A method for experimental replay and deduction

Abstract

The present invention provides a test review and deduction method, which belongs to the field of test management technology. The method includes clarifying the test objectives, determining the test background according to the test objectives, proposing test hypotheses in combination with the test objectives and the test background, and building a virtual test environment based on the test hypotheses; conducting tests in the virtual test environment, determining potential impacts that affect the test results, predicting the test development based on the potential impacts, and dividing the test development into stages; determining corresponding review main objectives according to the stage division results, obtaining solutions to each review main objective based on historical review meetings, and deriving stage products of each stage by combining all solutions; conducting multiple deductions on the stage products, generating deduction reports based on the stage deduction results, achieving more accurate data analysis and dynamic adjustment, improving the accuracy of the test, improving the effectiveness of the improvement plan, and strengthening the continuous use of data.

Description

Technical Field

[0001] The present invention relates to the technical field of test management, and in particular to a test review and deduction method. Background Art

[0002] When conducting experiments, it is usually necessary to use complex and changeable environments to simulate real situations, which puts higher requirements on the collection and analysis of experimental data. With the improvement of data analysis capabilities, more and more organizations rely on historical data and predictive analysis to make operational decisions in order to improve experimental efficiency and accuracy. Traditional experimental design is often weak in the ability to build an experimental environment, and the means of processing and analyzing experimental data are relatively simple. At the same time, there is often a lack of systematic methods to combine the lessons learned from historical data during the review process.

[0003] Therefore, the present invention provides a test review and deduction method. Summary of the invention

[0004] The present invention provides a test review and deduction method, which simulates real scenarios by building a virtual test environment, conducts development forecasts and divides stages in combination with test data, obtains solutions from historical data according to the review goals of each stage, generates interim results, and generates deduction reports through repeated deduction and analysis, thereby achieving more accurate data analysis and dynamic adjustment, improving the accuracy of the test, improving the effectiveness of the improvement plan, and strengthening the continuous use of data.

[0005] The present invention provides a test replay deduction method, comprising:

[0006] Step 1: Clarify the test objectives, determine the test background according to the test objectives, propose test hypotheses based on the test objectives and test background, and build a virtual test environment based on the test hypotheses;

[0007] Step 2: Conducting experiments in a virtual test environment, determining potential impacts that may affect the test results, predicting the development of the experiment based on the potential impacts, and dividing the development of the experiment into stages;

[0008] Step 3: Determine the corresponding main review objectives based on the stage division results, obtain solutions to each main review objective based on historical review meetings, and combine all solutions to derive the stage products of each stage;

[0009] Step 4: Perform multiple deductions on the stage products to obtain stage deduction results, and generate a deduction report based on the stage deduction results.

[0010] The present invention provides a test review and deduction method, which clarifies the test objectives, determines the test background according to the test objectives, proposes a test hypothesis in combination with the test objectives and the test background, and constructs a virtual test environment based on the test hypothesis, including:

[0011] Analyze the current experiment, derive the experiment goal based on the experiment analysis results, set the success criteria based on the experiment goal, and identify the existing research in the related field based on the success criteria, and then determine the experiment background related to the current experiment based on the existing research;

[0012] According to the experimental objectives and experimental background, the independent variables, dependent variables and control variables are identified, and then the experimental hypothesis is derived;

[0013] Determine the key functions required for constructing a primary virtual environment based on the test hypothesis, obtain test challenges existing in the same type of test according to the test background, and construct the primary virtual environment based on the test challenges and the required key functions;

[0014] Using the experimental hypothesis, a virtual experiment is conducted in a primary virtual environment to obtain a primary experimental result, which is compared with the theoretical expectation of the current experiment to generate a first difference. At the same time, the primary experimental result is compared with the actual data corresponding to the current experiment to generate a second difference. According to the first difference and the second difference, the primary virtual environment is refined to obtain a virtual experimental environment.

[0015] The present invention provides a test replay deduction method, which determines the key functions required for constructing a primary virtual environment based on the test hypothesis, obtains the test challenges existing in the same type of test according to the test background, and constructs the primary virtual environment based on the test challenges and the required key functions, including:

[0016] Performing variable relationship analysis on each variable in the test hypothesis to obtain the variable relationship analysis result, combining the variable relationship analysis result with the test goal to simulate the causal relationship between each test hypothesis, and deriving the key functions to be determined for constructing the primary virtual environment based on the causal relationship;

[0017] Screening the key functions to be determined based on the interdisciplinary opinion table to obtain the key functions required for building a primary virtual environment;

[0018] Acquire similar experiments according to the experimental background, determine the experimental challenges mentioned in the research process of the similar experiments, classify the experimental challenges, and generate a challenge-category table according to the classification results;

[0019] A challenge-solution table is obtained from the same type of experiments based on the challenge-category table, and a primary virtual environment is constructed according to the challenge-category table, the challenge-solution table and the required key functions.

[0020] The present invention provides a test replay deduction method, which refines the primary virtual environment according to the first difference and the second difference to obtain a virtual test environment, including:

[0021] For the first difference, analyze the first cause of the difference, and determine the first refinement condition by combining the test hypothesis and the first cause;

[0022] For the second difference, identify the data deviation of the primary test results and determine the second refinement conditions;

[0023] Compare the first difference with the second difference to find the common factors affecting the experimental hypothesis;

[0024] The primary virtual environment is refined based on the first refinement condition, the second refinement condition and the common factor to obtain a virtual test environment.

[0025] The present invention provides a test replay deduction method, which conducts a test in a virtual test environment, determines potential impacts that affect test results, predicts the test development based on the potential impacts, and divides the test development into stages, including:

[0026] Running the first virtual test based on the virtual test environment, obtaining the first test result, and comparing the first test result with the relevant data of the first test under the same conditions to obtain a virtual comparison result, identifying the first potential impact of the test process based on the virtual comparison result, and at the same time, identifying the key variables that have a significant impact on the test result based on the virtual comparison result;

[0027] Set different adjustment conditions for key variables, predict the test development under different adjustment conditions in the virtual test environment, and then determine the second potential impact of different adjustment conditions;

[0028] Based on the first potential impact and the second potential impact, a second virtual test is run in the virtual test environment to obtain a second test result, and based on the second test result and relevant data of the second test under the same conditions, a forecast adjustment is made to the predicted test development situation;

[0029] According to the forecast adjustment results, the experimental development situation is divided into different stages and the stage division results are obtained.

[0030] The present invention provides a test review deduction method, which determines the corresponding review main objectives according to the stage division results, obtains the solution to each review main objective based on the historical review meeting, and obtains the stage product of each stage by integrating all the solutions, including:

[0031] Identify the corresponding stage problems based on the results of each stage division, and then determine the main goals of the review at the corresponding stage;

[0032] Collect historical review meetings related to the main review objectives, extract historical review strategies from the historical review meetings, match the historical review strategies with the main review objectives, and derive historical review strategies that are applicable to the main review objectives;

[0033] Determine the solution to each major goal of the review by integrating all historical review strategies, simulate the actual effect of each solution based on the virtual test environment, and adjust the solution based on the simulation results;

[0034] According to the adjusted solution, the specific products that should be produced in each stage are defined, and the specific products are screened based on the product function standards to obtain the stage products.

[0035] The present invention provides a test replay deduction method, which matches the replay history strategy with the replay main goal to obtain the replay history strategy suitable for the replay main goal, including:

[0036] ;in, Represents the matching function between the historical strategy of review and the main goal of review; A represents the historical strategy data set of review; B represents the main goal data set of review; Represents the weight factor of the main objective of the review of the i-th stage for the overall test review; The indicator function representing the main goal of the review and the historical strategy of the review in the i-th stage; A data vector representing the main objective of the i-th stage review; The norm of the data vector representing the main objective of the replay at the i-th stage; v represents the data vector of the replay historical strategy; Represents the norm of the historical strategy data vector; Represents the weight factor of the j-th historical review strategy;

[0037] Determine the historical review strategy based on the matching function.

[0038] The present invention provides a test replay deduction method, which performs multiple deductions on the stage products to obtain stage deduction results, and generates a deduction report based on the stage deduction results, including:

[0039] The stage products are deduced multiple times to obtain stage deduction results, an evaluation index vector of each stage product is determined based on the experimental hypothesis, and the effectiveness of the stage deduction results is evaluated according to the evaluation index vector, thereby obtaining the stage results;

[0040] Generate a simulation report based on the results of all stages.

[0041] Compared with the existing technology, the beneficial effects of the present application are as follows: by constructing a virtual test environment to simulate real scenarios, combining test data to make development forecasts and divide stages, according to the review goals of each stage, obtaining solutions from historical data, generating interim results, and generating deduction reports through repeated deduction and analysis, more accurate data analysis and dynamic adjustment are achieved, the accuracy of the test is improved, the effectiveness of the improvement plan is improved, and the continuous use of data is strengthened.

[0042] Other features and advantages of the present invention will be described in the following description, and partly become apparent from the description, or understood by practicing the present invention. The purpose and other advantages of the present invention can be realized and obtained by the structures particularly pointed out in the written description and the accompanying drawings.

[0043] The technical solution of the present invention is further described in detail below through the accompanying drawings and embodiments. BRIEF DESCRIPTION OF THE DRAWINGS

[0044] The accompanying drawings are used to provide a further understanding of the present invention and constitute a part of the specification. Together with the embodiments of the present invention, they are used to explain the present invention and do not constitute a limitation of the present invention. In the accompanying drawings:

[0045] Figure 1 It is a flowchart of an experimental review and deduction method provided by an embodiment of the present invention. DETAILED DESCRIPTION

[0046] The preferred embodiments of the present invention are described below in conjunction with the accompanying drawings. It should be understood that the preferred embodiments described herein are only used to illustrate and explain the present invention, and are not used to limit the present invention.

[0047] Embodiment 1:

[0048] The embodiment of the present invention provides a test replay deduction method, such as Figure 1 As shown, including:

[0049] Step 1: Clarify the test objectives, determine the test background according to the test objectives, propose test hypotheses based on the test objectives and test background, and build a virtual test environment based on the test hypotheses;

[0050] Step 2: Conducting experiments in a virtual test environment, determining potential impacts that may affect the test results, predicting the development of the experiment based on the potential impacts, and dividing the development of the experiment into stages;

[0051] Step 3: Determine the corresponding main review objectives based on the stage division results, obtain solutions to each main review objective based on historical review meetings, and combine all solutions to derive the stage products of each stage;

[0052] Step 4: Perform multiple deductions on the stage products to obtain stage deduction results, and generate a deduction report based on the stage deduction results.

[0053] In this embodiment, the test objective is a clear goal formed based on the test analysis results, that is, what results are expected to be achieved through the test, such as improving the flight stability of the UAV under strong wind conditions; the test background combines the content of existing research to describe the scientific background and social significance of the current test. For example, when performing missions, UAVs often encounter complex meteorological conditions and changeable terrain environments. These factors will affect the flight stability of the UAV. For example, strong winds, turbulence, and changeable gusts will cause the UAV to experience bumps, yaw, or roll, affecting the accuracy and safety of its mission execution. Therefore, improving the flight stability of UAVs in complex environments is an important research direction; the test hypothesis is a testable statement based on the test objective and test background. For example, by optimizing the flight control algorithm and increasing sensor data fusion, the flight stability of the UAV under strong wind conditions will be significantly improved.

[0054] In this embodiment, by analyzing the causal relationship between the independent variables, dependent variables and control variables in the experimental hypothesis, the key functions are effectively determined, and screened in combination with interdisciplinary opinions. At the same time, the challenges faced by similar experiments are studied, a challenge solution table is established, and a primary virtual environment is constructed. By analyzing the differences in the experimental data, the reasons affecting the experimental results are explored, the first difference and the second difference are analyzed, and the root causes of the differences and data deviations are identified. Based on these causes and deviations, detailed conditions are determined, and common influencing factors are found. By making corresponding adjustments to the primary virtual environment, the experimental settings are optimized and a more accurate virtual experimental environment is established.

[0055] In this embodiment, the potential impacts include a first potential impact and a second potential impact. The first potential impact is a factor that may affect the test results, which is identified based on the results of the first test. For example, the flight control algorithm does not respond fast enough to gusts. The second potential impact is a possible impact based on the prediction after adjustment of key variables, which reflects the changes that may occur in the second test. For example, the optimized flight control algorithm will improve the flight stability of the UAV under strong wind conditions.

[0056] In this embodiment, the historical review meeting is a review meeting related to the current experiment conducted in the past. These meetings bring together the team's lessons learned and strategies from previous experiments. For example, reviewing the review meeting records of previous drone flight stability experiments and analyzing the performance of different algorithms in different environments.

[0057] In this embodiment, the main goal of the review is to determine the core goal based on the stage problems, and is intended to effectively solve the specific problems identified. For example, the main goal of the review in the result verification stage is to verify that the optimized flight control algorithm and dynamics model can maintain good flight stability under various environmental conditions.

[0058] In this embodiment, the solution is a specific solution strategy and plan formulated for the main goal of the review, which is used to solve the identified stage problems. For example, the solution in the result verification stage is to conduct multiple verification tests in a virtual test environment to verify that the optimized flight control algorithm and dynamic model can maintain good flight stability under various environmental conditions.

[0059] In this embodiment, the stage product is the most effective and valuable result obtained after screening according to the functional standards of the specific product. For example, in the key variable adjustment stage, the stage product may include optimized flight control algorithm parameter settings, improved UAV dynamics model, etc.

[0060] In this embodiment, the stage simulation results are the specific results obtained when performing each stage simulation based on the experimental hypothesis, including comparing the experimental results under different conditions and analyzing the response of the experimental subjects to the intervention measures. For example, in the result verification stage, through multiple simulations of the flight stability of the UAV under different environmental conditions, it was found that environmental factors have a significant impact on flight stability. Under crosswind conditions, the attitude stability of the UAV will be reduced.

[0061] In this embodiment, the deduction report is a final report formed by integrating the results of each stage, including a summary of the test process, analysis of results and research conclusions, with the purpose of demonstrating the integrity of the research and the reliability of the results, including the test background and purpose, evaluation indicators and measurement results of each stage, summary and comparison of stage results, final conclusions and suggestions, and possible follow-up research directions. For example, the experiment aims to optimize the flight stability of the UAV under strong wind conditions; through three stages of experiments, the parameters of the flight control algorithm are optimized and the dynamics model is improved, which significantly improves the flight stability of the UAV; the final conclusion is to recommend the use of the optimized flight control algorithm and dynamics model to improve the flight performance of the UAV under strong wind conditions; subsequent research can further explore the flight stability of the UAV under different environmental conditions, develop more advanced flight control algorithms and dynamics models, and study multi-sensor fusion technology to improve the UAV's perception of environmental changes.

[0062] The working principle and beneficial effects of the above technical solution are: by building a virtual test environment to simulate real scenarios, combining test data to make development forecasts and divide stages, obtaining solutions from historical data according to the review goals of each stage, generating interim results, and generating deduction reports through repeated deduction and analysis, so as to achieve more accurate data analysis and dynamic adjustment, improve the accuracy of the test, improve the effectiveness of the improvement plan, and strengthen the continuous use of data.

[0063] Embodiment 2:

[0064] The embodiment of the present invention provides a test review and deduction method, which clarifies the test objectives, determines the test background according to the test objectives, proposes a test hypothesis in combination with the test objectives and the test background, and constructs a virtual test environment based on the test hypothesis, including:

[0065] Analyze the current experiment, derive the experiment goal based on the experiment analysis results, set the success criteria based on the experiment goal, and identify the existing research in the related field based on the success criteria, and then determine the experiment background related to the current experiment based on the existing research;

[0066] According to the experimental objectives and experimental background, the independent variables, dependent variables and control variables are identified, and then the experimental hypothesis is derived;

[0067] Determine the key functions required for constructing a primary virtual environment based on the test hypothesis, obtain test challenges existing in the same type of test according to the test background, and construct the primary virtual environment based on the test challenges and the required key functions;

[0068] Using the experimental hypothesis, a virtual experiment is conducted in a primary virtual environment to obtain a primary experimental result, which is compared with the theoretical expectation of the current experiment to generate a first difference. At the same time, the primary experimental result is compared with the actual data corresponding to the current experiment to generate a second difference. According to the first difference and the second difference, the primary virtual environment is refined to obtain a virtual experimental environment.

[0069] In this embodiment, the test analysis result is a summary of the preliminary test data. For example, the preliminary test data shows that the flight stability of the UAV under specific wind speed conditions is lower than expected.

[0070] In this embodiment, the success criterion is a criterion for measuring whether the test is successful, and is usually determined by quantitative or qualitative indicators. For example, if the test goal is to improve the flight stability of the UAV under strong wind conditions, the success criterion is that the UAV can maintain a stable posture under strong wind conditions, and the change range of the yaw angle and the roll angle does not exceed a specific threshold, for example, the yaw angle change range does not exceed 5 degrees, and the roll angle change range does not exceed 3 degrees.

[0071] In this embodiment, the existing research refers to the related research results in the same field, which are usually used to support or refute the assumptions of the current experiment. For example, relevant studies have shown that the flight stability of the UAV in complex environments can be improved by optimizing the flight control algorithm of the UAV and increasing sensor data fusion.

[0072] In this example, the test challenges are obstacles that may be encountered during the design and implementation of the test, such as how to accurately simulate the aerodynamic effects under strong wind conditions and how to process sensor data in real time.

[0073] In this embodiment, independent variables are factors that are manipulated or changed during the experiment, such as parameter adjustment of the drone flight control algorithm; dependent variables are results observed and measured during the experiment, such as the flight stability of the drone (the range of change in yaw angle and roll angle); controlled variables are factors that remain unchanged to ensure the reliability of the test results, such as wind speed, drone model, sensor configuration, etc.

[0074] In this embodiment, the key functions required are to build the functional characteristics required for the test to ensure the effectiveness of the test, such as simulating different wind speed conditions, drone dynamics, sensor data processing, etc.

[0075] In this embodiment, the primary virtual environment is a preliminary constructed test environment, which may be a simulation or computer environment, for example, a drone flight environment based on computer simulation that can simulate basic wind speed and sensor data.

[0076] In this embodiment, the primary test results are the output data of the preliminary test, which are used to evaluate the test hypotheses. For example, the simulation results show that in the primary virtual environment, the flight stability of the UAV under strong wind conditions is improved, but still does not meet expectations. The first difference is the difference between the primary test results and the theoretical expectations. For example, there is still a gap between the primary test results and the theoretical expectations, indicating that there is a large room for algorithm optimization. The second difference is the difference between the primary test results and the actual data. For example, compared with the actual data, the primary test results show that the flight stability is lower than the actual performance under certain specific wind speed conditions.

[0077] In this embodiment, the virtual test environment is a new environment generated based on the refinement of the primary virtual environment, which is usually more complex and closer to the real situation. For example, based on the refinement of the primary virtual environment, a more complex aerodynamic model and a more advanced sensor data processing algorithm are added.

[0078] The working principle and beneficial effects of the above technical solution are: by clarifying independent variables, dependent variables and control variables to form hypotheses, determining key functions based on existing research and challenges, running virtual experiments and obtaining results on this basis, comparing theoretical expectations and actual data generation differences, and thus optimizing and adjusting the virtual environment to improve the efficiency and accuracy of subsequent experiments, reduce manual construction time, and achieve rapid iteration.

[0079] Embodiment 3:

[0080] The embodiment of the present invention provides a test replay deduction method, which determines the key functions required for constructing a primary virtual environment based on the test hypothesis, obtains the test challenges existing in the same type of test according to the test background, and constructs the primary virtual environment based on the test challenges and the required key functions, including:

[0081] Performing variable relationship analysis on each variable in the test hypothesis to obtain the variable relationship analysis result, combining the variable relationship analysis result with the test goal to simulate the causal relationship between each test hypothesis, and deriving the key functions to be determined for constructing the primary virtual environment based on the causal relationship;

[0082] Screening the key functions to be determined based on the interdisciplinary opinion table to obtain the key functions required for building a primary virtual environment;

[0083] Acquire similar experiments according to the experimental background, determine the experimental challenges mentioned in the research process of the similar experiments, classify the experimental challenges, and generate a challenge-category table according to the classification results;

[0084] A challenge-solution table is obtained from the same type of experiments based on the challenge-category table, and a primary virtual environment is constructed according to the challenge-category table, the challenge-solution table and the required key functions.

[0085] In this embodiment, variable relationship analysis is an analysis of the interaction between the variables in the experimental hypothesis, with the purpose of exploring the relationship between independent variables, dependent variables and controlled variables. For example, it analyzes how the parameter settings of the drone flight control algorithm (independent variables) affect the flight stability of the drone (dependent variables), and considers the impact of environmental factors (controlled variables) such as wind speed, temperature, etc. on flight stability.

[0086] In this embodiment, causality describes the impact of one variable on another variable, and is an important analysis to determine how independent variables lead to changes in dependent variables. For example, it simulates the impact of different flight control algorithm parameter settings on the flight stability of the UAV under strong wind conditions, and how these parameters respond to environmental changes by changing the attitude and speed of the UAV.

[0087] In this embodiment, the key functions to be determined are the functions that must be possessed in the test environment after variable relationships and causal analysis, such as the UAV dynamics model, which is used to simulate the flight characteristics of the UAV; the flight control algorithm, which is used to adjust the attitude and speed of the UAV; and the environmental simulation, which is used to simulate the wind speed, temperature and other conditions in the actual flight environment.

[0088] In this embodiment, the interdisciplinary opinion table is a tool that collects opinions of experts from different disciplines and is used to screen and determine key functions. For example, the opinions of experts from different disciplines such as aerospace engineers, control theory experts, and meteorologists are collected to screen and optimize key functions.

[0089] In this embodiment, the challenge-category table systematically categorizes the challenges encountered during the test to manage these challenges, such as environmental simulation challenges: accurately simulating wind speed, temperature and other conditions in the actual flight environment, control algorithm challenges: developing flight control algorithms that can adapt to complex environments, and dynamic model challenges: building an accurate UAV dynamics model.

[0090] In this embodiment, the challenge-solution table provides a table of corresponding solutions for various challenges, which is convenient for quick reference when encountering challenges. For example, environmental simulation challenges: using advanced fluid dynamics models and meteorological data, control algorithm challenges: using adaptive control algorithms and machine learning technology, and dynamic model challenges: conducting detailed drone design and parameter identification.

[0091] The working principle and beneficial effects of the above technical solution are: by analyzing the causal relationship between independent variables, dependent variables and control variables in the experimental hypothesis, the key functions are effectively determined, and screened in combination with interdisciplinary opinions. At the same time, by studying the challenges faced by similar experiments, a challenge solution table is established, and a primary virtual environment is constructed to improve the flexibility and efficiency of the experiment, realize dynamic adjustment of data collection and analysis, and ultimately improve the accuracy and efficiency of the experiment, enhance the flexibility of the experimental plan, and adapt to different experimental conditions.

[0092] Embodiment 4:

[0093] An embodiment of the present invention provides a test replay deduction method, which refines a primary virtual environment according to the first difference and the second difference to obtain a virtual test environment, including:

[0094] For the first difference, analyze the first cause of the difference, and determine the first refinement condition by combining the test hypothesis and the first cause;

[0095] For the second difference, identify the data deviation of the primary test results and determine the second refinement conditions;

[0096] Compare the first difference with the second difference to find the common factors affecting the experimental hypothesis;

[0097] The primary virtual environment is refined based on the first refinement condition, the second refinement condition and the common factor to obtain a virtual test environment.

[0098] In this embodiment, the first cause refers to the fundamental factor that can directly lead to the first difference when analyzing the first difference, and is the primary cause affecting the test results. The first refinement condition is an optimization condition determined based on the analysis result of the first difference and the test hypothesis. For example, the first difference is that the flight stability of the UAV under strong wind conditions in the primary test results is lower than expected; the first cause is that the flight control algorithm does not respond quickly enough to gusts, resulting in excessive yaw of the UAV when encountering strong winds; the first refinement condition is to optimize the parameters of the flight control algorithm to improve its response speed and stability to gusts.

[0099] In this embodiment, data deviation refers to the systematic difference between the actual observed value and the true value during the process of collecting and analyzing test data. The second refined condition is an optimization condition formulated for the second difference (based on the deviation between the actual data and the primary test results) to reduce or eliminate data deviation and improve data reliability. For example, the second difference is that compared with the actual flight data, the primary test results show that the attitude stability of the drone at a specific wind speed is poor; the data deviation is that the actual flight data contains more environmental variables and sensor noise, which are not fully simulated in the primary test; the second refined condition is to add more complex environmental simulations in the virtual environment, such as random wind speed changes, sensor noise simulation, etc.

[0100] In this embodiment, the common factors are the common factors that affect the test hypothesis identified when comparing the first difference and the second difference. They are usually those variables or influencing factors that exist in both conditions. For example, when analyzing the first difference and the second difference, it is found that the dynamic model of the drone may be overly simplified and fails to fully consider the nonlinear dynamic characteristics in actual flight.

[0101] In this embodiment, the specific content of the refinement is the specific measures to optimize the experimental design according to the first refinement condition and the second refinement condition, including technical, methodological or operational adjustments, such as using a more advanced dynamic model, considering the nonlinear dynamic characteristics of the UAV, and increasing the complexity of the environmental simulation, including variable gust patterns, temperature gradients, etc.

[0102] The working principle and beneficial effects of the above technical solution are: by analyzing the differences in the test data, exploring the reasons that affect the test results, analyzing the first difference and the second difference, identifying the root causes of the differences and data deviations, and determining the detailed conditions based on these reasons and deviations, and finding common influencing factors, by making corresponding adjustments to the primary virtual environment, optimizing the test settings, and establishing a more accurate virtual test environment, thereby improving the reliability and predictive ability of the test, achieving real-time feedback and adjustment of the test conditions, being able to cope with ever-changing test scenarios, and improving the effectiveness of the test results.

[0103] Embodiment 5:

[0104] The embodiment of the present invention provides a test replay deduction method, which conducts a test in a virtual test environment, determines potential impacts that affect test results, predicts the test development based on the potential impacts, and divides the test development into stages, including:

[0105] Running the first virtual test based on the virtual test environment, obtaining the first test result, and comparing the first test result with the relevant data of the first test under the same conditions to obtain a virtual comparison result, identifying the first potential impact of the test process based on the virtual comparison result, and at the same time, identifying the key variables that have a significant impact on the test result based on the virtual comparison result;

[0106] Set different adjustment conditions for key variables, predict the test development under different adjustment conditions in the virtual test environment, and then determine the second potential impact of different adjustment conditions;

[0107] Based on the first potential impact and the second potential impact, a second virtual test is run in the virtual test environment to obtain a second test result, and based on the second test result and relevant data of the second test under the same conditions, a forecast adjustment is made to the predicted test development situation;

[0108] According to the forecast adjustment results, the experimental development situation is divided into different stages and the stage division results are obtained.

[0109] In this embodiment, the first test result is the result data obtained after running the initial test in the virtual test environment. For example, in the virtual test environment, the flight stability test results of the UAV under simulated strong wind conditions show that when the UAV encounters a gust of wind, the change amplitude of its yaw angle and roll angle exceeds the expected threshold, indicating that the flight stability needs to be improved.

[0110] In this embodiment, the virtual comparison result is the result obtained by comparing the first virtual test result with the real test data under the same conditions. For example, by comparing the first virtual test result with the real test data under the same conditions, it is found that the virtual test result has a certain deviation from the actual data in terms of the attitude stability of the UAV. For example, the yaw angle change amplitude of the UAV under strong wind conditions in the virtual test is smaller than the change amplitude in the actual flight data.

[0111] In this embodiment, key variables are variables that are identified during the test as potentially having a significant impact on the results. For example, during the test, key variables may include parameter settings of the flight control algorithm (such as control gain, filter parameters, etc.), the dynamics model of the drone (such as mass distribution, aerodynamic characteristics, etc.), and the complexity of the environmental simulation (such as wind speed, turbulence intensity, etc.).

[0112] In this embodiment, the second test results are the results obtained by running the second test in a virtual test environment. These results are used to evaluate the effects of adjusting key variables. For example, the second test was conducted in a virtual test environment. In this test, the parameter settings of the flight control algorithm were adjusted to optimize the response speed of the UAV to gusts. The results showed that when the adjusted UAV encountered gusts, the changes in its yaw angle and roll angle were significantly reduced, and were closer to the expected threshold.

[0113] In this embodiment, the specific process of predictive adjustment is an adjustment process based on the first test results and the second potential impact, including identifying the test conditions that need to be modified and predicting the impact of these modifications on the results. For example, the predictive adjustment may include optimizing the parameter settings of the flight control algorithm to reduce the variation of the yaw angle and roll angle of the drone under strong wind conditions.

[0114] In this embodiment, the difference between the first test results and the second test results is that the difference between the first test results and the second test results is reflected in the change of the test data after the key variables are adjusted. For example, the first test results show that the yaw angle and roll angle of the drone under strong wind conditions have a large change range, while the second test results show that these changes are significantly reduced.

[0115] In this embodiment, the difference between the second potential impact and the first potential impact is that the first potential impact is derived from the identification factors obtained from the first test results, while the second potential impact is the expected result obtained based on the prediction model after adjusting the key variables. For example, the first potential impact is that the flight control algorithm does not respond fast enough to gusts, while the second potential impact is that the optimized flight control algorithm will improve the flight stability of the UAV under strong wind conditions.

[0116] The working principle and beneficial effects of the above technical solution are: by conducting multiple rounds of simulation tests in a virtual test environment, the first virtual test is run first and the results are obtained. On this basis, the results are compared with the actual data to identify potential influences and key variables. By setting adjustment conditions for key variables, their impact on the test results is predicted, and a second virtual test is conducted. According to the comparison and predicted adjustment results of the two tests, the test development is divided into stages, thereby optimizing the test design and result prediction, improving the accuracy and reliability of the test, and helping researchers to identify potential problems in a timely manner.

[0117] Embodiment 6:

[0118] The embodiment of the present invention provides a test review deduction method, which determines the corresponding review main goal according to the stage division result, obtains the solution to each review main goal based on the historical review meeting, and obtains the stage product of each stage by integrating all the solutions, including:

[0119] Identify the corresponding stage problems based on the results of each stage division, and then determine the main goals of the review at the corresponding stage;

[0120] Collect historical review meetings related to the main review objectives, extract historical review strategies from the historical review meetings, match the historical review strategies with the main review objectives, and derive historical review strategies that are applicable to the main review objectives;

[0121] Determine the solution to each major goal of the review by integrating all historical review strategies, simulate the actual effect of each solution based on the virtual test environment, and adjust the solution based on the simulation results;

[0122] According to the adjusted solution, the specific products that should be produced in each stage are defined, and the specific products are screened based on the product function standards to obtain the stage products.

[0123] In this embodiment, the stage problems are specific problems identified in each stage, which need to be analyzed and solved in subsequent reviews. For example, in the preliminary test stage, the UAV’s flight stability in strong wind conditions is insufficient.

[0124] In this embodiment, the historical review strategy is a strategy extracted from historical review meetings, which is usually an improvement measure for past failures or shortcomings. For example, the historical review strategy in the result verification phase can be to conduct multiple verification tests in a virtual test environment to verify that the optimized flight control algorithm and dynamics model can maintain good flight stability under various environmental conditions, and collect more actual flight data to further verify and improve the model.

[0125] In this embodiment, adjusting the solution is to optimize and adjust the solution after simulating the actual effect of the solution in a virtual test environment, for example, adjusting the parameters of the sensor data processing algorithm to improve the attitude stability of the drone under strong wind conditions.

[0126] In this embodiment, the specific products are the specific achievements, documents or data sets that should be produced in each stage. These products are the results of achieving the main goals of the review. For example, in the result verification stage, the corresponding specific products include the final test report: including the results, analysis and conclusions of all test stages; flight stability evaluation results: including the performance data of the UAV under various environmental conditions; verification report of the flight control algorithm and dynamic model: proving the reliability and effectiveness of the algorithm and model.

[0127] The working principle and beneficial effects of the above technical solution are: by gradually identifying the problems in the results of each stage division, determining the corresponding main goals of the review, and extracting relevant strategies from historical review meetings, matching the extracted historical review strategies with the review goals to formulate appropriate solutions, simulating the actual effects of each solution in a virtual test environment, and making adjustments based on the simulation results, ultimately defining the specific products that should be produced in each stage, and screening out stage products that meet the requirements through functional standards, ensuring targeted solutions for each stage and improving the practicality of the review.

[0128] Embodiment 7:

[0129] The embodiment of the present invention provides a test replay deduction method, which matches the replay history strategy with the replay main goal to obtain the replay history strategy suitable for the replay main goal, including:

[0130] ;in, Represents the matching function between the historical strategy of review and the main goal of review; A represents the historical strategy data set of review; B represents the main goal data set of review; Represents the weight factor of the main objective of the review of the i-th stage for the overall test review; The indicator function representing the main goal of the review and the historical strategy of the review in the i-th stage; A data vector representing the main objective of the i-th stage review; The norm of the data vector representing the main objective of the replay at the i-th stage; v represents the data vector of the replay historical strategy; Represents the norm of the historical strategy data vector; Represents the weight factor of the j-th historical review strategy;

[0131] Determine the historical review strategy based on the matching function.

[0132] In this embodiment, the indicator function quantifies the relationship between the historical strategy and the target matching. For example, in the study of drone flight stability, set A represents the historical strategy data set, such as data under specific environment or control parameters, and set B represents the main target data set, such as the flight performance standard expected to be achieved. The indicator function is In the ith test, the drone was in a completely windless environment (belonging to set A). At the same time, the test showed good attitude maintenance ability and no flutter (this behavior met the target standard and also entered set B). The ith test met all requirements because the record was located in an ideal climate and no inappropriate track appeared, so I(i∈A∩B)=1.

[0133] The working principle and beneficial effects of the above technical solution are: utilizing the matching function between the review history strategy and the review goal, by quantifying the importance and relevant data of the review goals at different stages, evaluating their influence in the overall experiment, and thus determining the most appropriate review history strategy, improving the scientific nature of the selection of solutions to current problems, reducing the investment required for traditional manual construction and single processing methods, and accelerating the entire test verification process.

[0134] Embodiment 8:

[0135] The embodiment of the present invention provides a test replay deduction method, which performs multiple deductions on the stage products to obtain stage deduction results, and generates a deduction report based on the stage deduction results, including:

[0136] The stage products are deduced multiple times to obtain stage deduction results, an evaluation index vector of each stage product is determined based on the experimental hypothesis, and the effectiveness of the stage deduction results is evaluated according to the evaluation index vector, thereby obtaining the stage results;

[0137] Generate a simulation report based on the results of all stages.

[0138] In this embodiment, the evaluation index vector is a set of quantitative standards or indicators defined to judge the effectiveness of the stage product, including quantitative indicators: such as quantitative indicators: for example, the flight time of the UAV at a specific wind speed, the number of successful completion of tasks; quality indicators: for example, the attitude stability of the UAV under strong wind conditions (the change amplitude of the yaw angle and the roll angle), the smoothness of the flight trajectory; timeliness indicators: for example, the average time of the UAV to respond to gusts, the time required to complete the task; cost-effectiveness indicators: for example, the ratio of the UAV maintenance cost, upgrade cost and performance improvement. Therefore, the evaluation index vector can be: [flight time (minutes), attitude stability (degrees / second), response time (seconds), cost-effectiveness (ratio)].

[0139] In this embodiment, effectiveness evaluation is the process of analyzing the results of the trial phase to determine whether they have achieved the expected goals and standards, including significance tests (such as p-values), effect sizes, and comparisons of results with expectations, for example, the degree of improvement in the drone's attitude stability, and the comparison of the drone's actual flight stability under strong wind conditions with the expected stability.

[0140] In this embodiment, the stage results are the final conclusions and decisions for each test stage, which are usually based on the screening and refinement of the deduction results, including recommendations or conclusions after effectiveness evaluation, and confirmation of the best conditions and solutions to use. For example, after multiple verification tests, the optimized flight control algorithm and dynamic model can maintain good flight stability under various environmental conditions.

[0141] The working principle and beneficial effects of the above technical solution are: by conducting multiple deductions on the products of each stage, determining the evaluation index vector based on the experimental hypothesis, evaluating and screening the effectiveness of the deduction results, and finally forming a comprehensive stage result. All these steps are summarized to generate a detailed deduction report to support decision-making and experimental optimization, improve the overall experimental quality, help to adjust the strategy in a timely manner, and achieve flexible response to market changes.

[0142] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit it. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or make equivalent replacements for some of the technical features therein. However, these modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. A test replay deduction method, characterized in that: include: Step 1: Clarify the test objectives, determine the test background according to the test objectives, propose test hypotheses based on the test objectives and test background, and build a virtual test environment based on the test hypotheses; Step 2: Conducting experiments in a virtual test environment, determining potential impacts that may affect the test results, predicting the development of the experiment based on the potential impacts, and dividing the development of the experiment into stages; Step 3: Determine the corresponding main review objectives based on the stage division results, obtain solutions to each main review objective based on historical review meetings, and combine all solutions to derive the stage products of each stage; Step 4: Perform multiple deductions on the stage product to obtain stage deduction results, and generate a deduction report based on the stage deduction results; Among them, conducting the test in the virtual test environment, determining the potential impact that affects the test results, predicting the test development based on the potential impact, and dividing the test development into stages, including: Running the first virtual test based on the virtual test environment, obtaining the first test result, and comparing the first test result with the relevant data of the first test under the same conditions to obtain a virtual comparison result, identifying the first potential impact of the test process based on the virtual comparison result, and at the same time, identifying the key variables that have a significant impact on the test result based on the virtual comparison result; Set different adjustment conditions for key variables, predict the test development under different adjustment conditions in the virtual test environment, and then determine the second potential impact of different adjustment conditions; Based on the first potential impact and the second potential impact, a second virtual test is run in the virtual test environment to obtain a second test result, and based on the second test result and relevant data of the second test under the same conditions, a forecast adjustment is made to the predicted test development situation; According to the forecast adjustment results, the experimental development situation is divided into different stages and the stage division results are obtained.

2. A test replay deduction method according to claim 1, characterized in that: Clarify the test objectives, determine the test background according to the test objectives, propose test hypotheses based on the test objectives and test background, and build a virtual test environment based on the test hypotheses, including: Analyze the current experiment, derive the experiment goal based on the experiment analysis results, set the success criteria based on the experiment goal, and identify the existing research in the related field based on the success criteria, and then determine the experiment background related to the current experiment based on the existing research; According to the experimental objectives and experimental background, the independent variables, dependent variables and control variables are identified, and then the experimental hypothesis is derived; Determine the key functions required for constructing a primary virtual environment based on the test hypothesis, obtain test challenges existing in the same type of test according to the test background, and construct the primary virtual environment based on the test challenges and the required key functions; Using the experimental hypothesis, a virtual experiment is conducted in a primary virtual environment to obtain a primary experimental result, which is compared with the theoretical expectation of the current experiment to generate a first difference. At the same time, the primary experimental result is compared with the actual data corresponding to the current experiment to generate a second difference. According to the first difference and the second difference, the primary virtual environment is refined to obtain a virtual experimental environment.

3. A test replay deduction method according to claim 2, characterized in that: Determining the key functions required for constructing a primary virtual environment based on the test hypothesis, obtaining test challenges existing in the same type of test according to the test background, and constructing a primary virtual environment based on the test challenges and the required key functions, including: Performing variable relationship analysis on each variable in the test hypothesis to obtain the variable relationship analysis result, combining the variable relationship analysis result with the test goal to simulate the causal relationship between each test hypothesis, and deriving the key functions to be determined for constructing the primary virtual environment based on the causal relationship; Screening the key functions to be determined based on the interdisciplinary opinion table to obtain the key functions required for building a primary virtual environment; Acquire similar experiments according to the experimental background, determine the experimental challenges mentioned in the research process of the similar experiments, classify the experimental challenges, and generate a challenge-category table according to the classification results; A challenge-solution table is obtained from the same type of experiments based on the challenge-category table, and a primary virtual environment is constructed according to the challenge-category table, the challenge-solution table and the required key functions.

4. A test replay deduction method according to claim 2, characterized in that: The primary virtual environment is refined according to the first difference and the second difference to obtain a virtual test environment, including: For the first difference, analyze the first cause of the difference, and determine the first refinement condition by combining the test hypothesis and the first cause; For the second difference, identify the data deviation of the primary test results and determine the second refinement conditions; Compare the first difference with the second difference to find the common factors affecting the experimental hypothesis; The primary virtual environment is refined based on the first refinement condition, the second refinement condition and the common factor to obtain a virtual test environment.

5. The test replay deduction method according to claim 1, characterized in that: Determine the corresponding main review objectives based on the stage division results, obtain solutions to each main review objective based on historical review meetings, and combine all solutions to derive the stage products of each stage, including: Identify the corresponding stage problems based on the results of each stage division, and then determine the main goals of the review at the corresponding stage; Collect historical review meetings related to the main review objectives, extract historical review strategies from the historical review meetings, match the historical review strategies with the main review objectives, and derive historical review strategies that are applicable to the main review objectives; Determine the solution to each major goal of the review by integrating all historical review strategies, simulate the actual effect of each solution based on the virtual test environment, and adjust the solution based on the simulation results; According to the adjusted solution, the specific products that should be produced in each stage are defined, and the specific products are screened based on the product function standards to obtain the stage products.

6. A test replay deduction method according to claim 5, characterized in that: The historical review strategy is matched with the main review goal to obtain a historical review strategy suitable for the main review goal, including: ;in, Represents the matching function between the historical strategy of review and the main goal of review; A represents the historical strategy data set of review; B represents the main goal data set of review; Represents the weight factor of the main objective of the review of the i-th stage for the overall test review; The indicator function representing the main goal of the review and the historical strategy of the review in the i-th stage; A data vector representing the main objective of the i-th stage review; The norm of the data vector representing the main objective of the replay at the i-th stage; v represents the data vector of the replay historical strategy; Represents the norm of the historical strategy data vector; Represents the weight factor of the j-th historical review strategy; Determine the historical review strategy based on the matching function.

7. The test replay deduction method according to claim 1, characterized in that: The stage product is deduced multiple times to obtain stage deduction results, and a deduction report is generated based on the stage deduction results, including: The stage products are deduced multiple times to obtain stage deduction results, an evaluation index vector of each stage product is determined based on the experimental hypothesis, and the effectiveness of the stage deduction results is evaluated according to the evaluation index vector, thereby obtaining the stage results; Generate a simulation report based on the results of all stages.

Images