CN109741007A - A kind of subject personnel's screening technique of aviation cockpit workload test - Google Patents
A kind of subject personnel's screening technique of aviation cockpit workload test Download PDFInfo
- Publication number
- CN109741007A CN109741007A CN201811516969.9A CN201811516969A CN109741007A CN 109741007 A CN109741007 A CN 109741007A CN 201811516969 A CN201811516969 A CN 201811516969A CN 109741007 A CN109741007 A CN 109741007A
- Authority
- CN
- China
- Prior art keywords
- knowledge
- fuzzy
- data
- rule
- input
- 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.)
- Pending
Links
Landscapes
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
Abstract
A kind of subject personnel's screening technique of aviation cockpit workload test, i.e., subject personnel's screening technique of a kind of aviation cockpit workload test based on fuzzy neural network, steps are as follows: 1: initial knowledge-representation system is obtained according to data sample；2: the pretreatment of input data；After this step, the knowledge-representation system of a discretization is obtained；3: carrying out Reduction of Knowledge and obtain minimum decision set；4: after obtaining the minimum decision set after reduction, being further processed the fuzzy system model promoted；5: defining optimality criterion；By above step, invention achieves the effect screened simultaneously using quantitative data and qualitative data to subject, solve the problems, such as that subject state before experiment starts is inconsistent in aviation human-computer interaction experiment；It can have a good evaluation to the workload of subject personnel, and the perfect experiment flow of aviation human-computer interaction experiment has very high practical value.
Description
Technical field:
The present invention provides a kind of subject personnel's screening technique of aviation cockpit workload test, it is a kind of based on mould
The subject personnel's screening technique for pasting the aviation cockpit workload test of neural network, is in filter out workload
The subject personnel of normal condition.Due to aviation human-computer interaction test for the more demanding of subject, do not require nothing more than subject possess compared with
Good physiological status also requires subject to possess the preferable state of mind, such as suitable workload situations.Good physiological status
It can be realized by the methods of sufficient rest, and workload situations only pass through certain then without preferable adjusting method
Screening technique this inappropriate subject is screened out so that the workload situations of subject personnel are in substantially similar journey
Degree is carried out convenient for experiment.
Background technique:
The core of aviation human-computer interaction experimenter's screening technique is exactly using the qualitative data for being tested personnel and to determine
Data are measured, subject is screened.Separately below from the angle of qualitative data and quantitative data, to why using both data
It is illustrated.
It is exactly directly to be inquired to subject using the method for questionnaire in the most straightforward procedure of the workload confirmation to people,
Qualitative data of the result of inquiry namely described in us.And under certain conditions, the possible accuracy of qualitative data is higher.It is international
Upper mainstream to based on questionnaire form workload measurement table there are three types of, be US National Aeronautics and Space Administration (NASA) respectively
NASA-TLX scale, the SWAT scale of United States Air Force Institute of Aeronautics and Astronautics and the car steering workload level amount of Japan
Table.NASA-TLX scale is up to the present using relatively broad one of workload measurement comprising six dimensions
Degree evaluates workload, is cognitive load, physical load, time requirement, level of performance, level of effort and setback respectively
Degree the experiment has found that the sensibility of the scale is higher, preliminary can test and assess to workload, and can with certain
Reliability.The evaluation and test dimension of SWAT scale not as good as NASA-TLX scale it is extensive, it includes assessment entry include time load, psychology
Effort and three entries of psychological stress load, and three entries are ranked up according to certain significance level, each entry is again
Be divided into it is light, in, weigh three kinds of states, combine the state in 27 that shares in this way, and corresponding 27 kinds of marking states, which has biggish
The specific descriptions of respective entries can be arranged in autonomous operation space according to the particular state of oneself.The car steering work of Japan
Making load level scale includes four kinds of states of workload situations before sleep state, mood, physical condition, work, Mei Gexiang
Mesh has 5 classes.The car steering workload level scale of Japan can describe the preceding different shape of experiment to a certain extent
State, and first two method is more suitable for being suitable for afterwards for the self-appraisal of workload.However in reality, for qualitative parameter
Use is there are two problem, 1) evaluation of qualitative factor grade has very big ambiguity, and individual can not provide accurate value, can only provide
One relatively reasonable range；2) type entry is more in qualitative parameter, needs cautiously to choose reasonable entry to improve measurement
As a result accuracy.
In addition to quantitative parameter, in order to monitor workload situations in real time, experimenter can utilize some external parameters, such as raw
Manage parameter and eye movement parameter, Lai Fanying workload.That is, workload itself is the index for being difficult to quantify, it is real
The personnel of testing must find a ruler, be capable of measuring the size of workload.One possible ruler be exactly some physical signs and
Eye movement index.Physiological parameter includes brain wave (EEG), event related potential (ERP), electromyogram (EMG), photoplethysmographic
(PPG) and breathing (RSP), and eye movement index includes that frequency of wink, wink time, pupil diameter, sight watch track and phase attentively
Close region fixation time.These parameters respectively have with workload and have certain relationship, such as the alpha wave in brain wave exists
Certain reduction is presented in high work load state, and delta wave then will increase；And event correlation point (ERP) is also negative with work
Obvious correlativity is presented in lotus or fatigue, and in brainfag experiment, after 24 hours sleep deprivations, ERP is presented subject
It is decreased obviously.For eye index, studies have shown that pupil dilation is connected with the cognitive load of people, and the frequency expanded with
Task difficulty increases and increases.In simulated flight task, pupil diameter, eyelid aperture is also chosen as commenting for test job load
Estimate, with the increase of load level, both shows as first increasing reducing afterwards, and different load significant difference.And for electrocardio
Index, heart rate variability (HRV) is a reliable evaluation index, more sensitive to workload change.With sympathetic nerve
Arousal level increases, and HRV is reduced with the increase of Mental Workload.And under simulated flight simulated environment, with Mental Workload
Increase, HRV time domain index SDNN, AVNN, RRCV etc. have certain reduction, and its frequency-domain index also has change.
But there is certain risks for single parameter measurement workload, be these parameters are first certain
There can be accurate reflection under scene to workload, scene change be added, then this mapping relations just may no longer be set up；
Followed by different personnel, the effect that these parameters play may vary, that is to say, that for an experimenter some
Parameter may be well suited for measuring its workload situations, and for another experimenter, this parameter may
It is just less accurate to its workload situations of reflection, so needing to carry out for the method for one-parameter prediction work load
It improves.
When using a parameter carry out assessment be unable to satisfy demand when, by these parameter combinations together using be one ratio
Preferable method, this combined method are also referred to as data fusion.In domain of data fusion, there are two types of main data fusion sides
Method, the i.e. data fusion based on data Layer and the data fusion based on functional layer.Data fusion based on data Layer is lowermost layer
Secondary fusion method.This method refers in data processing, directly merges to initial data.In practice, this
The main processing data volume that method uses is very huge, and its calculation method is sufficiently complex, it is difficult to calculate final fusion
Data, therefore and be not frequently used；Data fusion method based on characteristic layer is a kind of higher level fusion method.This method
Refer to the characteristic value for first extracting initial data in data processing, is merged to characteristic value.This method it is main excellent
Point is that, by extracting the feature of data, the primitive character that we remain data is excessively huge without regard to data volume is made,
This brings conveniently for further analysis.Herein, we carry out data fusion in this way.
In a practical situation, in order to describe the state that this qualitative and quantitative is inputted and deposited, based on the data of characteristic layer
Fusion method introduces the fuzzy neural network based on rough set to classify to subject.
(1) summary of the invention:
(1) purpose: the present invention provides a kind of subject personnel's screening technique of aviation cockpit workload test, it is one
Subject personnel's screening technique of aviation cockpit workload test of the kind based on fuzzy neural network, utilizes determining for subject personnel
Amount measurement data and observational measurement data screen subject according to resulting workload situations value.The purpose of screening is
Make to be tested before experiment in relatively uniform state, in order to test development and analysis.
(2) technical solution:
1, qualitative data
As it was noted above, mainly using the third evaluating method as the qualitative evaluating tool of working condition before testing.?
It includes the personal base case such as 5 classifications, including age, height, weight and driving age in table that the driving work-load, which is evaluated and tested,；
The mind & bodies situation such as workload before sleep quality, mood, physical condition, driving；Driving task condition；Road and ring
Border condition.The actual use of the evaluation table is shown, the workload that mind & body situation is caused accounts for overall assessment
13% or more, this is most important accounting factor.In addition, the evaluation and test of the part can reflect participant's work before experiment well
The difference of state, so mainly applying the evaluation and test entry of mind & body status sections in actual use.
2, quantitative data
Described in as discussed above, before using fuzzy neural network, we first have to extract the characteristic parameter of data.Such as
If fruit considers the data type of physical measurement, many available methods that characteristic parameters are extracted from initial data are had.
Herein using the methods experiment for utilizing statistical value.Experiment mainly acquires two kinds of physiological datas, is photoplethysmographic respectively
(PPG) and electrocardio (EMG).In conjunction with chapter 2 experiment output main indicator with above described in workload relationship compared with
For close some values, the measured value EMG-Average of digitlization statement the AVNN parameter and EMG of HRV measured by PPG is chosen
Input value of the parameter as quantitative Treatment data.The mean value at all intervals aroused in interest, Average represent table during AVNN representative record
Show average amplitude in EMG/ECG signal.
3, fuzzy neural network is arranged
In the data fusion method based on characteristic layer, neural network is because of its more quick calculating speed, fault-tolerant ability
Great application is obtained.In addition, neural network also has good classification performance, wanting for this experiment can be met well
It asks.But the method as traditional neural network and support vector machines cannot calculate qualitative input, it is therefore desirable to find new
Tool for processing this problem.This patent proposes the fuzzy neural network of an extension to handle qualitative input.This method
The main practice it is as follows:
A kind of subject personnel's screening technique of aviation cockpit workload test of the present invention, i.e., it is a kind of to be based on fuzzy neural
Subject personnel's screening technique of the aviation cockpit workload test of network, implementation step are as follows:
Step 1: initial knowledge-representation system is obtained according to data sample；This, which represents, needs to establish using mode standard
Comprising inputting the knowledge-representation system with output；
Step 2: the pretreatment of input data；This means that the discretization variable is to obtain for qualitative output and input
To value type, there are many available discretization method, as interval method, blur method, clustering method, based on the discrete of entropy
Change method etc.；For quantitatively outputting and inputting, fuzzy division is carried out using fuzzy decision discrete method, is divided into different obscure
Subset, and corresponding discrete value is assigned to each fuzzy subset, and choosing there is maximum to obscure membership function value attribute value
Fuzzy subset corresponding to discrete value as its discrete value；After this step, the knowledge representation of an available discretization
System；
Step 3: on the basis of discretization knowledge-representation system, needing to carry out Reduction of Knowledge and obtain minimum decision set；For
The smallest decision set of acquisition, needs to handle knowledge-representation system, excludes incompatible rule first and delete redundancy rule；Its
It is secondary for the same terms but there is the knowledge representation entries of different outputs, retain its rule with high confidence level, and
The rule of low confidence level is deleted, to eliminate the attribute for occurring incompatible rule in regular collection；
Step 4: after obtaining the minimum decision set after reduction, using the decision set, with reference to fuzzy reasoning weighted sum method
Fuzzy system model, the fuzzy system model promoted can be further processed:
Y in formulakRepresent output parameter, μkRepresent the confidence level coefficient of kth rule；
As shown in Figure 1, the numerical value of input layer is transmitted directly to middle layer from input layer after treatment in Fig. 1；It is defeated
Two kinds of input can be calculated by entering layer: quantitative input and qualitative input；There are two kinds of neurons to handle respectively not
The input of same type；Wherein, numerical value input uses blur method, and each node passes through Gaussian function as subordinating degree function:
Wherein mkiAnd σkiIt is midpoint and the width of subordinating degree function；For qualitative parameter, subordinating degree function regards normal as
Amount；
The each node of middle layer can be regarded as a multiplier, it calculates the confidence level of kth rule, that is, calculates and determine
Determine the product of the degree of membership of all conditions attribute of kth rule；Output layer is exported its result using corresponding rule；
Step 5: defining optimality criterion；Neural network can optimize iteration to analog value according to this index,
It can stop iterative process after reaching corresponding the number of iterations or required precision；
The essentially inverse convergence rate for propagating (Back Propagation, BP) method is slower, and there are Local Minimums to ask
Topic, therefore we use L-M method (Levenberg-Marquardt) Lai Tigao network training speed；L-M method is under steepest
A kind of method that drop method and Newton method combine not only can provide the speed of Newton method but also guarantee the convergence by steepest descent method；
Its iterative learning method is as follows:
θk+1=θk-[ATA+μkI]-1ATe
Wherein A is Jacobi (Jacobian) matrix, the first differential including network error item relative to weight；Work as ratio
Coefficient μkWhen larger, for this method close to steepest descent method, and when dropping to 0, this method becomes Newton method；Every success when iteration
One step reduces, μkIn this way when close to error, this method is gradually similar to Newton method, and Newton method is when error is smaller
Calculating speed is quickly.
Wherein, " Reduction of Knowledge " described in step 3 refers under conditions of keeping knowledge classification ability constant, deletes it
In uncorrelated and unessential knowledge, it enormously simplifies the complexity of database structure, improves people to lying in data
The awareness of various information under the huge data volume in library.
By above step, invention achieves the effects screened simultaneously using quantitative data and qualitative data to subject
Fruit solves the problems, such as that subject state before experiment starts is inconsistent in aviation human-computer interaction experiment；It can be to subject personnel
Workload have a good evaluation, the perfect experiment flow of aviation human-computer interaction experiment has very high practical value.
(3) advantages of the present invention and effect:
The invention proposes a kind of aviation human-computer interaction experimenter's screening technique based on fuzzy neural network, realizes
Screening based on quantitative data and qualitative data to subject personnel, perfect aviation human-computer interaction experiment flow.Its effect is main
It is following three aspects:
1, quantitative data and qualitative data are used simultaneously, there can be one preferably to comment the workload of subject personnel
It is fixed.
2, the classification method based on fuzzy neural network has been used, can have been obtained in the case where there is sample data set preferably
Classifying quality.
3, the experiment flow of the perfect aviation human-computer interaction experiment of this method, so that the experimental state of subject is in substantially phase
As experimental state.
(2) Detailed description of the invention:
Fig. 1 fuzzy neural network schematic diagram.
Fig. 2 screening technique flow diagram of the present invention.
Serial number, symbol, code name are described as follows in figure:
AVNN represents the NN interval average value of heart rate variability (HRV) measured by photoplethysmographic (PPG)；
EMG-Average represents electrocardiographicdata data average value；X1, X2, X3, X4 represent nerve net network inputs, and Y represents neural network output
(3) specific embodiment:
1, qualitative data
As it was noted above, mainly using the third evaluating method as the qualitative evaluating tool of working condition before testing.?
It includes the personal base case such as 5 classifications, including age, height, weight and driving age in table that the driving work-load, which is evaluated and tested,；
The mind & bodies situation such as workload before sleep quality, mood, physical condition, driving；Driving task condition；Road and ring
Border condition.The actual use of the evaluation table is shown, the workload that mind & body situation is caused accounts for overall assessment
13% or more, this is most important accounting factor.In addition, the evaluation and test of the part can reflect participant's work before experiment well
The difference of state, so mainly applying the evaluation and test entry of mind & body status sections in actual use.
In the experiment that this group carries out, due to being single experiment and guaranteeing preferable sleep state daily, so sleep
Situation and physical condition are essentially a constant.So we only acquire the data of two projects, that is, work before mood and experiment
Make load.Workload has 5 grades before mood and experiment, is indicated with 1,2,3,4,5.Output is simply divided into two
Class, high work load and low workload are indicated with number 0 and 1.For the angle of classification, this is a simple binary
Classification problem.
2, quantitative data
Described in as discussed above, before using fuzzy neural network, we first have to extract the characteristic parameter of data.Such as
If fruit considers the data type of physical measurement, many available methods that characteristic parameters are extracted from initial data are had.
Herein using the methods experiment for utilizing statistical value.Experiment mainly acquires two kinds of physiological datas, is photoplethysmographic respectively
(PPG) and electrocardio (EMG).In conjunction with chapter 2 experiment output main indicator and 3.1 described in workload relationship compared with
For close some values, the measurement average value EMG- of digitlization statement the AVNN parameter and EMG of HRV measured by PPG is chosen
Input value of the Average parameter as quantitative Treatment data.The mean value at all intervals aroused in interest during AVNN representative record,
Average, which is represented, indicates average amplitude in EMG/ECG signal.
Table 1 show 129 groups of data that the verifying present invention tests.The data are derived from the Experiment of Psychology of University Of Augsburg
Data set is more authoritative data source.Wherein, preceding 100 groups of data are used for neural metwork training, and then 29 groups of data are used for
Prediction verifying for neural network.
1 qualitative data of table and quantitative data
3, fuzzy neural network is arranged
In the data fusion method based on characteristic layer, neural network is because of its more quick calculating speed, fault-tolerant ability
Great application is obtained.In addition, neural network also has good classification performance, wanting for this experiment can be met well
It asks.But the method as traditional neural network and support vector machines cannot calculate qualitative input, it is therefore desirable to find new
Tool for processing this problem.Related scholar proposes the fuzzy neural network of an extension to handle qualitative input.The party
The main practice of method is as follows:
A kind of subject personnel's screening technique of aviation cockpit workload test of the present invention, i.e., it is a kind of to be based on fuzzy neural
Subject personnel's screening technique of the aviation cockpit workload test of network, as shown in Figure 2, implementation step is as follows:
Step 1: initial knowledge-representation system is obtained according to data sample.This, which represents, needs to establish using mode standard
Comprising inputting the knowledge-representation system with output.
Step 2: the pretreatment of input data.This means that the discretization variable is to obtain for qualitative output and input
To value type, there are many available discretization method, as interval method, blur method, clustering method, based on the discrete of entropy
Change method etc.；For quantitatively outputting and inputting, fuzzy division is carried out using fuzzy decision discrete method, is divided into different obscure
Subset, and corresponding discrete value is assigned to each fuzzy subset, and choosing there is maximum to obscure membership function value attribute value
Fuzzy subset corresponding to discrete value as its discrete value.After this step, the knowledge representation of an available discretization
System.
Step 3: on the basis of discretization knowledge-representation system, needing to obtain minimum decision set.It is the smallest in order to obtain
Decision set needs to handle knowledge-representation system.Incompatible rule is excluded first and deletes redundancy rule.Secondly for phase
Still there is the knowledge representation entry of different outputs with condition, retain its rule with high confidence level, and delete low confidence level
Rule, to eliminate the attribute for occurring incompatible rule in regular collection
Step 4: after obtaining the minimum decision set after reduction, using the decision set, with reference to fuzzy reasoning weighted sum method
Fuzzy system model, the fuzzy system model promoted can be further processed:
Y in formulakRepresent output parameter, μkRepresent the confidence level coefficient of kth rule.
According to the practice, we can establish fuzzy neural network.The structure of fuzzy neural network is as shown in Figure 1:；
The numerical value of each node of input layer is transmitted directly to middle detection from input layer after treatment in Fig. 1.Input
Layer can calculate two kinds of input: quantitative input and qualitative input.There are two kinds of neurons to handle difference respectively
The input of type.Wherein, numerical value input uses blur method, and each node passes through Gaussian function as subordinating degree function:
Wherein mkiAnd σkiIt is midpoint and the width of subordinating degree function.For qualitative parameter, subordinating degree function regards normal as
Amount.
Each node of middle layer can be regarded as a multiplier, it calculates the confidence level of kth rule, that is, calculate
Determine the product of the degree of membership of all conditions attribute of kth rule.Last decision-making level using corresponding rule by its result into
Row output.
Step 5: defining optimality criterion.Neural network can optimize iteration to analog value according to this index,
It can stop iterative process after reaching corresponding the number of iterations or required precision.
The convergence rate of basic BP algorithm is slower, and there are problems that Local Minimum, therefore we use L-M algorithm
(Levenberg-Marquardt) Lai Tigao network training speed.L-M algorithm is one that steepest descent method and Newton method combine
Kind method not only can provide the speed of Newton method but also guarantee to pass through the convergence of steepest descent method.Its Iterative Algorithm is as follows:
θk+1=θk-[ATA+μkI]-1ATe
Wherein A is Jacobian matrix, the first differential including network error item relative to weight.As proportionality coefficient μkCompared with
When big, for the algorithm close to steepest descent method, and when dropping to 0, algorithm becomes Newton method.Every one step of success when iteration reduces
It is some, μkIn this way when close to error, algorithm is gradually similar to Newton method, and calculating speed is very when error is smaller for Newton method
Fastly.
4, data are analyzed
Using previously described University Of Augsburg's data set, this method can be verified:
(1) when quantitative parameter is only used only, accuracy rate 89.69%.Simultaneous quantitative data are also not more more more prepare,
When we fully enter all related datas in data set, there is sharp fall in accuracy rate, only
69.38%.
(2) when quantitative data and qualitative data is used in combination, accuracy rate rises to 93% or more.
Claims (2)
1. a kind of subject personnel's screening technique of aviation cockpit workload test, i.e., a kind of boat based on fuzzy neural network
Subject personnel's screening technique of empty cockpit workload test, it is characterised in that: implementation step is as follows:
Step 1: initial knowledge-representation system is obtained according to data sample；This represents to need to establish using mode standard
The knowledge-representation system of input and output；
Step 2: the pretreatment of input data；This means that the discretization variable is to be counted for qualitative output and input
Value Types, there are many usable discretization method, such as interval method, blur method, clustering method and based on the discretization of entropy
Method；For quantitatively outputting and inputting, fuzzy division is carried out using fuzzy decision discrete method, is divided into different fuzzy sons
Collection, and corresponding discrete value is assigned to each fuzzy subset, and choose for attribute value with maximum fuzzy membership function value
Discrete value corresponding to fuzzy subset is as its discrete value；After this step, the knowledge-representation system of a discretization is obtained；
Step 3: on the basis of discretization knowledge-representation system, needing to carry out Reduction of Knowledge and obtain minimum decision set；In order to obtain
The smallest decision set is obtained, needs to handle knowledge-representation system, excludes incompatible rule first and delete redundancy rule；Secondly right
In the knowledge representation entry still with the same terms with different outputs, retain its rule with high confidence level, and delete
The rule of low confidence level, to eliminate the attribute for occurring incompatible rule in regular collection；
Step 4: after obtaining the minimum decision set after reduction, using the decision set, with reference to the mould of fuzzy reasoning weighted sum method
System model is pasted, the fuzzy system model promoted can be further processed:
Y in formulakRepresent output parameter, μkRepresent the confidence level coefficient of kth rule；
Input layer can calculate two kinds of input: quantitative input and qualitative input；There are two kinds of neurons to locate respectively
Manage different types of input；Wherein, numerical value input uses blur method, and each node passes through Gaussian function as degree of membership letter
Number:
Wherein mkiAnd σkiIt is midpoint and the width of subordinating degree function；For qualitative parameter, subordinating degree function regards constant as；
Third layer is rules layer；Each node can regard a multiplier as, it calculates the confidence level of kth rule, that is, count
Calculate the product for determining the degree of membership of all conditions attribute of kth rule；4th layer is decision-making level, it means that using accordingly
Rule exports its result；
Step 5: defining optimality criterion；Neural network can optimize iteration to analog value according to this index, reach
It can stop iterative process after corresponding the number of iterations and required precision；
The essentially inverse convergence rate for propagating i.e. BP method is slower, and there are problems that Local Minimum, therefore we use L-M method
To improve network training speed；L-M method is a kind of method that steepest descent method and Newton method combine, and can provide Newton method
Speed guarantee the convergence by steepest descent method again；Its iterative learning method is as follows:
6k+l=θk-[ATA+μkI]-1ATe
Wherein A is Jacobi, that is, Jacobian matrix, the first differential including network error item relative to weight；Work as proportionality coefficient
μkWhen larger, for this method close to steepest descent method, and when dropping to 0, this method becomes Newton method；Every success one when iteration
Step reduces, μkIn this way when close to error, this method is gradually similar to Newton method, and Newton method is counted when error is smaller
Calculate speed quickly.
2. a kind of subject personnel's screening technique of aviation cockpit workload test according to claim 1, i.e., a kind of
Subject personnel's screening technique of aviation cockpit workload test based on fuzzy neural network, it is characterised in that:
" Reduction of Knowledge " described in step 3 refers under conditions of keeping knowledge classification ability constant, deletes wherein uncorrelated
With unessential knowledge, it enormously simplifies the complexity of database structure, improves people to lying in the huge number of database
According to the awareness of the various information under amount.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201811516969.9A CN109741007A (en) | 2018-12-12 | 2018-12-12 | A kind of subject personnel's screening technique of aviation cockpit workload test |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201811516969.9A CN109741007A (en) | 2018-12-12 | 2018-12-12 | A kind of subject personnel's screening technique of aviation cockpit workload test |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN109741007A true CN109741007A (en) | 2019-05-10 |
Family
ID=66359327
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201811516969.9A Pending CN109741007A (en) | 2018-12-12 | 2018-12-12 | A kind of subject personnel's screening technique of aviation cockpit workload test |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN109741007A (en) |
Cited By (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110598789A (en) * | 2019-09-12 | 2019-12-20 | 首都师范大学 | Human fatigue state prediction method and system based on fuzzy perceptron |
| CN111387976A (en) * | 2020-03-30 | 2020-07-10 | 西北工业大学 | Cognitive load assessment method based on eye movement and electroencephalogram data |
| CN111407292A (en) * | 2020-03-30 | 2020-07-14 | 西北工业大学 | Pilot workload assessment method based on eye movement and multi-parameter physiological data information |
| CN111580500A (en) * | 2020-05-11 | 2020-08-25 | 吉林大学 | Evaluation method for safety of automatic driving automobile |
| CN112733772A (en) * | 2021-01-18 | 2021-04-30 | 浙江大学 | Real-time cognitive load and fatigue degree detection method and system in storage sorting task |
Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20140143200A1 (en) * | 2010-11-23 | 2014-05-22 | Novell, Inc. | System and method for determining fuzzy cause and effect relationships in an intelligent workload management system |
| CN104239968A (en) * | 2014-09-02 | 2014-12-24 | 浙江大学 | Short-term load predicting method based on quick fuzzy rough set |
| US20150339587A1 (en) * | 2014-05-21 | 2015-11-26 | Prophetstor Data Services, Inc. | Adaptive fuzzy rule controlling system for software defined storage system for controlling performance parameter |
| CN105631532A (en) * | 2015-12-07 | 2016-06-01 | 江苏省电力公司检修分公司 | Power system load prediction method using fuzzy decision-based neural network model |
| CN106971241A (en) * | 2017-03-17 | 2017-07-21 | 浙江工商大学 | The method that sewage quality data are predicted based on fuzzy neural network |
| CN108074004A (en) * | 2016-11-12 | 2018-05-25 | 华北电力大学（保定） | A kind of GIS-Geographic Information System short-term load forecasting method based on gridding method |
-
2018
- 2018-12-12 CN CN201811516969.9A patent/CN109741007A/en active Pending
Patent Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20140143200A1 (en) * | 2010-11-23 | 2014-05-22 | Novell, Inc. | System and method for determining fuzzy cause and effect relationships in an intelligent workload management system |
| US20150339587A1 (en) * | 2014-05-21 | 2015-11-26 | Prophetstor Data Services, Inc. | Adaptive fuzzy rule controlling system for software defined storage system for controlling performance parameter |
| CN104239968A (en) * | 2014-09-02 | 2014-12-24 | 浙江大学 | Short-term load predicting method based on quick fuzzy rough set |
| CN105631532A (en) * | 2015-12-07 | 2016-06-01 | 江苏省电力公司检修分公司 | Power system load prediction method using fuzzy decision-based neural network model |
| CN108074004A (en) * | 2016-11-12 | 2018-05-25 | 华北电力大学（保定） | A kind of GIS-Geographic Information System short-term load forecasting method based on gridding method |
| CN106971241A (en) * | 2017-03-17 | 2017-07-21 | 浙江工商大学 | The method that sewage quality data are predicted based on fuzzy neural network |
Non-Patent Citations (3)
| Title |
|---|
| XIN LU ETAL.: "A human workload monitoring method considering qualitative and quantitative data fusion", 《2017 SECOND INTERNATIONAL CONFERENCE ON RELIABILITY SYSTEMS ENGINEERING (ICRSE)》 * |
| 田景文 等: "《人工神经网络算法研究及应用》", 31 July 2006, 北京理工大学出版社 * |
| 邵莹: "基于神经网络的电力系统短期负荷预测研究", 《中国优秀博硕士学位论文全文数据库 (硕士) 工程科技Ⅱ辑》 * |
Cited By (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110598789A (en) * | 2019-09-12 | 2019-12-20 | 首都师范大学 | Human fatigue state prediction method and system based on fuzzy perceptron |
| CN111387976A (en) * | 2020-03-30 | 2020-07-10 | 西北工业大学 | Cognitive load assessment method based on eye movement and electroencephalogram data |
| CN111407292A (en) * | 2020-03-30 | 2020-07-14 | 西北工业大学 | Pilot workload assessment method based on eye movement and multi-parameter physiological data information |
| CN111387976B (en) * | 2020-03-30 | 2022-11-29 | 西北工业大学 | Cognitive load assessment method based on eye movement and electroencephalogram data |
| CN111580500A (en) * | 2020-05-11 | 2020-08-25 | 吉林大学 | Evaluation method for safety of automatic driving automobile |
| CN111580500B (en) * | 2020-05-11 | 2022-04-12 | 吉林大学 | Evaluation method for safety of automatic driving automobile |
| CN112733772A (en) * | 2021-01-18 | 2021-04-30 | 浙江大学 | Real-time cognitive load and fatigue degree detection method and system in storage sorting task |
| CN112733772B (en) * | 2021-01-18 | 2024-01-09 | 浙江大学 | Method and system for detecting real-time cognitive load and fatigue degree in warehouse picking task |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN109741007A (en) | A kind of subject personnel's screening technique of aviation cockpit workload test | |
| Heard et al. | A survey of workload assessment algorithms | |
| CN108052973B (en) | Map symbol user interest analysis method based on multiple items of eye movement data | |
| CN102293656B (en) | Emotional stability evaluation system based on magnetic resonance imaging and evaluation method thereof | |
| Kuravsky et al. | Mathematical foundations of flight crew diagnostics based on videooculography data | |
| CN112927782B (en) | Heart health state early warning system based on text emotion analysis | |
| Li et al. | A novel method for the evaluation of fashion product design based on data mining | |
| Long et al. | Drawings as a window into developmental changes in object representations | |
| CN114154592A (en) | Flight training quality and efficiency evaluation method fusing physiological data and flight parameter data | |
| Boonyanusith et al. | Blood donor classification using neural network and decision tree techniques | |
| Wang et al. | Cognitive Load Identification of Pilots Based on Physiological‐Psychological Characteristics in Complex Environments | |
| CN109976530A (en) | A kind of VR system cognizant load quantitative estimation method | |
| Zou et al. | Integrating biometric sensors, VR, and machine learning to classify EEG signals in alternative architecture designs | |
| CN114693170A (en) | Multi-source parameter fusion-based psychological assessment method for parachuting training of pilots | |
| CN109508653A (en) | A kind of subjective and objective individual combat Emotion identification method merged based on EEG signals with psychology | |
| CN117547270A (en) | Pilot cognitive load feedback system with multi-source data fusion | |
| Krzyżaniak-Luiza et al. | Validity and reliability of single-item self-report measures of meaning in life and satisfaction with life | |
| Chen et al. | Evaluating aircraft cockpit emotion through a neural network approach | |
| CN111887803B (en) | Multi-dimensional monitoring and evaluating system for artificial efficiency of aircraft cockpit | |
| Jiang et al. | Mental workload artificial intelligence assessment of pilots’ EEG based on multi-dimensional data fusion and LSTM with attention mechanism model | |
| CN108763459A (en) | Professional trend analysis method and system based on psychological test and DNN algorithms | |
| CN114783552A (en) | Method and system for detecting fatigue of operator before post | |
| Huang et al. | Modeling and analysis of fatigue detection with multi-channel data fusion | |
| Gao et al. | Mental Health Evaluation of College Students Based on Similar Trajectory Clustering Algorithm | |
| Huang et al. | Fatigue detection and analysis based on multi-channel data fusion |
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 | ||
| RJ01 | Rejection of invention patent application after publication |
Application publication date: 20190510 |
|
| RJ01 | Rejection of invention patent application after publication |