: The overall core performance score when the three types of data undergo unit changes using the same method. Change data , , Then determine the target weight of the corresponding data item. , , .

[0019] Preferably, the analysis steps for the performance evaluation module to predict the final weights are as follows:

[0020] J1: Based on total sample size The weighting coefficients corresponding to the sub-performance scores are set to the average weights of the dataset for initial prediction. Then the first The first iteration obtained Prediction weights for each sample During model iteration, residual calculations and decision tree fitting are performed for each iteration, and prediction weights are updated. This continues until the preset number of iterations is reached. Stop iterating when the validation set loss no longer decreases, and obtain the final prediction weights. ;

[0021] J2: The final trained model is a weighted sum of decision trees. ,in , The initial weights and the learning rate are... , For the first The first round A decision tree fitted with weight dimensions, whose output is the input features. Mapped leaf node scores; final weights And ensure that the sum of the three weights is 1.

[0022] Preferably, the analysis steps for the overall environmental adaptability score by the environmental adaptability module are as follows:

[0023] K1: Overall score for environmental adaptability , , , These are the weighting coefficients;

[0024] K2: Meteorological Environment Score Wind resistance score , and These are the actual stable wind resistance speed and the design wind resistance speed, respectively; temperature and humidity adaptation score. , and These are the deviations between the actual operating temperature / humidity range and the design range; and the precipitation / dust protection score. , and These are the dustproof / waterproof ratings, respectively.

[0025] K3: Geographical Environment Score Terrain adaptation score , Terrain variation with height; Electromagnetic immunity score , The bit error rate in communication under electromagnetic interference conditions;

[0026] K4: Special Environment Score High-altitude adaptability score , Dynamic attenuation rate at an altitude of 5000m; marine adaptability score. , For sea surface positioning deviation, The corrosion rate is shown after the salt spray test.

[0027] Preferably, the analysis steps for the environment adaptation module to predict the type of interference source are as follows:

[0028] L1: Building the dataset , For interference feature vectors, Label the interference sources; construct the training model: Constraints: , ,in The normal vector of the classification hyperplane, To adjust the bias term at the classification boundary position, As slack variables, A penalty coefficient to control the strictness of classification;

[0029] L2: Solve the constructed trained model to obtain... and The optimal value is then the final model after training: , To collect interference features in real time, by... Inputting the data into the model can accurately predict the type of interference source. .

[0030] Preferably, the environmental adaptation module takes the following countermeasures based on the predicted interference source type:

[0031] Q1: If the interference source is determined to be WiFi interference, the spectrum analysis function is integrated into the communication module of the drone. The spectrum is scanned at set time intervals, and the communication is dynamically switched to the channel with the least interference based on the signal-to-noise ratio or bit error rate.

[0032] Q2: If the interference source is determined to be base station interference, the radio module of the drone can be used to monitor the spectrum occupancy in real time and dynamically avoid the concentrated frequency bands of the base station for communication.

[0033] Q3: If the interference source is determined to be GPS suppression interference, the "safe mode" logic is integrated into the drone's flight control system. When the interference continues to exceed the threshold, the drone will automatically execute a return-to-home or hovering command and activate a preset safety strategy to prevent the drone from going out of control.

[0034] Preferably, the drone performance testing method includes the following steps:

[0035] U1: After aligning all collected data by timestamp, the system performs statistical analysis on multiple sets of measurements of the same parameter at the same time. First, it calculates the mean and standard deviation of the data set. Then, based on the "three-times-standard-deviation principle," it determines the reasonable fluctuation range of the data. Values ​​exceeding this range are marked as outliers. If the proportion of outliers exceeds 30%, the system determines that the data acquisition process is abnormal and automatically triggers a retest mechanism. If the retest result is still abnormal, the system determines that the acquisition equipment is faulty, automatically triggers an audible and visual alarm, and displays "Acquisition equipment abnormal" on the monitoring interface, guiding staff to perform timely repairs.

[0036] U2: Core performance is broken down into three main categories: flight performance, propulsion system performance, and payload performance. Each category is further subdivided into multiple sub-indicators. For each sub-indicator, clear scoring rules are set, and the performance scores for each category are obtained through weighted summation. Historical test data is retrieved and grouped for analysis according to mission type. By statistically analyzing the impact of various performance indicators on overall performance in different missions, the initial weights of each performance category are preliminarily determined. A weight prediction model is constructed and trained using the mapping relationship between mission characteristics and optimal weights in historical data, enabling the system to dynamically adjust performance weights based on the characteristics of the current test mission.

[0037] U3: Real-time monitoring of electromagnetic signal characteristics in the environment, combined with the current communication error rate, positioning error and other status parameters of the UAV, to identify the type of interference source through a pre-trained classification model; once the interference type is identified, the system automatically matches and executes the preset response strategy.

[0038] The beneficial effects of this invention are:

[0039] 1. Through a modular and intelligent data processing and evaluation system, the accuracy, comprehensiveness, and adaptability of UAV performance testing have been significantly improved. The performance evaluation module constructs a scoring model that decomposes core performance into three categories: flight performance, power system performance, and payload performance. Each category has multiple sub-indicators with clear calculation formulas and dynamic weights. By combining historical data to train a weight prediction model, the weights are adaptively adjusted according to the task type, making the evaluation results more task-specific and scientific. The environment adaptation module further incorporates environmental factors into the evaluation system, constructs three scoring models for meteorological, geographical, and special environments, and achieves intelligent judgment of interference source types and matching of response strategies through signal feature recognition and classification models, significantly improving the performance stability and safety of UAVs in complex environments.

[0040] 2. By integrating machine learning models through the environment adaptation module, the system can dynamically optimize weight allocation and interference identification based on historical data and real-time features, and continuously learn. The environment adaptation module not only stays at the evaluation level, but also realizes an intelligent response chain of "identification-matching-execution". It automatically switches communication channels, avoids frequency bands, or initiates a safe return procedure according to the identified interference type, thereby verifying and enhancing the adaptability of UAVs in real complex environments during the testing process. Attached Figure Description

[0041] Figure 1 This is a system flowchart of the present invention. Detailed Implementation

[0042] The specific embodiments of the present invention will be further described below with reference to the accompanying drawings. It should be noted that these descriptions are for the purpose of aiding understanding the present invention, but do not constitute a limitation thereof. Furthermore, the technical features involved in the various embodiments of the present invention described below can be combined with each other as long as they do not conflict with each other.

[0043] Please see Figure 1 This invention provides a technical solution: a UAV performance testing system and method, comprising;

[0044] Core performance parameters are collected using a laser positioning system, RTK module, IMU, thrust test bench, battery capacity tester, and dedicated payload testing equipment (such as resolution card and point cloud analyzer). This includes data on hovering deviation, maximum speed / acceleration, endurance, dynamic positioning deviation, maximum thrust of the power system, response delay, battery capacity and low-temperature degradation rate, fault switching time, and maximum payload mass, resolution, frame rate, and other operational indicators. Environmental adaptability parameters are collected using wind tunnels, high and low temperature humidity chambers, dust / salt spray test chambers, lidar, and EMI signal generators. This includes data on wind resistance speed, operating temperature and humidity range, dust and water resistance rating, terrain following deviation, communication error rate under electromagnetic interference, high-altitude power attenuation rate, sea surface positioning deviation, and salt spray corrosion rate. The corresponding data collected simultaneously are preprocessed.

[0045] Data preprocessing: The collected data is sorted according to the collection time, and corresponding items collected at the same time are processed. averaging the data and standard deviation The calculation, and the mean obtained from the calculation. and standard deviation Collect data fluctuation range for corresponding items The system is configured to compare the collected data for a given item with its fluctuation range, mark data outside the fluctuation range as outliers, and record the number of outliers. ,like If the collected data is abnormal, the data will be re-tested; if If outliers are removed, the mean of the remaining corresponding test data after outlier removal is calculated. The calculation, and the mean obtained from the calculation. As the corresponding data detected at the corresponding time;

[0046] Re-examine the corresponding data; if the comparison result is still negative... If the problem is detected, it is determined that the data acquisition device is malfunctioning, a device warning signal is generated, and the device warning signal is transmitted to the warning module.

[0047] After receiving the equipment warning signal, the warning module will issue a buzzer warning through the buzzer module of the test system and display "Acquisition equipment abnormal" on the display screen of the test system, so that the staff can perform timely maintenance operations on the equipment.

[0048] Core performance These are the basic capabilities of drones, covering flight performance. Powertrain performance Load performance Overall core performance score , , , These are the weighting coefficients;

[0049] Flight performance score Hovering accuracy score , and These are vertical hovering deviation and horizontal hovering deviation; maneuverability score. , and These are the actual maximum horizontal velocity and the actual maximum acceleration, respectively. and These are the reference speed and reference acceleration defined by drone type; battery life score. , and These are the actual battery life and the target battery life, respectively; navigation accuracy score. , For dynamic positioning deviation;

[0050] Powertrain performance score Power output score , and These are the actual maximum thrust and the target thrust, respectively. For dynamic response delay; energy system score , and These are the actual battery capacity and the designed capacity, respectively. Low-temperature discharge capacity decay rate; power redundancy score , This refers to the failover time.

[0051] Load performance score Payload score , and These are the actual maximum load and the target load, respectively; load performance score. , This represents the total number of load performance indicators. For the first The weighting coefficient of the load performance index, and The first The actual maximum load and target load of the load performance index.

[0052] Obtain historical data and score flight performance from the historical data. Powertrain performance score Load performance score The data is acquired and categorized according to test time to obtain several test group data sets. These test group data sets are then further categorized according to the task type tested, resulting in several task group data sets. Test group data sets within the same task group are retrieved. Flight performance scores are then calculated from the retrieved test group data. Powertrain performance score Load performance score For comparison, data with the same two types of data in the test group are grouped together, and the change in the other type of data in the same group is compared with the total core performance score. The corresponding changes are recorded to obtain the total core performance score caused by another type of data unit change. Change amount; using the same method to obtain the total core performance score when the three types of data undergo a unit change. Change data , , Then determine the target weight of the corresponding data item. , , ;

[0053] Let the training dataset be... , No. Feature vector of each sample The corresponding true weight label ,and , This represents the total number of samples. The feature dimension is defined; the initial prediction weights are set to the average weights of the dataset. Then the first The first iteration obtained Prediction weights for each sample During model iteration, residual calculations and decision tree fitting are performed for each iteration, and prediction weights are updated. This continues until the preset number of iterations is reached. Stop iterating when the validation set loss no longer decreases, and obtain the final prediction weights. The final trained model is a weighted sum of decision trees. ,in , The initial weights and the learning rate are... , For the first The first round A decision tree fitted with weight dimensions, whose output is the input features. Mapped leaf node scores; final weights And ensure that the sum of the three weights is 1.

[0054] Environmental adaptability The assessment evaluates the ability to maintain performance under complex environments, including meteorological conditions. Geographical environment Special environment Overall score for environmental adaptability , , , These are the weighting coefficients;

[0055] Meteorological environment score Wind resistance score , and These are the actual stable wind resistance speed and the design wind resistance speed, respectively; temperature and humidity adaptation score. , and These are the deviations between the actual operating temperature / humidity range and the design range; and the precipitation / dust protection score. , and These are the dustproof / waterproof ratings, respectively.

[0056] Geographical environment score Terrain adaptation score , Terrain variation with height; Electromagnetic immunity score , The bit error rate in communication under electromagnetic interference conditions;

[0057] Special Environment Score High-altitude adaptability score , Dynamic attenuation rate at an altitude of 5000m; marine adaptability score. , For sea surface positioning deviation, The corrosion rate is shown after the salt spray test.

[0058] The obtained interference signal parameters (frequency) ,strength Duration ) and UAV parameter status (bit error rate) Positioning error of navigation module Electromagnetic interference immunity score The signal is correlated based on timestamps. The interference type is determined by the signal's time-domain waveform. The signal frequency is matched with the frequency bands of common interference sources to identify the interference frequency band. Each interference parameter is then analyzed. degree of influence , Representing variables and Pearson correlation coefficient, , ;

[0059] Building a dataset Interference feature vector Includes the lowest frequency of the interfering signal, the equivalent strength of the interfering signal, the pulse characteristics of the interfering signal, and the interference source label. Includes WiFi interference, base station interference, and GPS suppression interference; Build a training model: Constraints: , ,in The normal vector of the classification hyperplane, To adjust the bias term at the classification boundary position, As slack variables, To control the severity of classification, a penalty coefficient is applied; the constructed training model is then solved to obtain... and The optimal value is then the final model after training: , To collect interference features in real time, by... Inputting the data into the model can accurately predict the type of interference source. ;

[0060] If the interference source is determined to be WiFi interference, the drone's communication module integrates spectrum analysis functionality to scan the spectrum at set time intervals and dynamically switches to the channel with the least interference based on the signal-to-noise ratio or bit error rate. If the interference source is determined to be base station interference, the drone's radio module monitors spectrum occupancy in real time and dynamically avoids the concentrated frequency bands of base stations for communication. If the interference source is determined to be GPS suppression interference, the drone's flight control system integrates "safe mode" logic to automatically execute return-to-home or hover commands and activate preset safety strategies when interference continuously exceeds a threshold, preventing the drone from going out of control.

[0061] The embodiments of the present invention have been described in detail above with reference to the accompanying drawings, but the present invention is not limited to the described embodiments. For those skilled in the art, various changes, modifications, substitutions, and variations can be made to these embodiments without departing from the principles and spirit of the present invention, and these variations still fall within the protection scope of the present invention.

Claims

1. A UAV performance testing system, characterized by: It includes a preprocessing module, a performance evaluation module, and an environment adaptation module; The preprocessing module preprocesses the collected data, determines the fluctuation range, identifies outliers based on the fluctuation range, removes outliers, and improves the accuracy of the collected data. The performance evaluation module categorizes core performance based on sub-performance categories that affect core performance, calculates scores for each sub-performance category, and obtains the total core performance score by weighted summation of the sub-performance scores. It also acquires historical data, analyzes the sub-performance score data within the historical data, and determines the target weight coefficients for the corresponding sub-performance scores. Finally, it builds a model, trains it on historical data based on the target weight coefficients, obtains multiple training models, and dynamically adjusts the corresponding weight coefficients for the sub-performance scores based on the final trained model. The environment adaptation module categorizes environment adaptability based on the environmental categories that affect it, calculates scores for each environmental category, and obtains the total environment adaptability score by weighted summation of the environmental category scores. It acquires test data, analyzes the environmental category scores in the test data, identifies the type of interference source by combining interference signal characteristics, and correlates them with environment adaptability performance. Based on the type of interference source, it matches preset response strategies and dynamically adjusts the UAV's working mode or communication strategy under corresponding environmental conditions to improve its performance maintenance capability in complex environments.

2. The UAV performance testing system of claim 1, wherein: The preprocessing module performs the following data preprocessing steps: E1: sorting the collected data according to the collection time, calculating the mean value and standard deviation of the corresponding data collected at the same time, and setting the fluctuation range of the corresponding data collected according to the calculated mean value and standard deviation, and comparing the collected data of the corresponding items with the fluctuation range of the corresponding items; ;​​​​​ E2: mark the corresponding item data not in the fluctuation range as an abnormal value, and record the number of abnormal values , if , it is determined that the collected data is abnormal, and the detection of data is re-performed; if , the abnormal value is eliminated, the mean of the detection data of the corresponding item remaining after the elimination of the abnormal value is calculated, and the calculated mean is taken as the detection data of the corresponding item at the corresponding moment.

3. The UAV performance testing system of claim 1, wherein: The steps for analyzing the core performance score in the performance evaluation module are as follows: G1: total score of core performance , , , is a weight coefficient; G2: flight performance score , hovering accuracy score , and are respectively vertical and horizontal hovering deviation; maneuverability score , and are respectively actual maximum horizontal speed and actual maximum acceleration, and are respectively reference speed and reference acceleration defined by the drone type, is the gravity acceleration; endurance score , and are respectively actual endurance time and target endurance time; navigation accuracy score , is the dynamic positioning deviation; G3: Power system performance score G4: Power output score , G5: Power response delay , G6: Energy system score , G7: Low temperature discharge capacity decay rate , G8: Power redundancy score , G9: Fault switchover time G4: load performance score , load performance score , and are the actual maximum load and target load, respectively; load performance score , is the total number of load performance indicators, is the weight coefficient of the th load performance indicator, and are the actual maximum load and target load of the th load performance indicator, respectively.

4. The UAV performance testing system according to claim 3, characterized in that: The performance evaluation module performs the target weight analysis as follows: H1: Obtain historical data and score flight performance from the historical data. Powertrain performance score Load performance score The data is acquired and categorized according to the test time to obtain several test group data. The test group data is then categorized again according to the task type tested in each test group to obtain several task group data. Test group data within the same task group is then retrieved. Flight performance scores were calculated based on the retrieved test group data. Powertrain performance score Load performance score For comparison, data with the same two types of data in the test group are grouped together, and the change in the other type of data in the same group is compared with the total core performance score. The corresponding changes are recorded to obtain the total core performance score caused by another type of data unit change. Change; H2: The overall core performance score when the three types of data undergo unit changes using the same method. Change data , , Then determine the target weight of the corresponding data item. , , .

5. The UAV performance testing system according to claim 4, characterized in that: The analysis steps for the performance evaluation module to predict the final weights are as follows: J1: Based on total sample size The weighting coefficients corresponding to the sub-performance scores are set to the average weights of the dataset for initial prediction. Then the first The first iteration obtained Prediction weights for each sample During model iteration, residual calculations and decision tree fitting are performed for each iteration, and prediction weights are updated. This continues until the preset number of iterations is reached. Stop iterating when the validation set loss no longer decreases, and obtain the final prediction weights. ; J2: The final trained model is a weighted sum of decision trees. ,in , The initial weights and the learning rate are... , For the first The first round A decision tree fitted with weight dimensions, whose output is the input features. Mapped leaf node scores; final weights And ensure that the sum of the three weights is 1.

6. The UAV performance testing system according to claim 1, characterized in that: The steps for analyzing the overall environment adaptability score of the environment adaptation module are as follows: K1: Overall score for environmental adaptability , , , These are the weighting coefficients; K2: Meteorological Environment Score Wind resistance score , and These are the actual stable wind resistance speed and the design wind resistance speed, respectively; temperature and humidity adaptation score. , and These are the deviations between the actual operating temperature / humidity range and the design range; and the precipitation / dust protection score. , and These are the dustproof / waterproof ratings, respectively. K3: Geographical Environment Score Terrain adaptation score , Terrain variation with height; Electromagnetic interference immunity score , The bit error rate in communication under electromagnetic interference conditions; K4: Special Environment Score High-altitude adaptability score , Dynamic attenuation rate at an altitude of 5000m; marine adaptability score. , For sea surface positioning deviation, The corrosion rate is shown after the salt spray test.

7. The UAV performance testing system according to claim 6, characterized in that: The analysis steps for the environment adaptation module to predict interference source types are as follows: L1: Building the dataset , For interference feature vectors, Label the interference sources; construct the training model: Constraints: , ,in The normal vector of the classification hyperplane, To adjust the bias term at the classification boundary position, As slack variables, A penalty coefficient to control the strictness of classification; L2: Solve the constructed trained model to obtain... and The optimal value is then the final model after training: , To collect interference features in real time, by... Inputting the data into the model can accurately predict the type of interference source. .

8. The UAV performance testing system according to claim 7, characterized in that: The environmental adaptation module takes the following countermeasures based on the predicted interference source type: Q1: If the interference source is determined to be WiFi interference, the spectrum analysis function is integrated into the communication module of the drone. The spectrum is scanned at set time intervals, and the communication is dynamically switched to the channel with the least interference based on the signal-to-noise ratio or bit error rate. Q2: If the interference source is determined to be base station interference, the radio module of the drone can be used to monitor the spectrum occupancy in real time and dynamically avoid the concentrated frequency bands of the base station for communication. Q3: If the interference source is determined to be GPS suppression interference, the "safe mode" logic is integrated into the drone's flight control system. When the interference continues to exceed the threshold, the drone will automatically execute a return-to-home or hovering command and activate a preset safety strategy to prevent the drone from going out of control.

9. A method for testing the performance of a drone, using the drone performance testing system according to any one of claims 1-8, characterized in that: The drone performance testing method includes the following steps: U1: After aligning all collected data by timestamp, the system performs statistical analysis on multiple sets of measurements of the same parameter at the same time. First, it calculates the mean and standard deviation of the data set. Then, based on the "three-times-standard-deviation principle," it determines the reasonable fluctuation range of the data. Values ​​exceeding this range are marked as outliers. If the proportion of outliers exceeds 30%, the system determines that the data acquisition process is abnormal and automatically triggers a retest mechanism. If the retest result is still abnormal, the system determines that the acquisition equipment is faulty, automatically triggers an audible and visual alarm, and displays "Acquisition equipment abnormal" on the monitoring interface, guiding staff to perform timely repairs. U2: Core performance is broken down into three main categories: flight performance, propulsion system performance, and payload performance. Each category is further subdivided into multiple sub-indicators. For each sub-indicator, clear scoring rules are set, and the performance scores for each category are obtained through weighted summation. Historical test data is retrieved and grouped for analysis according to mission type. By statistically analyzing the impact of various performance indicators on overall performance in different missions, the initial weights of each performance category are preliminarily determined. A weight prediction model is constructed and trained using the mapping relationship between mission characteristics and optimal weights in historical data, enabling the system to dynamically adjust performance weights based on the characteristics of the current test mission. U3: Real-time monitoring of electromagnetic signal characteristics in the environment, combined with the current communication error rate, positioning error and other status parameters of the UAV, to identify the type of interference source through a pre-trained classification model; once the interference type is identified, the system automatically matches and executes the preset response strategy.

Images