CN117633722B - Detection control method and system based on intelligent detection robot - Google Patents
Detection control method and system based on intelligent detection robot Download PDFInfo
- Publication number
- CN117633722B CN117633722B CN202410106551.XA CN202410106551A CN117633722B CN 117633722 B CN117633722 B CN 117633722B CN 202410106551 A CN202410106551 A CN 202410106551A CN 117633722 B CN117633722 B CN 117633722B
- Authority
- CN
- China
- Prior art keywords
- data
- test
- calibration
- result
- algorithm
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Active
Links
- 238000000034 method Methods 0.000 title claims abstract description 132
- 238000001514 detection method Methods 0.000 title claims abstract description 70
- 238000012360 testing method Methods 0.000 claims abstract description 391
- 238000004422 calculation algorithm Methods 0.000 claims abstract description 153
- 238000004458 analytical method Methods 0.000 claims abstract description 103
- 230000002159 abnormal effect Effects 0.000 claims abstract description 97
- 230000007613 environmental effect Effects 0.000 claims abstract description 54
- 238000005457 optimization Methods 0.000 claims abstract description 49
- 238000013528 artificial neural network Methods 0.000 claims abstract description 35
- 230000004927 fusion Effects 0.000 claims abstract description 26
- 238000005516 engineering process Methods 0.000 claims abstract description 24
- 238000012417 linear regression Methods 0.000 claims abstract description 20
- 230000006978 adaptation Effects 0.000 claims abstract description 16
- 238000012549 training Methods 0.000 claims abstract description 13
- 238000010801 machine learning Methods 0.000 claims abstract description 10
- 230000008569 process Effects 0.000 claims description 57
- 238000012706 support-vector machine Methods 0.000 claims description 56
- 238000010998 test method Methods 0.000 claims description 37
- 230000008859 change Effects 0.000 claims description 34
- 230000003044 adaptive effect Effects 0.000 claims description 27
- 239000000463 material Substances 0.000 claims description 22
- 238000012545 processing Methods 0.000 claims description 20
- 238000013433 optimization analysis Methods 0.000 claims description 19
- 238000000605 extraction Methods 0.000 claims description 18
- 238000003066 decision tree Methods 0.000 claims description 17
- 238000013135 deep learning Methods 0.000 claims description 14
- 238000000513 principal component analysis Methods 0.000 claims description 14
- 238000009826 distribution Methods 0.000 claims description 13
- 230000000694 effects Effects 0.000 claims description 13
- 230000010354 integration Effects 0.000 claims description 13
- 230000009467 reduction Effects 0.000 claims description 11
- 238000012544 monitoring process Methods 0.000 claims description 9
- 238000012098 association analyses Methods 0.000 claims description 8
- 230000008901 benefit Effects 0.000 claims description 8
- 230000004044 response Effects 0.000 claims description 8
- 238000011156 evaluation Methods 0.000 claims description 7
- 230000002068 genetic effect Effects 0.000 claims description 7
- 238000003064 k means clustering Methods 0.000 claims description 7
- 238000007726 management method Methods 0.000 claims description 7
- 230000003993 interaction Effects 0.000 claims description 6
- 238000000611 regression analysis Methods 0.000 claims description 6
- 238000003324 Six Sigma (6σ) Methods 0.000 claims description 5
- 230000006399 behavior Effects 0.000 claims description 5
- 238000010276 construction Methods 0.000 claims description 5
- 230000006872 improvement Effects 0.000 claims description 5
- 238000013468 resource allocation Methods 0.000 claims description 5
- 238000012300 Sequence Analysis Methods 0.000 claims description 4
- 230000008303 genetic mechanism Effects 0.000 claims description 4
- 238000005065 mining Methods 0.000 claims description 4
- 230000008030 elimination Effects 0.000 claims description 2
- 238000003379 elimination reaction Methods 0.000 claims description 2
- 238000007405 data analysis Methods 0.000 abstract description 4
- 239000013598 vector Substances 0.000 abstract description 4
- 239000000126 substance Substances 0.000 description 8
- 238000004519 manufacturing process Methods 0.000 description 6
- 239000002131 composite material Substances 0.000 description 4
- 239000002184 metal Substances 0.000 description 4
- 229910052751 metal Inorganic materials 0.000 description 4
- 238000013136 deep learning model Methods 0.000 description 3
- 230000002708 enhancing effect Effects 0.000 description 3
- 230000006870 function Effects 0.000 description 3
- 238000003062 neural network model Methods 0.000 description 3
- 230000009286 beneficial effect Effects 0.000 description 2
- 238000010219 correlation analysis Methods 0.000 description 2
- 230000007547 defect Effects 0.000 description 2
- 238000010586 diagram Methods 0.000 description 2
- 238000013100 final test Methods 0.000 description 2
- 238000007689 inspection Methods 0.000 description 2
- 230000007774 longterm Effects 0.000 description 2
- 230000004048 modification Effects 0.000 description 2
- 238000012986 modification Methods 0.000 description 2
- 238000012502 risk assessment Methods 0.000 description 2
- 230000004913 activation Effects 0.000 description 1
- 238000000540 analysis of variance Methods 0.000 description 1
- 230000001364 causal effect Effects 0.000 description 1
- 238000010224 classification analysis Methods 0.000 description 1
- 238000011217 control strategy Methods 0.000 description 1
- 238000012937 correction Methods 0.000 description 1
- 230000003247 decreasing effect Effects 0.000 description 1
- 239000000284 extract Substances 0.000 description 1
- 238000011990 functional testing Methods 0.000 description 1
- 238000013178 mathematical model Methods 0.000 description 1
- 150000002739 metals Chemical class 0.000 description 1
- 239000000203 mixture Substances 0.000 description 1
- 210000002569 neuron Anatomy 0.000 description 1
- 238000003909 pattern recognition Methods 0.000 description 1
- 238000011056 performance test Methods 0.000 description 1
- 230000008092 positive effect Effects 0.000 description 1
- 230000002035 prolonged effect Effects 0.000 description 1
- 230000003014 reinforcing effect Effects 0.000 description 1
- 238000010845 search algorithm Methods 0.000 description 1
- 230000001932 seasonal effect Effects 0.000 description 1
- 238000004088 simulation Methods 0.000 description 1
- 239000007787 solid Substances 0.000 description 1
- 238000013179 statistical model Methods 0.000 description 1
- 230000009897 systematic effect Effects 0.000 description 1
- 238000012731 temporal analysis Methods 0.000 description 1
- 238000012422 test repetition Methods 0.000 description 1
- 238000000700 time series analysis Methods 0.000 description 1
Landscapes
- Feedback Control In General (AREA)
Abstract
The invention relates to the technical field of automatic testing, in particular to a detection control method and system based on an intelligent detection robot, comprising the following steps: based on intelligent detection robot data, a linear regression algorithm is adopted to analyze data trend, a deviation analysis method is adopted to compare the data with a standard value in real time, data deviation is identified and quantized, and the data is calibrated and adjusted through a self-adaptive filter, so that noise is reduced, errors are eliminated, and calibration data is generated. According to the invention, the data analysis accuracy is improved by applying linear regression and deviation analysis, so that trend judgment is more accurate, training and enhancement mode recognition of a vector machine model is supported, the calibration efficiency is improved, a neural network algorithm is rapidly adapted to environmental changes in real-time calibration, data accuracy is ensured, abnormal prediction and test flow optimization are enhanced by a Bayesian network and a graph theory algorithm, the detection reliability is improved, a data fusion technology is combined with machine learning, a test strategy is optimized, and accurate adaptation of results is ensured.
Description
Technical Field
The invention relates to the technical field of automatic testing, in particular to a detection control method and system based on an intelligent detection robot.
Background
The technical field of automated testing is focused on performing test tasks by automated tools and systems with the aim of improving the efficiency and accuracy of the test process. This area covers from simple automated test script writing to complex automated management of test procedures. The automatic test technology is widely applied to a plurality of industries such as software development, manufacturing industry, electronic product detection and the like, so that the requirement of manual testing is reduced, the testing speed and consistency are improved, and meanwhile, the error rate and cost are reduced.
The detection control method based on the intelligent detection robot is a detection method combining the intelligent robot technology and an automatic test principle, and aims to automatically execute detection tasks by using the intelligent robot so as to achieve the effect of improving detection efficiency, accuracy and reliability.
In the aspect of a test flow, the traditional detection control method lacks effective algorithm support, so that the test flow is often not efficient enough, redundant steps exist, the test efficiency is affected, the adaptability of the traditional method in response to environmental changes is poor, the test strategy is difficult to quickly and accurately adjust, the instability and the reliability of a test result are reduced, and the defects lead to prolonged test period, increased cost and inaccurate test result.
Disclosure of Invention
The invention aims to solve the defects in the prior art, and provides a detection control method and system based on an intelligent detection robot.
In order to achieve the above purpose, the present invention adopts the following technical scheme: a detection control method based on an intelligent detection robot comprises the following steps:
S1: based on intelligent detection robot data, a linear regression algorithm is adopted to analyze data trend, a deviation analysis method is adopted to compare the data with a standard value in real time, data deviation is identified and quantized, and the data is calibrated and adjusted through a self-adaptive filter, so that noise is reduced, errors are eliminated, and calibration data is generated;
s2: based on the calibration data, training a support vector machine model by using a support vector machine algorithm, identifying and learning historical test data, optimizing the accuracy and efficiency of a data calibration process and generating an optimized calibration model by using modes and correlations in a calibration result;
S3: based on the optimized calibration model, combining the temperature and humidity changes monitored in real time, applying a neural network algorithm to perform real-time calibration strategy adjustment, and generating environment adaptability calibration data;
s4: based on the environment adaptability calibration data, the prediction and recognition of the abnormal state are carried out through a Bayesian network model, the abnormal processing capacity in the test process is enhanced, and an abnormal state prediction result is generated;
S5: based on the abnormal state prediction result, a shortest path algorithm in graph theory is applied to construct a graph model of the test flow, the dependency relationship between test links is identified and analyzed, and a test path is calculated to generate an optimized test flow;
s6: based on the optimized test flow, analyzing sample characteristics and real-time data by adopting a decision tree algorithm, automatically adjusting test parameters, selecting a test method, and generating a self-adaptive test result;
S7: based on the self-adaptive test result, combining material science and physical data, performing data integration and feature extraction by using a data fusion technology, performing test strategy optimization by using a machine learning algorithm, optimizing test accuracy and adaptability, and generating a comprehensive optimization test result;
The calibration data comprises calibration values of temperature, pressure and humidity, the optimization calibration model specifically refers to a model based on support vector machine learning, the environment adaptability calibration data comprises response data to temperature change and humidity change factors, the abnormal state prediction result comprises fault points, abnormal indexes and occurrence probability, the optimization test flow comprises a sequence of test steps, a dependency relationship and an optimization path of a test link, the self-adaptation test result comprises an adjustment result of test parameters and a test method effect, and the comprehensive optimization test result comprises a data fusion analysis result, a test strategy optimization scheme and a test efficiency improvement scheme.
As a further scheme of the invention, based on intelligent detection robot data, a linear regression algorithm is adopted to analyze data trend, a deviation analysis method is applied to compare the data with a standard value in real time, data deviation is identified and quantized, and the data is calibrated and adjusted through an adaptive filter, so that noise is reduced, errors are eliminated, and the step of generating calibration data is specifically as follows:
s101: based on intelligent detection robot data, analyzing trends in the data set by adopting a linear regression algorithm, fitting trend lines and calculating slopes of the data points, identifying a data change mode, and generating a data trend analysis result;
S102: based on the data trend analysis result, comparing the data points with a preset standard value by using a deviation analysis method, calculating deviation values of the data points, analyzing deviation distribution of a data set, and generating a data deviation recognition result;
S103: based on the data deviation recognition result, a Kalman filter algorithm is applied to filter and eliminate errors of noise in a data set, the data is adjusted to be close to an actual value, the data quality is optimized, and noise reduction data is generated;
S104: based on the noise reduction data, a Gaussian process regression algorithm is adopted to calibrate the data, the data accuracy is optimized, a least square method is applied to adjust errors of the data, system deviation is eliminated, and calibration data is generated.
As a further scheme of the invention, based on the calibration data, a support vector machine algorithm is applied to train a support vector machine model, historical test data are identified and learned, patterns and correlations in a calibration result are identified, the accuracy and efficiency of a data calibration process are optimized, and the steps for generating an optimized calibration model are specifically as follows:
S201: based on the calibration data, performing model training by adopting a support vector machine algorithm, determining basic classification boundaries and decision rules, analyzing data characteristics and class distribution, and generating a basic support vector machine model;
s202: based on the basic support vector machine model, learning historical test data and a current calibration result by using a deep learning technology, enhancing the recognition capability of the model on a data mode, and generating a deep learning enhancement model;
S203: based on the deep learning enhancement model, analyzing modes and relations in the calibration result by using an association rule learning method, identifying the relation between data and generating a mode association analysis result;
S204: based on the mode association analysis result, the support vector machine model is optimized, parameters are adjusted, the structure is improved, the accuracy and the efficiency of the calibration process are improved, and an optimized calibration model is generated.
As a further scheme of the invention, based on the optimized calibration model, by combining the temperature and humidity changes monitored in real time, and applying a neural network algorithm, the method for performing real-time calibration strategy adjustment, and the method for generating the environment adaptive calibration data comprises the following steps:
S301: based on the optimized calibration model, a real-time data stream processing technology is adopted, environmental data are collected through a temperature and humidity sensor, and real-time change conditions of temperature and humidity are analyzed to generate environmental monitoring data;
S302: based on the environment monitoring data, adjusting the input parameters of the neural network by using a regression analysis method, adapting to the current environment change, adjusting the weight of the temperature and humidity data and standardizing the input, and generating the adjusted input parameters by reflecting the environment change through the input data;
s303: based on the adjustment input parameters, performing calibration strategy optimization by using a neural network, performing neural network weight adjustment by using a gradient descent algorithm, and generating an optimized calibration strategy by adapting to environmental changes by using the calibration strategy;
s304: based on the optimized calibration strategy, the environmental adaptation data is adjusted and refined using a data calibration technique to generate the environmental adaptation calibration data.
As a further aspect of the present invention, the step of predicting and identifying an abnormal state based on the environmental adaptability calibration data by using a bayesian network model to generate an abnormal state prediction result specifically includes:
S401: based on the environment adaptability calibration data, a Bayesian network model is constructed, a model structure and probability distribution are defined, network nodes, edges and probability relations are determined, data characteristics and potential relations are represented through the model, and a Bayesian network model structure is generated;
s402: based on the Bayesian network model structure, an expected maximization algorithm is applied to analyze an abnormal mode in the calibration data, and abnormal state prediction is carried out by calculating the conditional probability of the node to generate a probability inference result;
S403: based on the probability inference result, adopting principal component analysis and variance analysis to verify the abnormal mode, analyzing the frequency and influence factors of the abnormal mode, and generating an abnormal mode analysis result;
S404: based on the abnormal mode analysis result, the Bayesian decision theory is used for carrying out abnormal state prediction and identification, and probability evaluation and decision making are carried out on the differential abnormal state to generate an abnormal state prediction result.
As a further scheme of the invention, based on the abnormal state prediction result, a shortest path algorithm in graph theory is applied to construct a graph model of a test flow, the dependency relationship between test links is identified and analyzed, and a test path is calculated, and the steps for generating an optimized test flow are specifically as follows:
s501: based on the abnormal state prediction result, a graph model of a test flow is constructed by using graph theory analysis and Dijkstra algorithm, a test link is determined as a node, a relation between tests is taken as an edge, connectivity between the nodes is analyzed, and a test flow graph model is generated;
S502: based on the test flow chart model, a Bellman-Ford algorithm is applied, the dependency relationship between test links is identified, paths among multiple nodes are calculated, the structure of the test flow is optimized, and a test path analysis result is generated;
S503: based on the analysis result of the test path, a six-sigma method is applied to optimize the test flow, redundant test steps in the flow are identified and reduced through a DMAIC framework, test links are strengthened, test efficiency is improved, and a strengthened test flow is generated;
s504: based on the enhanced test flow, a system dynamic method is used for carrying out flow construction, and the optimized test flow is generated by simulating the dynamic behavior and interaction of the test flow, analyzing the test efficiency and coverage rate, balancing the resource allocation and time management.
As a further scheme of the invention, based on the optimized test flow, a decision tree algorithm is adopted to analyze sample characteristics and real-time data, automatically adjust test parameters, and select a test method, and the step of generating a self-adaptive test result is specifically as follows:
S601: based on the optimized test flow, adopting a decision tree algorithm, and combining the sample characteristics and real-time data to analyze the physical characteristics of the sample and the dynamic data in the test process so as to generate a sample characteristic analysis result;
S602: based on the sample characteristic analysis result, adopting an adaptive control theory and a model reference adaptive control method, combining real-time feedback data of the sample characteristic, dynamically adjusting the test rate and the pressure parameter, adapting to the sample characteristic, and generating an automatic adjustment test parameter;
s603: based on the automatic adjustment test parameters, adopting a genetic algorithm to simulate natural selection and a genetic mechanism, evaluating and selecting a differential test method, analyzing sample characteristics and test targets, selecting a preset test method, and generating a test method selection result;
S604: based on the test method selection result, a fuzzy logic control method is used, a test method and a fuzzy set of sample characteristics are combined, test parameters are adjusted, and the accuracy and the efficiency of the test are improved through the matching of the test parameters and the method, so that a self-adaptive test result is generated.
As a further scheme of the invention, based on the self-adaptive test result, the data integration and feature extraction are performed by combining material science and physical data and applying a data fusion technology, and the test strategy optimization, the test accuracy and adaptability are optimized by a machine learning algorithm, and the step of generating the comprehensive optimization test result comprises the following steps:
S701: based on the self-adaptive test result, combining material science and physical data, carrying out integration of material science and physical data by using principal component analysis, enriching test background information, and generating a data fusion result;
S702: based on the data fusion result, adopting K-means clustering analysis to identify groups and modes in the data, applying an Apriori association rule mining algorithm to analyze the relation and interdependence among multiple groups, carrying out data integration and feature extraction, capturing information and features, and generating a feature extraction result;
S703: based on the feature extraction result, a support vector machine algorithm is used for carrying out optimization analysis on the test strategy, the relation between the feature data and the test result is analyzed, the test strategy is optimized, the accuracy and the adaptability of the test effect are improved, and a test strategy optimization analysis result is generated;
s704: based on the test strategy optimization analysis result, a multi-criterion decision analysis method is adopted to evaluate the test strategy, and the test scheme is weighted and optimized by analyzing the cost benefit, risk evaluation, efficiency and reliability, so that the comprehensiveness and practicability of the test scheme are enhanced, and the comprehensive optimization test result is generated.
The detection control system based on the intelligent detection robot is used for executing the detection control method based on the intelligent detection robot, and comprises a data trend analysis calibration module, an environment adaptability calibration module, an abnormal state prediction module, a test flow optimization module, an adaptive test module and a comprehensive optimization analysis module.
As a further scheme of the invention, the data trend analysis and calibration module is based on intelligent detection robot data, adopts time sequence analysis and linear regression algorithm to analyze data trend, adopts a deviation analysis method to compare data with a standard value, identifies and quantifies data deviation, and adopts a Kalman filter and a Gaussian process regression algorithm to reduce noise and eliminate error of the data to generate calibration data;
The environment adaptability calibration module is used for carrying out real-time calibration strategy adjustment by utilizing a neural network algorithm based on calibration data and combining with the temperature and humidity changes monitored in real time, adapting to the environment changes and generating environment adaptability calibration data;
The abnormal state prediction module predicts and identifies the abnormal state by adopting a Bayesian network and an expected maximization algorithm based on the environmental adaptability calibration data, and generates an abnormal state prediction result;
the test flow optimization module builds a graph model of the test flow based on the abnormal state prediction result by using graph theory analysis and Dijkstra algorithm, identifies and optimizes the dependency relationship between test links, and generates an optimized test flow;
The self-adaptive test module is based on an optimized test flow, applies a decision tree algorithm and a fuzzy logic control method, analyzes sample characteristics and real-time data, automatically adjusts test parameters and selects a test method, improves the adaptability and the accuracy of the test, and generates a self-adaptive test result;
The comprehensive optimization analysis module performs optimization analysis of a test strategy based on a self-adaptive test result by combining material science and physical data and applying a data fusion technology and a support vector machine algorithm, ensures the accuracy and adaptability of a test scheme and generates a comprehensive optimization test result.
Compared with the prior art, the invention has the advantages and positive effects that:
According to the invention, the accuracy of data analysis is improved through the application of the linear regression and deviation analysis method, so that the judgment of data trend is more accurate, the error of data analysis is effectively reduced, the training of the support vector machine model strengthens the mode recognition capability of data, the accuracy and efficiency of the calibration process are improved, the neural network algorithm shows extremely strong adaptability in the adjustment of the real-time calibration strategy, the environment change can be responded rapidly, the data accuracy is ensured, the application of the Bayesian network model and the graph theory algorithm is ensured, the capability of abnormal state prediction and test flow optimization is enhanced, the reliability and efficiency of the whole detection process are improved, the combination of the data fusion technology and the machine learning algorithm further optimizes the test strategy, and the accuracy and the adaptability of the test result are ensured.
Drawings
FIG. 1 is a schematic workflow diagram of the present invention;
FIG. 2 is a S1 refinement flowchart of the present invention;
FIG. 3 is a S2 refinement flowchart of the present invention;
FIG. 4 is a S3 refinement flowchart of the present invention;
FIG. 5 is a S4 refinement flowchart of the present invention;
FIG. 6 is a S5 refinement flowchart of the present invention;
FIG. 7 is a S6 refinement flowchart of the present invention;
FIG. 8 is a S7 refinement flowchart of the present invention;
Fig. 9 is a system flow diagram of the present invention.
Detailed Description
The present invention will be described in further detail with reference to the drawings and examples, in order to make the objects, technical solutions and advantages of the present invention more apparent. It should be understood that the specific embodiments described herein are for purposes of illustration only and are not intended to limit the scope of the invention.
In the description of the present invention, it should be understood that the terms "length," "width," "upper," "lower," "front," "rear," "left," "right," "vertical," "horizontal," "top," "bottom," "inner," "outer," and the like indicate orientations or positional relationships based on the orientation or positional relationships shown in the drawings, merely to facilitate describing the present invention and simplify the description, and do not indicate or imply that the devices or elements referred to must have a specific orientation, be configured and operated in a specific orientation, and therefore should not be construed as limiting the present invention. Furthermore, in the description of the present invention, the meaning of "a plurality" is two or more, unless explicitly defined otherwise.
Example 1
Referring to fig. 1, the present invention provides a technical solution: a detection control method based on an intelligent detection robot comprises the following steps:
S1: based on intelligent detection robot data, a linear regression algorithm is adopted to analyze data trend, a deviation analysis method is adopted to compare the data with a standard value in real time, data deviation is identified and quantized, and the data is calibrated and adjusted through a self-adaptive filter, so that noise is reduced, errors are eliminated, and calibration data is generated;
S2: based on the calibration data, training a support vector machine model by using a support vector machine algorithm, identifying and learning historical test data, optimizing the accuracy and efficiency of a data calibration process and generating an optimized calibration model by using modes and correlations in a calibration result;
s3: based on the optimized calibration model, combining the temperature and humidity changes monitored in real time, applying a neural network algorithm to perform real-time calibration strategy adjustment, and generating environment adaptability calibration data;
S4: based on the environment adaptability calibration data, the prediction and recognition of the abnormal state are carried out through a Bayesian network model, the abnormal processing capacity in the test process is enhanced, and an abnormal state prediction result is generated;
S5: based on the abnormal state prediction result, a shortest path algorithm in graph theory is applied to construct a graph model of the test flow, the dependency relationship between test links is identified and analyzed, and a test path is calculated to generate an optimized test flow;
S6: based on the optimized test flow, adopting a decision tree algorithm to analyze sample characteristics and real-time data, automatically adjusting test parameters, selecting a test method, and generating a self-adaptive test result;
s7: based on the self-adaptive test result, combining material science and physical data, performing data integration and feature extraction by using a data fusion technology, performing test strategy optimization by using a machine learning algorithm, optimizing test accuracy and adaptability, and generating a comprehensive optimization test result;
The calibration data comprises calibration values of temperature, pressure and humidity, the optimization calibration model specifically refers to a model based on support vector machine learning, the environment adaptability calibration data comprises response data to temperature change and humidity change factors, the abnormal state prediction result comprises fault points, abnormal indexes and occurrence probability, the optimization test flow comprises the sequence of test steps, the dependence relationship and the optimization path of a test link, the self-adaptation test result comprises the adjustment result of test parameters and the test method effect, and the comprehensive optimization test result comprises a data fusion analysis result, a test strategy optimization scheme and a test efficiency improvement scheme.
In step S1, trend analysis is carried out on time series data by intelligently detecting data collected by a robot, including sensor readings of temperature, pressure, humidity and the like, a linear regression algorithm is applied, the process comprises the steps of determining linear relation of the data, calculating regression coefficients and slopes to identify long-term trend and seasonal change of the data, a deviation analysis method is used for comparing real-time data with preset standard values, calculating deviation amounts to identify abnormal points in the data, an adaptive filter adjusts the data according to the result of the deviation analysis, reduces noise influence, eliminates data errors, improves accuracy and usability of the data, and calibration data generated in the step reflects the environment and operation conditions subjected to accurate correction, so that accurate basic data is provided for subsequent steps.
In step S2, based on the calibration data, further pattern recognition and learning are performed on the data by using a support vector machine algorithm, which classifies the data by constructing one or more hyperplanes, maximizing the interval between different types of data, and the training process involves selecting appropriate kernel functions, adjusting parameters such as regularization coefficients and kernel parameters, so as to enhance the recognition capability of the model on patterns in the historical test data.
In the step S3, based on an optimized calibration model and combined with temperature and humidity change data monitored in real time, a neural network algorithm is applied to carry out dynamic calibration strategy adjustment, the neural network adjusts the weight and deviation of the neural network according to input environmental data such as temperature and humidity changes, quick response to the environmental change is realized, the neural network can learn how to keep the accuracy and consistency of the data under different environmental conditions through training and adjusting the network, the environmental adaptability calibration data generated in the step provides effective response to the real-time environmental change, and the adaptability and accuracy of data processing are ensured.
In step S4, prediction and recognition of the abnormal state are performed using the bayesian network model based on the environmental adaptive calibration data. The bayesian network models causal relationships between various variables by constructing a probabilistic graph model. The Bayesian network can predict occurrence probabilities of different abnormal states by calculating the conditional probabilities of each node, such as fault points, abnormal indexes and the like, and the result of the step is an abnormal state prediction result, including fault points, abnormal indexes, occurrence probabilities and the like, so that powerful support is provided for abnormal processing in the test process.
In step S5, based on the abnormal state prediction result, a shortest path algorithm in graph theory, such as Dijkstra algorithm, is adopted to optimize the test flow, this step involves regarding the test flow as a graph model, wherein nodes represent the test steps, edges represent the dependency relationships between the steps, the shortest path algorithm is used to calculate the most effective path from the starting point to the end point, so as to optimize the sequence and the dependency relationship of the test steps, and the result of this step is an optimized test flow, including the sequence of the test steps, the dependency relationship and the optimized path of the test links, so that the efficiency and coverage rate of the whole test process are improved.
In step S6, based on the optimized test flow, the decision tree algorithm is applied to analyze the sample characteristics and real-time data, so as to automatically adjust the test parameters, and select the most suitable test method, the decision tree analyzes the data by creating a series of rules, and determines the optimal test parameter setting according to the characteristics of the sample, such as the material type, the size, and the like, and the real-time data, such as the current operation conditions, the environmental variables, and the like, and the result of this step is an adaptive test result, including the adjustment result of the test parameters and the effect of the test method, thereby ensuring the flexibility and the adaptability of the test process.
In step S7, based on the self-adaptive test result, the data integration and feature extraction are performed by using the data fusion technology in combination with the material science and physics data, the process includes merging data from different sources, such as experimental data and theoretical model data, and analyzing the merged data set by using a machine learning algorithm to identify key features and modes, the machine learning algorithm such as a support vector machine and a neural network is used for identifying complex modes in the data, and optimizing the test strategy, and the result of the step is comprehensive optimization test result, including data fusion analysis result, test strategy optimization scheme and test efficiency improvement scheme, so that the overall performance of the test process is effectively improved.
Referring to fig. 2, based on intelligent detection robot data, a linear regression algorithm is adopted to analyze data trend, a deviation analysis method is applied to compare the data with a standard value in real time, data deviation is identified and quantized, and the data is calibrated and adjusted through an adaptive filter, so that noise is reduced, errors are eliminated, and the step of generating calibration data is specifically as follows:
s101: based on intelligent detection robot data, analyzing trends in the data set by adopting a linear regression algorithm, fitting trend lines and calculating slopes of the data points, identifying a data change mode, and generating a data trend analysis result;
s102: based on the data trend analysis result, comparing the data points with a preset standard value by using a deviation analysis method, calculating deviation values of the data points, analyzing deviation distribution of a data set, and generating a data deviation recognition result;
s103: based on the data deviation recognition result, a Kalman filter algorithm is applied to filter and eliminate errors of noise in a data set, the data is adjusted to be close to an actual value, the data quality is optimized, and noise reduction data is generated;
s104: based on the noise reduction data, a Gaussian process regression algorithm is adopted to calibrate the data, the data accuracy is optimized, a least square method is adopted to carry out error adjustment on the data, the system deviation is eliminated, and the calibration data is generated.
In the sub-step S101, the trend analysis is performed on the time series data by using the data collected from the intelligent detection robot, such as the readings of temperature, pressure, humidity and the like, and a linear regression algorithm is adopted to firstly determine whether a linear relation exists in the data, and the most suitable straight line is fitted by calculating the minimum square difference between the data points and the trend line, wherein the process comprises the steps of calculating the slope and intercept of the trend line so as to identify the main trend of the data change, if the data is the change of the temperature with time, the linear regression can reveal whether the temperature steadily rises or falls with the time, and the data trend analysis result generated in the step provides insight into the long-term change mode and provides a basis for the subsequent data processing.
In the sub-step S102, based on the data trend analysis result, a deviation analysis method is applied to further refine the data processing, this process involves comparing each data point with a preset standard value, calculating a deviation value, and analyzing the distribution of deviations in the whole data set, the deviation analysis can reveal differences between the data points and expected values, help identify abnormal points or patterns in the data, for example, the standard value is a certain temperature range, and the deviation analysis can reveal which data points deviate from this range, by this method, the generated data deviation identification result helps the subsequent step to more accurately adjust and calibrate the data.
In the step S103, based on the data deviation recognition result, a kalman filter algorithm is applied to filter and eliminate errors in the data set, and the kalman filter is an efficient algorithm for estimating the state of the dynamic system and can provide optimal estimation under the condition of noise and uncertainty.
In the step S104, based on noise reduction data, a Gaussian process regression algorithm is adopted for final calibration, the Gaussian process regression is a strong non-parameter statistical model and is used for prediction and interpolation, the method is particularly suitable for complex data with noise, the Gaussian process regression optimizes data accuracy by considering correlation among data points, a least square method is used for further adjusting the data, system deviation is eliminated, and finally generated calibration data is guaranteed to be high in accuracy and high in reliability.
Assume that a set of temperature data collected from a smart inspection robot is as follows: 22.1, 22.3, 22.5, 22.7, 22.4, 22.8, 22.6, in step S101, the linear regression algorithm analyzes the data to calculate a trend line with a slope of 0.1, indicating that the temperature generally increases, in step S102, if the standard temperature value is 22.5 ℃, the deviation analysis shows that most data points are close to the standard value, but individual points deviate farther, in step S103, the kalman filter processes the data to reduce noise effects, the adjusted data is smoother, in step S104, the gaussian process regression further optimizes the data, the data after the least square method adjustment is closer to the real temperature, and the finally generated calibration data reflects a more accurate temperature variation trend.
Referring to fig. 3, based on calibration data, using a support vector machine algorithm, training a support vector machine model, identifying and learning historical test data, and optimizing the accuracy and efficiency of the data calibration process, the steps of generating an optimized calibration model are specifically as follows:
S201: based on the calibration data, performing model training by adopting a support vector machine algorithm, determining basic classification boundaries and decision rules, analyzing data characteristics and class distribution, and generating a basic support vector machine model;
S202: based on a basic support vector machine model, learning historical test data and a current calibration result by using a deep learning technology, enhancing the recognition capability of the model on a data mode, and generating a deep learning enhancement model;
S203: based on the deep learning enhancement model, analyzing modes and relations in the calibration result by using an association rule learning method, identifying the relation between data, and generating a mode association analysis result;
S204: based on the mode association analysis result, the support vector machine model is optimized, parameters are adjusted, the structure is improved, the accuracy and the efficiency of the calibration process are improved, and an optimized calibration model is generated.
In the step S201, model training is carried out on calibration data collected based on the intelligent detection robot by adopting a support vector machine algorithm, the support vector machine algorithm identifies characteristics in the data, basic classification boundaries and decision rules are determined according to the characteristics, the support vector machine constructs a high-dimensional space, each data point is mapped into a point in the space, the support vector machine algorithm searches an optimal hyperplane to separate data points of different categories, and in order to determine the hyperplane, the support vector machine calculates support vectors, namely data points closest to the classification boundaries, and maximizes the distance between the support vectors and the boundaries, in this way, the support vector machine can identify and distinguish different categories in a complex data set, and a basic support vector machine model is generated, which is critical to understanding modes and relations in the data, and provides a basis for subsequent deep learning enhancement and mode association analysis.
In the S202 substep, based on a basic support vector machine model, deep learning technology is introduced to deep learn historical test data and a current calibration result, in the process, a deep learning algorithm extracts advanced features and complex modes of data through a multi-layer neural network structure, the step involves setting parameters such as the number of layers, the number of neurons, an activation function, a learning rate and the like of a neural network, training the network, so that the network can learn a fine mode in the data, and the deep learning model can gradually improve the understanding of the deep learning model to the data by continuously adjusting network parameters, thereby enhancing the recognition capability of the support vector machine model to the data mode, and the deep learning enhancement model generated in the process is more accurate and efficient in terms of data calibration and is beneficial to improving the performance of the whole system.
In the step S203, based on the generated deep learning enhancement model, the patterns and relationships in the calibration result are analyzed by using an association rule learning method, and in this step, using an association rule learning technology such as Apriori algorithm, the relationships between various items are found from the data, for example, the association rule may reveal a specific pattern of the data under a specific environmental condition, and this analysis helps to understand the implicit relationship between the data, so that the model can better predict and respond to future data, and the generated pattern association analysis result provides a deep understanding of the internal structure of the data, and provides a basis for further optimizing the support vector machine model.
In the sub-step S204, the support vector machine model is further optimized based on the pattern correlation analysis result, which includes adjusting the kernel function, regularization parameters and fault tolerance of the support vector machine to improve the accuracy and efficiency of the model in processing new data, and in addition, the structure of the support vector machine model is adjusted, e.g. the number of support vectors is increased or decreased, according to the pattern correlation analysis result to adapt to the new pattern identified in the data, such optimization makes the support vector machine model more suitable for specific data sets and tasks, and the generated optimized calibration model shows higher accuracy and efficiency in the data calibration process.
Assuming that the intelligent detection robot collects a set of data about the operation state of the mechanical device, including temperature, pressure, vibration intensity, etc., the data may be in the form of time series data, the operation state of the mechanical device at different time points is recorded, in step S201, the data is analyzed by a support vector machine algorithm, the relationship between the vibration intensity and the mechanical fault is determined, for example, the support vector machine may find that when the vibration intensity exceeds a certain threshold value, the probability of the device failure increases significantly, in step S202, the deep learning model further analyzes the historical data, identifies the association between the specific vibration mode and the fault type, in step S203, the association rule learning reveals the relationship between the vibration intensity and the temperature change, for example, the small change of the vibration intensity may indicate the impending fault under the condition of temperature rise, in step S204, the support vector machine model is adjusted to reflect these new findings, an optimized calibration model is generated, and the device fault can be predicted and maintained more accurately.
Referring to fig. 4, based on the optimized calibration model, in combination with the temperature and humidity changes monitored in real time, a neural network algorithm is applied to perform real-time calibration policy adjustment, and the step of generating the environment adaptive calibration data specifically includes:
S301: based on an optimized calibration model, a real-time data stream processing technology is adopted, environmental data are collected through a temperature and humidity sensor, and real-time change conditions of temperature and humidity are analyzed to generate environmental monitoring data;
S302: based on the environment monitoring data, adjusting the input parameters of the neural network by using a regression analysis method, adapting to the current environment change, adjusting the weight of the temperature and humidity data, standardizing the input, and generating the adjusted input parameters by reflecting the environment change through the input data;
s303: based on the adjustment input parameters, the neural network is utilized to perform calibration strategy optimization, the neural network weight adjustment is performed through a gradient descent algorithm, the environment change is adapted through the calibration strategy, and an optimized calibration strategy is generated;
S304: based on the optimized calibration strategy, the environmental adaptation data is adjusted and refined using a data calibration technique to generate the environmental adaptation calibration data.
In step S301, environmental data, in particular temperature and humidity information, of the intelligent detection robot is collected by a real-time data stream processing technology, the process involves collecting the environmental data by using a temperature and humidity sensor, the data format generally includes a time stamp and corresponding temperature and humidity readings, the collected data is transmitted and analyzed in real time to monitor real-time changes of environmental conditions, for example, the data stream processing may involve performing time series analysis on the temperature and humidity data to identify trends and modes of environmental changes, and the generated environmental monitoring data can reflect the current environmental conditions in detail, so as to provide an accurate basis for adjustment of subsequent calibration strategies.
In the step S302, based on the environmental monitoring data, the regression analysis method is used to adjust the input parameters of the neural network model, in this step, by analyzing the relationship between the temperature and humidity data and the performance of the equipment, it is determined how to adjust the input weight of the neural network and the standardized input data to reflect the change of the current environment, for example, if the analysis shows that the change of the temperature has a great influence on the performance of the equipment, the influence weight of the temperature data on the neural network is increased, and the adjusted input parameters generated in this step can ensure that the neural network model can still accurately perform data processing and analysis under different environmental conditions.
In the step S303, based on the adjustment input parameters, optimization of the calibration strategy is performed by using the neural network, and this process includes adjusting the weights of the neural network by using a gradient descent algorithm to ensure that the network output is as close as possible to the desired calibration result, and by iteratively adjusting the network weights, the neural network can learn how to better adapt to environmental changes and perform effective data calibration, and the optimized calibration strategy is more flexible and accurate, can quickly respond to changes in environmental conditions, and improves accuracy and efficiency of data processing.
In a sub-step S304, the environmental adaptation data is further refined and adjusted using a data calibration technique based on an optimized calibration strategy, which involves applying the output of the neural network to the actual data to adjust and optimize the environmental adaptation data, e.g. making necessary adjustments to the temperature and humidity data according to the calibration strategy provided by the neural network to ensure that the data is still accurate and consistent under different environmental conditions, the generated environmental adaptation calibration data being critical to ensure reliable operation of the system under different environments.
Assuming a practical scenario, the intelligent detection robot is deployed in a factory to monitor the operating conditions of the equipment, the robot collects temperature and humidity data, such as temperature data over an hour: [22 ℃, 22.5 ℃, 23 ℃, 22.7 ℃, 23.2 ℃), humidity data: in the step S302, the relation between the temperature and the equipment performance is found to be more obvious through regression analysis, so that a neural network model is adjusted, the weight of temperature data is increased, in the step S303, the neural network is optimized by using a gradient descent algorithm, the influence of environmental change on the equipment performance is ensured to be effectively reflected, in the step S304, the original temperature and humidity data is adjusted according to an optimized calibration strategy, and the generated environmental adaptability calibration data more accurately reflects the equipment operation environment, so that key support is provided for ensuring the stable operation of equipment.
Referring to fig. 5, based on the environmental adaptability calibration data, the abnormal state is predicted and identified through the bayesian network model, so as to enhance the abnormal processing capability in the test process, and the step of generating the abnormal state prediction result specifically includes:
S401: based on the environment adaptability calibration data, a Bayesian network model is constructed, a model structure and probability distribution are defined, network nodes, edges and probability relations are determined, data characteristics and potential relations are represented through the model, and the Bayesian network model structure is generated;
s402: based on a Bayesian network model structure, an expected maximization algorithm is applied to analyze an abnormal mode in the calibration data, and abnormal state prediction is carried out by calculating the conditional probability of the node to generate a probability inference result;
S403: based on the probability inference result, adopting principal component analysis and variance analysis to verify the abnormal mode, and analyzing the frequency and influence factors of the abnormal mode to generate an abnormal mode analysis result;
S404: based on the analysis result of the abnormal mode, the Bayesian decision theory is used for carrying out abnormal state prediction and identification, and carrying out probability evaluation and decision making on the differential abnormal state to generate an abnormal state prediction result.
In a sub-step S401, a bayesian network model is constructed by calibrating data based on environmental adaptation. The process first defines the structure and probability distribution of the model, including determining nodes and edges in the network, the nodes represent various environmental parameters such as temperature and humidity, the edges represent probability relations among the nodes, the probability distribution of each node is estimated according to historical data and priori knowledge in the construction process, the edges among the nodes reflect probability dependence among the variables, for example, the temperature and the humidity can influence each other, the relations can be represented in the form of edges in the model, in this way, the Bayesian network can capture the characteristics and potential relations of the data, and the generated model structure is beneficial to effectively carrying out probability reasoning on complex environmental data.
In the step S402, based on the Bayesian network model structure, the abnormal mode in the calibration data is analyzed by using an expectation maximization algorithm, wherein the expectation maximization algorithm is an iterative algorithm and is used for finding the optimal estimation of the probability model parameters, especially when the unobservable potential variables exist in the model, in the step, the probability of the abnormal state possibly occurring in equipment under different environments can be predicted by the expectation maximization algorithm through calculating the conditional probability of each node in the network, and the probability inference result generated by the method reveals the occurrence probability of the abnormal state under various environmental conditions and provides scientific basis for the subsequent recognition and processing of the abnormal state.
In the step S403, based on the probability inference result, the abnormal mode is further verified and analyzed by adopting principal component analysis and variance analysis, wherein the principal component analysis is a dimension reduction technology, key features and structures of the data are revealed by extracting principal components in the data, and the variance analysis is used for evaluating the influence degree and significance among different variables.
In the S404 substep, based on the analysis result of the abnormal mode, the abnormal state is predicted and identified by using a Bayesian decision theory, which is a decision making method, the optimal decision is selected by calculating the expected loss under different decisions, in this step, probability evaluation is performed on the differentiated abnormal state, and reasonable response strategies are formulated by combining factors such as cost, risk and influence, so that the prediction result of the abnormal state generated by the method provides accurate decision support for equipment operation staff, and the influence caused by the abnormal state can be effectively prevented and reduced.
Assuming that in a chemical plant, the intelligent detection robot collects a series of environmental data related to the production process, including temperature, humidity, chemical concentration, etc., the temperature data may include, for example, continuously recorded average temperature per hour: [30 ℃, 31 ℃, 29 ℃, 32 ℃), humidity data: [60%, 62%, 58%, 65% ], in step S401, a bayesian network model is constructed that considers the interplay between temperature, humidity and chemical concentration, in step S402, these data are analyzed by a expectation maximization algorithm, it is found that the probability of chemical concentration anomalies is significantly increased at a particular temperature and humidity combination, in step S403, principal component analysis and analysis of variance reveal the primary contributors to chemical concentration anomalies by temperature and humidity changes, in step S404, bayesian decision theory is used to evaluate the effects of different response strategies, such as choosing whether to immediately stop the production line for inspection upon detection of potential chemical concentration anomalies, which together provide an effective set of environmental monitoring and anomaly prediction systems for chemical plants, significantly improving production safety and efficiency.
Referring to fig. 6, based on the abnormal state prediction result, a shortest path algorithm in graph theory is applied to construct a graph model of a test flow, identify and analyze the dependency relationship between test links, and calculate a test path, and the steps for generating an optimized test flow are specifically as follows:
S501: based on the abnormal state prediction result, constructing a graph model of a test flow by using graph theory analysis and Dijkstra algorithm, determining test links as nodes, determining the relation between tests as edges, and analyzing the connectivity between the nodes to generate the test flow graph model;
S502: based on a test flow chart model, a Bellman-Ford algorithm is applied, the dependency relationship between test links is identified, paths among multiple nodes are calculated, the structure of the test flow is optimized, and a test path analysis result is generated;
S503: based on the analysis result of the test path, a six-sigma method is applied to optimize the test flow, redundant test steps in the flow are identified and reduced through a DMAIC framework, test links are strengthened, test efficiency is improved, and a strengthened test flow is generated;
s504: based on the reinforced test flow, a system dynamic method is applied to process construction, and the test efficiency and coverage rate are analyzed by simulating the dynamic behavior and interaction of the test flow, so that the resource allocation and time management are balanced, and the optimized test flow is generated.
In the sub-step S501, a graph model of the test flow is constructed by combining graph theory analysis and Dijkstra algorithm, each link in the test flow is defined as a node in the graph according to the abnormal state prediction result, the dependency relationship between the nodes is used as an edge, for example, if the result of one test step directly affects the start of another step, the two steps are connected through one edge in the graph, dijkstra algorithm is used to find the shortest path from one node to other nodes in the graph, which helps to identify the most efficient test flow, in this way, the graph model can clearly show the connection between the structure and links of the test flow, and the generated graph model of the test flow helps to understand and optimize the whole test flow.
In the step S502, based on the test flow chart model, the Bellman-Ford algorithm is applied to identify the dependency relationship among the test links, and can process the negative weight edge in the graph, so that the Bellman-Ford algorithm can calculate the shortest path among a plurality of nodes in the graph, even in the complex dependency relationship, the method helps to determine the optimal execution sequence and path of each link in the test flow, and optimize the structure of the test flow. The generated test path analysis result reveals key links and potential bottlenecks in the test flow, and provides basis for further flow optimization.
In the sub-step S503, based on the analysis result of the test path, the test flow is optimized by applying a six sigma method, which is a methodology aimed at improving quality and efficiency, and its DMAIC framework is used to systematically optimize the test flow, this step involves identifying redundant test steps in the flow, and proposing improvement measures to cut down these steps, while reinforcing the key test links.
In the step S504, the flow construction is performed based on the reinforced test flow and by using a system dynamics method, the system dynamics is a method for analyzing and simulating complex system behaviors, the analysis and optimization of the test efficiency and coverage rate are helped by simulating the dynamic behaviors and interactions of the test flow, in this step, the test flow is simulated by establishing a mathematical model, and the resource allocation, time management and flow interaction are analyzed, so that the efficiency and effect of the whole flow are balanced, and the generated optimized test flow provides an effective scheme for realizing efficient and comprehensive test.
Assuming a home electronics product manufacturing company, a complex series of quality tests are required for its products. In step S501, a graph model is built based on the abnormal prediction result of the test flow, wherein the graph model comprises test items such as functional test and durability test as nodes, the dependency relationship among the nodes is used as edges, the shortest path from the preliminary test to the final test is found through Dijkstra algorithm, in step S502, the test path is further optimized by applying Bellman-Ford algorithm, the key test links and potential bottlenecks are identified, in step S503, the test flow is optimized through six sigma method, the redundancy steps are reduced, the key test links are strengthened, in step S504, the flow model is built through the systematic dynamic method, the resource allocation and time management are balanced through simulation analysis, and therefore, a high-efficiency and comprehensive test flow is generated, and the efficiency and quality of the product test are improved.
Referring to fig. 7, based on the optimized test flow, the decision tree algorithm is adopted to analyze the sample characteristics and real-time data, automatically adjust the test parameters, and select the test method, and the steps for generating the adaptive test result are specifically as follows:
s601: based on the optimized test flow, a decision tree algorithm is adopted, and the physical characteristics of the sample and the dynamic data in the test process are analyzed by combining the characteristics of the sample with real-time data to generate a sample characteristic analysis result;
S602: based on the analysis result of the sample characteristics, adopting an adaptive control theory and a model reference adaptive control method, combining real-time feedback data of the sample characteristics, dynamically adjusting the test rate and the pressure parameters, adapting to the sample characteristics, and generating automatic adjustment test parameters;
s603: based on automatic adjustment of test parameters, adopting a genetic algorithm to simulate natural selection and a genetic mechanism, evaluating and selecting a differential test method, analyzing sample characteristics and test targets, selecting a preset test method, and generating a test method selection result;
S604: based on the test method selection result, a fuzzy logic control method is used, a test method and a fuzzy set of sample characteristics are combined, test parameters are adjusted, and the test accuracy and efficiency are improved through test parameter and method matching, so that a self-adaptive test result is generated.
In the S601 substep, analysis is performed by combining sample characteristics and real-time data through a decision tree algorithm, a sample characteristic analysis result is generated, the decision tree algorithm analyzes the data by creating a tree structure, wherein each internal node represents an attribute test, each branch represents a test result, and each leaf node represents a class label.
In the step S602, based on the analysis result of the sample characteristics, the test parameters are dynamically adjusted by applying an adaptive control theory and a model reference adaptive control method, wherein the model reference adaptive control is a method for automatically adjusting the parameters of a controller to adapt to the dynamic change of a system, the model reference adaptive control is particularly suitable for processing a system with unknown or changed parameters, in the step, the model reference adaptive control automatically adjusts the test rate and the pressure parameters by comparing the characteristics of the current sample with a desired reference model, so that the test process is ensured to adapt to the actual conditions of the sample, and the automatically adjusted test parameters generated by the method can obviously improve the accuracy and the efficiency of the test, and ensure that the test result can accurately reflect the actual conditions of the sample.
In the step S603, based on automatic adjustment of test parameters, a genetic algorithm is applied to simulate natural selection and a genetic mechanism, different test methods are evaluated and selected, the genetic algorithm is a heuristic search algorithm, the optimization problem is solved by simulating genetic and natural selection in the biological evolution process, in the step, the algorithm codes different test methods, such as pressure test and temperature test, and evaluates and selects the different test methods in the simulated evolution process so as to find the test method which is most suitable for the current sample characteristics and the test targets, and the generated test method selection result can ensure that the test flow can meet the test targets and adapt to the change of the sample characteristics.
In the step S604, based on the selection result of the test method, the further adjustment of the test parameters is performed by using a fuzzy logic control method, the fuzzy logic control is a control strategy based on the fuzzy logic principle, and is suitable for processing a system with higher uncertainty and ambiguity.
Assuming a specific application scenario, an automotive part manufacturer needs to perform a series of quality tests on its products, in step S601, decision tree algorithms are used to analyze the physical characteristics of the part, such as metal type, weight, and dynamic data during the test, such as stress test readings, for example, algorithms may find that a particular type of metal is more prone to fatigue under high pressure, based on this analysis, in step S602, model reference adaptive control is used to dynamically adjust the parameters of the stress test to better accommodate the characteristics of different types of metals, in step S603, genetic algorithms are used to select the most suitable test method, for example, for certain metal types, select a higher frequency stress test, in step S604, fuzzy logic control is used to fine tune the test parameters, such as adjusting the duration and frequency of the stress test, to ensure accuracy and reliability of the test results, by which the manufacturer can ensure that its product quality test is both accurate and efficient, thereby improving product quality and reducing the risk of rejects.
Referring to fig. 8, based on the self-adaptive test result, the data integration and feature extraction are performed by combining material science and physical data and applying a data fusion technology, and the test strategy optimization is performed by a machine learning algorithm, so as to optimize the test accuracy and adaptability, and the steps for generating the comprehensive optimization test result are specifically as follows:
S701: based on the self-adaptive test result, combining material science and physical data, carrying out integration of material science and physical data by using principal component analysis, enriching test background information, and generating a data fusion result;
S702: based on the data fusion result, adopting K-means clustering analysis to identify groups and modes in the data, applying an Apriori association rule mining algorithm to analyze the relation and interdependence among multiple groups, carrying out data integration and feature extraction, capturing information and features, and generating a feature extraction result;
S703: based on the feature extraction result, carrying out optimization analysis on the test strategy by using a support vector machine algorithm, analyzing the relation between the feature data and the test result, optimizing the test strategy, improving the accuracy and adaptability of the test effect, and generating a test strategy optimization analysis result;
S704: based on the test strategy optimization analysis result, a multi-criterion decision analysis method is adopted to evaluate the test strategy, and the test scheme is weighted and optimized by analyzing the cost benefit, risk evaluation, efficiency and reliability, so that the comprehensiveness and practicability of the test scheme are enhanced, and the comprehensive optimization test result is generated.
In the sub-step S701, data fusion is performed by combining the adaptive test result and the material science and physics data, in this process, material science data, such as composition, density, and physics data, such as thermal conductivity and resistivity, related to the sample are collected, then the data are subjected to dimension reduction processing by applying principal component analysis, and the principal component analysis reduces the complexity of the data by identifying principal variables and modes in the data, while retaining the most important information, for example, if the original data set includes 100 variables, the principal component analysis may reduce the principal component analysis to 10 principal components, and these components still capture the core information of the sample characteristics, and the data fusion result generated by this method enriches the test background information, provides a more comprehensive view angle for the subsequent test, and helps to improve the accuracy and effectiveness of the test strategy.
In the S702 substep, based on the data fusion result, a K-means clustering analysis and an Apriori association rule mining algorithm are adopted, the K-means clustering analysis identifies modes and trends in data by grouping data points into K groups, in the process, the algorithm iteratively distributes each data point to the nearest group center until the distribution of the groups is stable, for sample groups with different material characteristics, the K-means clustering can help identify which attributes are key factors for distinguishing different groups, the relation and interdependence among the groups are analyzed by applying the Apriori algorithm, the association rule among the groups is mined, and the feature extraction result generated by the method captures key information and features of the sample data and provides an important basis for optimizing a test strategy.
In the step S703, based on the feature extraction result, the test strategy is optimized and analyzed by using a support vector machine algorithm, which is a supervised learning algorithm widely used for classification and regression analysis, and in this step, the support vector machine analyzes the relationship between the feature data and the test result by constructing a hyperplane that can separate different groups of samples to the maximum extent, the algorithm trains a model using the extracted feature data, and then predicts the effects of different test strategies using the model, and the generated test strategy optimizes the analysis result to provide specific guidance on how to adjust the test parameters to achieve the optimal effect, thereby improving the accuracy and adaptability of the test effect.
In the sub-step S704, based on the test strategy optimization analysis result, the test strategy is comprehensively evaluated by adopting a multi-criterion decision analysis method, the process involves analyzing various aspects of the test strategy, including cost benefit, risk assessment, efficiency and reliability, for example, the cost benefit can be evaluated by comparing the cost of different test strategies with the detection accuracy thereof, the risk assessment involves considering potential risks and uncertainties possibly brought by the test strategy, by adopting the method, the multi-criterion decision analysis is helpful for balancing various factors of the test strategy, optimizing the comprehensiveness and practicability of the test scheme, and the generated comprehensive optimization test result provides comprehensive guidance for the final test scheme, so that the efficiency, accuracy and reliability of the test are ensured.
Assuming an aircraft equipment manufacturing company, a series of strict quality tests are required for composite materials, in the step S701, the company performs data fusion by combining chemical component and physical characteristic data of the composite materials and using principal component analysis, so that the data are simplified and key characteristics are highlighted, in the step S702, K-means clustering analysis helps to identify groups of different types of composite materials, the Apriori algorithm reveals association rules between different material characteristics, in the step S703, a support vector machine algorithm is used for analyzing the relationship between the characteristics and performance test results of the materials, and an optimized test strategy is provided, in the step S704, multi-criterion decision analysis is used for evaluating cost benefits, risks and efficiency of different test strategies, a comprehensive optimized test scheme is generated, and the company can ensure that the test of the composite materials is comprehensive and efficient, and the product quality and safety are greatly improved.
Referring to fig. 9, a detection control system based on an intelligent detection robot is used for executing the detection control method based on the intelligent detection robot, and the system comprises a data trend analysis calibration module, an environment adaptability calibration module, an abnormal state prediction module, a test flow optimization module, an adaptive test module and a comprehensive optimization analysis module.
The data trend analysis and calibration module is used for analyzing data trend by adopting a time sequence analysis and linear regression algorithm based on intelligent detection robot data, comparing the data with a standard value by using a deviation analysis method, identifying and quantifying data deviation, and carrying out noise reduction and error elimination on the data by adopting a Kalman filter and a Gaussian process regression algorithm to generate calibration data;
The environment adaptability calibration module is based on calibration data, combines the temperature and humidity changes monitored in real time, and utilizes a neural network algorithm to perform real-time calibration strategy adjustment, adapt to the environment changes and generate environment adaptability calibration data;
The abnormal state prediction module predicts and identifies the abnormal state by adopting a Bayesian network and an expected maximization algorithm based on the environment adaptability calibration data, and generates an abnormal state prediction result;
The test flow optimization module builds a graph model of the test flow based on the abnormal state prediction result by using graph theory analysis and Dijkstra algorithm, identifies and optimizes the dependency relationship between test links, and generates an optimized test flow;
The self-adaptive test module is based on an optimized test flow, applies a decision tree algorithm and a fuzzy logic control method, analyzes sample characteristics and real-time data, automatically adjusts test parameters and selects a test method, improves the adaptability and the accuracy of the test, and generates a self-adaptive test result;
The comprehensive optimization analysis module performs optimization analysis of the test strategy by combining material science and physical data and applying a data fusion technology and a support vector machine algorithm based on the self-adaptive test result, ensures the accuracy and adaptability of the test scheme and generates a comprehensive optimization test result.
The data trend analysis calibration module is used, the time sequence analysis and the linear regression algorithm are combined, so that data processing becomes more accurate and efficient, through deep analysis of data trend and deviation, accuracy and reliability of data analysis are greatly improved, the Kalman filter and the Gaussian process regression algorithm are adopted, data noise is effectively reduced, errors are eliminated, quality and accuracy of data are guaranteed, a solid foundation is provided for subsequent decision making and analysis, the environmental adaptation calibration module is introduced, the system can flexibly adapt to various environmental changes, consistency and accuracy of test data under different environments are guaranteed, the application of the abnormal state prediction module is combined with a Bayesian network and an expected maximization algorithm, a powerful tool is provided for predicting and identifying potential abnormal states, the early warning capability and the risk management efficiency of the system are improved, the structure of a test process is optimized through the application of the graph theory analysis and the Dijkstra algorithm, the efficiency and the coverage rate of the test process are improved, unnecessary test repetition and unnecessary links are reduced, the system has important significance for saving decisions and time, the application of the whole system is provided with the application of the adaptive control tree module, the comprehensive analysis method is improved, the comprehensive adaptation and the comprehensive adaptation performance of the system is guaranteed, the comprehensive adaptation and the comprehensive analysis module is provided for the application of the data is improved, and the comprehensive adaptation and the comprehensive analysis module is provided.
The present invention is not limited to the above embodiments, and any equivalent embodiments which can be changed or modified by the technical disclosure described above can be applied to other fields, but any simple modification, equivalent changes and modification made to the above embodiments according to the technical matter of the present invention will still fall within the scope of the technical disclosure.
Claims (8)
1. The detection control method based on the intelligent detection robot is characterized by comprising the following steps of:
Based on intelligent detection robot data, a linear regression algorithm is adopted to analyze data trend, a deviation analysis method is adopted to compare the data with a standard value in real time, data deviation is identified and quantized, and the data is calibrated and adjusted through a self-adaptive filter, so that noise is reduced, errors are eliminated, and calibration data is generated;
Based on the calibration data, training a support vector machine model by using a support vector machine algorithm, identifying and learning historical test data, optimizing the accuracy and efficiency of a data calibration process and generating an optimized calibration model by using modes and correlations in a calibration result;
based on the optimized calibration model, combining the temperature and humidity changes monitored in real time, applying a neural network algorithm to perform real-time calibration strategy adjustment, and generating environment adaptability calibration data;
based on the environment adaptability calibration data, carrying out prediction and identification of an abnormal state through a Bayesian network model, and generating an abnormal state prediction result;
based on the abnormal state prediction result, a shortest path algorithm in graph theory is applied to construct a graph model of the test flow, the dependency relationship between test links is identified and analyzed, and a test path is calculated to generate an optimized test flow;
based on the optimized test flow, analyzing sample characteristics and real-time data by adopting a decision tree algorithm, automatically adjusting test parameters, selecting a test method, and generating a self-adaptive test result;
based on the self-adaptive test result, combining material science and physical data, performing data integration and feature extraction by using a data fusion technology, performing test strategy optimization by using a machine learning algorithm, optimizing test accuracy and adaptability, and generating a comprehensive optimization test result;
the calibration data comprises calibration values of temperature, pressure and humidity, the optimized calibration model specifically refers to a model based on support vector machine learning, the environment adaptability calibration data comprises response data to temperature change and humidity change factors, the abnormal state prediction result comprises fault points, abnormal indexes and occurrence probability, the optimized test flow comprises a sequence of test steps, a dependence relationship and an optimized path of a test link, the self-adaptive test result comprises an adjustment result of test parameters and a test method effect, and the comprehensive optimization test result comprises a data fusion analysis result, a test strategy optimization scheme and a test efficiency improvement scheme;
Based on the abnormal state prediction result, a shortest path algorithm in graph theory is applied to construct a graph model of a test flow, the dependency relationship between test links is identified and analyzed, a test path is calculated, and the steps for generating an optimized test flow are specifically as follows:
based on the abnormal state prediction result, a graph model of a test flow is constructed by using graph theory analysis and Dijkstra algorithm, a test link is determined as a node, a relation between tests is taken as an edge, connectivity between the nodes is analyzed, and a test flow graph model is generated;
Based on the test flow chart model, a Bellman-Ford algorithm is applied, the dependency relationship between test links is identified, paths among multiple nodes are calculated, the structure of the test flow is optimized, and a test path analysis result is generated;
Based on the analysis result of the test path, a six-sigma method is applied to optimize the test flow, and redundant test steps in the flow are identified and reduced through a DMAIC framework to generate an enhanced test flow;
based on the reinforced test flow, a system dynamic method is applied to process construction, and the test efficiency and coverage rate are analyzed, resource allocation and time management are balanced through simulating the dynamic behavior and interaction of the test flow, so that an optimized test flow is generated;
Based on the self-adaptive test result, combining material science and physical data, carrying out data integration and feature extraction by using a data fusion technology, carrying out test strategy optimization by a machine learning algorithm, optimizing test accuracy and adaptability, and generating a comprehensive optimization test result specifically comprises the following steps:
based on the self-adaptive test result, combining material science and physical data, and carrying out integration of material science and physical data by using principal component analysis to generate a data fusion result;
Based on the data fusion result, adopting K-means clustering analysis to identify groups and modes in the data, applying an Apriori association rule mining algorithm to analyze the relation and interdependence among multiple groups, carrying out data integration and feature extraction, capturing information and features, and generating a feature extraction result;
based on the feature extraction result, carrying out optimization analysis on the test strategy by using a support vector machine algorithm, analyzing the relation between feature data and the test result, optimizing the test strategy, and generating a test strategy optimization analysis result;
Based on the test strategy optimization analysis result, a multi-criterion decision analysis method is adopted to evaluate the test strategy, and the test scheme is weighed and optimized by analyzing the cost benefit, risk evaluation, efficiency and reliability to generate a comprehensive optimization test result.
2. The detection control method based on the intelligent detection robot according to claim 1, wherein based on the intelligent detection robot data, a linear regression algorithm is adopted to analyze data trend, a deviation analysis method is applied to compare the data with a standard value in real time, data deviation is identified and quantized, and the data is calibrated and adjusted through an adaptive filter, so that noise is reduced, errors are eliminated, and the step of generating calibration data is specifically as follows:
based on intelligent detection robot data, analyzing trends in the data set by adopting a linear regression algorithm, fitting trend lines and calculating slopes of the data points, identifying a data change mode, and generating a data trend analysis result;
based on the data trend analysis result, comparing the data points with a preset standard value by using a deviation analysis method, calculating deviation values of the data points, analyzing deviation distribution of a data set, and generating a data deviation recognition result;
based on the data deviation recognition result, a Kalman filter algorithm is applied to filter and eliminate errors of noise in a data set, and noise reduction data are generated;
Based on the noise reduction data, a Gaussian process regression algorithm is adopted to calibrate the data, the data accuracy is optimized, a least square method is applied to adjust errors of the data, system deviation is eliminated, and calibration data is generated.
3. The detection control method based on the intelligent detection robot according to claim 1, wherein based on the calibration data, a support vector machine algorithm is applied to train a support vector machine model, historical test data is identified and learned, patterns and correlations in a calibration result are recognized, accuracy and efficiency of a data calibration process are optimized, and the step of generating an optimized calibration model is specifically as follows:
based on the calibration data, performing model training by adopting a support vector machine algorithm, determining basic classification boundaries and decision rules, analyzing data characteristics and class distribution, and generating a basic support vector machine model;
based on the basic support vector machine model, learning historical test data and a current calibration result by using a deep learning technology to generate a deep learning enhancement model;
based on the deep learning enhancement model, analyzing modes and relations in the calibration result by using an association rule learning method, identifying the relation between data and generating a mode association analysis result;
based on the mode association analysis result, the support vector machine model is optimized, parameters are adjusted, the structure is improved, the accuracy and the efficiency of the calibration process are improved, and an optimized calibration model is generated.
4. The detection control method based on the intelligent detection robot according to claim 1, wherein based on the optimized calibration model, in combination with the temperature and humidity changes monitored in real time, a neural network algorithm is applied to perform real-time calibration strategy adjustment, and the step of generating the environmental adaptive calibration data specifically includes:
Based on the optimized calibration model, a real-time data stream processing technology is adopted, environmental data are collected through a temperature and humidity sensor, and real-time change conditions of temperature and humidity are analyzed to generate environmental monitoring data;
based on the environment monitoring data, adjusting the input parameters of the neural network by using a regression analysis method, adapting to the current environment change, adjusting the weight of the temperature and humidity data and standardizing the input, and generating the adjusted input parameters by reflecting the environment change through the input data;
Based on the adjustment input parameters, performing calibration strategy optimization by using a neural network, performing neural network weight adjustment by using a gradient descent algorithm, and generating an optimized calibration strategy by adapting to environmental changes by using the calibration strategy;
Based on the optimized calibration strategy, the environmental adaptation data is adjusted and refined using a data calibration technique to generate the environmental adaptation calibration data.
5. The detection control method based on the intelligent detection robot according to claim 1, wherein the step of predicting and identifying an abnormal state based on the environmental adaptability calibration data through a bayesian network model, and generating an abnormal state prediction result specifically comprises:
based on the environment adaptability calibration data, a Bayesian network model is constructed, a model structure and probability distribution are defined, network nodes, edges and probability relations are determined, data characteristics and potential relations are represented through the model, and a Bayesian network model structure is generated;
Based on the Bayesian network model structure, an expected maximization algorithm is applied to analyze an abnormal mode in the calibration data, and abnormal state prediction is carried out by calculating the conditional probability of the node to generate a probability inference result;
Based on the probability inference result, adopting principal component analysis and variance analysis to verify the abnormal mode, analyzing the frequency and influence factors of the abnormal mode, and generating an abnormal mode analysis result;
based on the abnormal mode analysis result, the Bayesian decision theory is used for carrying out abnormal state prediction and identification, and probability evaluation and decision making are carried out on the differential abnormal state to generate an abnormal state prediction result.
6. The detection control method based on the intelligent detection robot according to claim 1, wherein based on the optimized test flow, a decision tree algorithm is adopted to analyze sample characteristics and real-time data, automatically adjust test parameters, and select a test method, and the step of generating a self-adaptive test result specifically comprises the following steps:
Based on the optimized test flow, adopting a decision tree algorithm, and combining the sample characteristics and real-time data to analyze the physical characteristics of the sample and the dynamic data in the test process so as to generate a sample characteristic analysis result;
based on the sample characteristic analysis result, adopting an adaptive control theory and a model reference adaptive control method, combining real-time feedback data of the sample characteristic, dynamically adjusting the test rate and the pressure parameter, adapting to the sample characteristic, and generating an automatic adjustment test parameter;
Based on the automatic adjustment test parameters, adopting a genetic algorithm to simulate natural selection and a genetic mechanism, evaluating and selecting a differential test method, analyzing sample characteristics and test targets, selecting a preset test method, and generating a test method selection result;
Based on the test method selection result, a fuzzy logic control method is used, a test parameter is adjusted by combining the test method and a fuzzy set of sample characteristics, and a self-adaptive test result is generated through the matching of the test parameter and the method.
7. The detection control system based on the intelligent detection robot is characterized by comprising a data trend analysis and calibration module, an environment adaptability calibration module, an abnormal state prediction module, a test flow optimization module, an adaptive test module and a comprehensive optimization analysis module according to the detection control method based on the intelligent detection robot of any one of claims 1-6.
8. The detection control system based on the intelligent detection robot according to claim 7, wherein the data trend analysis calibration module analyzes data trend based on the intelligent detection robot data by using a time sequence analysis and a linear regression algorithm, compares data with a standard value by using a deviation analysis method, identifies and quantifies data deviation, and performs noise reduction and error elimination on the data by using a kalman filter and a gaussian process regression algorithm to generate calibration data;
The environment adaptability calibration module is used for carrying out real-time calibration strategy adjustment by utilizing a neural network algorithm based on calibration data and combining with the temperature and humidity changes monitored in real time, adapting to the environment changes and generating environment adaptability calibration data;
The abnormal state prediction module predicts and identifies the abnormal state by adopting a Bayesian network and an expected maximization algorithm based on the environmental adaptability calibration data, and generates an abnormal state prediction result;
the test flow optimization module builds a graph model of the test flow based on the abnormal state prediction result by using graph theory analysis and Dijkstra algorithm, identifies and optimizes the dependency relationship between test links, and generates an optimized test flow;
the self-adaptive test module is based on an optimized test flow, applies a decision tree algorithm and a fuzzy logic control method, analyzes sample characteristics and real-time data, automatically adjusts test parameters and selects a test method, and generates a self-adaptive test result;
The comprehensive optimization analysis module performs optimization analysis of a test strategy based on the self-adaptive test result by combining material science and physical data and applying a data fusion technology and a support vector machine algorithm to generate a comprehensive optimization test result.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202410106551.XA CN117633722B (en) | 2024-01-25 | 2024-01-25 | Detection control method and system based on intelligent detection robot |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202410106551.XA CN117633722B (en) | 2024-01-25 | 2024-01-25 | Detection control method and system based on intelligent detection robot |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN117633722A CN117633722A (en) | 2024-03-01 |
| CN117633722B true CN117633722B (en) | 2024-04-23 |
Family
ID=90029008
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202410106551.XA Active CN117633722B (en) | 2024-01-25 | 2024-01-25 | Detection control method and system based on intelligent detection robot |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN117633722B (en) |
Families Citing this family (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN117826887B (en) * | 2024-03-05 | 2024-06-14 | 深圳鼎晶科技有限公司 | Dynamic feedback pressure control method and system for display module bonding |
| CN118017508B (en) * | 2024-04-10 | 2024-06-07 | 山西艾米德自动化科技有限公司 | Regional power intelligent control method and system based on power distribution |
Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110163233A (en) * | 2018-02-11 | 2019-08-23 | 陕西爱尚物联科技有限公司 | A method of so that machine is competent at more complex works |
| WO2022079278A2 (en) * | 2020-10-16 | 2022-04-21 | Quadsat Aps | Antenna evaluation test system |
| EP4141669A1 (en) * | 2021-08-31 | 2023-03-01 | Siemens Aktiengesellschaft | Method and system for determining optimal computing configuration for executing computing operation |
| CN117171602A (en) * | 2023-10-31 | 2023-12-05 | 广州市林业和园林科学研究院 | Method and system for monitoring biodiversity protection area |
| CN117314389A (en) * | 2023-09-25 | 2023-12-29 | 南京衡臻交通科技有限公司 | Municipal administration management and maintenance system based on internet of things |
| CN117408536A (en) * | 2023-12-15 | 2024-01-16 | 山东省物化探勘查院 | AI-based underwater mapping real-time analysis system |
Family Cites Families (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US10417528B2 (en) * | 2018-02-18 | 2019-09-17 | Sas Institute Inc. | Analytic system for machine learning prediction model selection |
-
2024
- 2024-01-25 CN CN202410106551.XA patent/CN117633722B/en active Active
Patent Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110163233A (en) * | 2018-02-11 | 2019-08-23 | 陕西爱尚物联科技有限公司 | A method of so that machine is competent at more complex works |
| WO2022079278A2 (en) * | 2020-10-16 | 2022-04-21 | Quadsat Aps | Antenna evaluation test system |
| EP4141669A1 (en) * | 2021-08-31 | 2023-03-01 | Siemens Aktiengesellschaft | Method and system for determining optimal computing configuration for executing computing operation |
| CN117314389A (en) * | 2023-09-25 | 2023-12-29 | 南京衡臻交通科技有限公司 | Municipal administration management and maintenance system based on internet of things |
| CN117171602A (en) * | 2023-10-31 | 2023-12-05 | 广州市林业和园林科学研究院 | Method and system for monitoring biodiversity protection area |
| CN117408536A (en) * | 2023-12-15 | 2024-01-16 | 山东省物化探勘查院 | AI-based underwater mapping real-time analysis system |
Non-Patent Citations (3)
| Title |
|---|
| "Application of GA Feature Selection on Naive Bayes, Random Forest and SVM for Credit Card Fraud Detection";Yakub K. Saheed et.al.;《2020 International Conference on Decision Aid Sciences and Application (DASA)》;20210115;全文 * |
| "轮式移动采摘机器人控制与故障检测研究";易泽仁;《万方在线出版》;20230715;全文 * |
| 基于状态指标向量化及霍夫曼编码的变压器缺陷模式识别;袁耀;张曦;杨家辉;邹林;王颂;赵林杰;;广东电力;20180905(08);第174-181页 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN117633722A (en) | 2024-03-01 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN117633722B (en) | Detection control method and system based on intelligent detection robot | |
| US8732100B2 (en) | Method and apparatus for event detection permitting per event adjustment of false alarm rate | |
| Malik et al. | Fuzzy reinforcement learning based intelligent classifier for power transformer faults | |
| Di Maio et al. | Ensemble-approaches for clustering health status of oil sand pumps | |
| CN110571792A (en) | Analysis and evaluation method and system for operation state of power grid regulation and control system | |
| EP2978095A1 (en) | Power system operation | |
| CN111652310B (en) | Storage life prediction method based on Bagging framework SVM relay | |
| CN110757510B (en) | Method and system for predicting remaining life of robot | |
| Abaei et al. | A fuzzy logic expert system to predict module fault proneness using unlabeled data | |
| CN113408659A (en) | Building energy consumption integrated analysis method based on data mining | |
| CN111340269B (en) | Real-time optimization method for process industrial process | |
| CN116227786A (en) | Unmanned aerial vehicle comprehensive efficiency evaluation system | |
| Lutska et al. | Forecasting the efficiency of the control system of the technological object on the basis of neural networks | |
| CN117639602A (en) | Self-adaptive motor running state adjusting method and system | |
| CN117666367A (en) | Remote control method and system for underground coal mine machinery | |
| Ashfahani et al. | Autonomous deep quality monitoring in streaming environments | |
| CN116522065A (en) | Coal mill health degree assessment method based on deep learning | |
| CN116227673A (en) | Self-adaptive temperature early warning method and system based on hydrocracking device | |
| CN114970674A (en) | Time sequence data concept drift adaptation method based on relevance alignment | |
| CN117724348B (en) | Accurate pressure regulation and control system based on explosion testing machine | |
| Angelov et al. | Evolving fuzzy inferential sensors for process industry | |
| Karagiorgou et al. | On making factories smarter through actionable predictions based on time-series data | |
| Verbeke et al. | Fleet-Based Remaining Useful Life Prediction of Safety-critical Electronic Devices | |
| CN118014165B (en) | Traceability management method and traceability management system for lithium ion battery production | |
| CN112099477B (en) | Fault tracing method in lithium ion battery production process |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |