CN105045665A - Operator functional status based cockpit dynamic function allocation method - Google Patents

Abstract

The invention provides an operator functional status based cockpit dynamic function allocation method. The method comprises the steps of: firstly, determining a function permission of an automation level, and carrying out fuzzy modeling; secondly, carrying out fuzzy division on a dual-input single-output variable field, and choosing a suitable fuzzy set and a form and position parameters of the fuzzy set; and finally, determining a specific form of a fuzzy rule, generating the fuzzy rule of a system, and performing defuzzification after fuzzy reasoning settlement. According to the allocation method, by carrying out dynamic function allocation on a man-machine automation system of a cockpit, the shortcomings of conventional static function allocation can be overcome, and a system function allocation policy can be accurately adjusted in real time.

Description

Based on the driving cabin dynamic function-allocation method of operator's functional status

Technical field

The present invention relates to a kind of the Automation Design technology of civil aircraft driving cabin.

Background technology

Function distributes (FunctionAllocation, FA) Objective Concept P.M.Fitts proposed in nineteen fifty-one, refer to the process function (Function) in system or task (Task) point being tasked people or machine, it mainly emphasizes the Function Decomposition between each constituent of system.Aircraft cockpit man-machine function allocation is exactly that the function that will be realized by aircraft cockpit is reasonably distributed between pilot and driving cabin automation subsystem.Reasonably carry out the function/task matching of pilot and automation subsystem, people can be made full use of, functional resources that both machines have, enable the context aware level that pilot remains higher, reduce its work fatigue and operation error simultaneously, automated system advantage can be given full play to again, thus complete aerial mission efficiently.

Traditional function assigning method, by comparing people and machine completes the capacity superiority of specific function, determines that this function distributes to people or machine.This allocation scheme just determines people and machine role separately at system design stage, and does not allow any reallocation when system cloud gray model, is called that static function is distributed.But it is infeasible for determining that the function allocative decision of an optimum static state ensures that the performance of driving cabin system remains at optimum condition.On the one hand, if give more task to pilot, so in the process of executing the task, once run into task environment complicated and changeable, pilot work load may exceed threshold value, causes work efficiency drop, can cause aircraft accident time serious.On the other hand, if give the authority that automation subsystem is higher, under pilot is in monitor state for a long time, situational awareness will be caused to decline, rapid reaction can not be made to emergency.Compare static function to distribute, another kind of function forms of distribution---dynamic function-allocation allows system in operation phase change according to circumstances, systemic-function is dynamically redistributed between people, machine, thus make people, machine place of working coordinates more, improves the functional effect of entire system.At present, dynamic function-allocation is studied in the fields such as traffic control aloft, process control, Military Command and Control system to obtain and is applied widely.

China's large-sized civil transporter project is formally set up the project already, and the large-sized civil transporter that development has an independent intellectual property right needs the support of a large amount of correlation technique, and dynamic function-allocation is one of gordian technique wherein just.In view of the susceptibility of its application background, have no open report about its core technology at present, do not form system and complete method.From the angle of adaptation dynamic change environment, because physiological parameter measurement value can the duty of mirror operation person in real time, the trigger mechanism based on measurement is most suitable.By the measurement to these physiological signals of operator, extract the index of energy mirror operation person load, functional status (the OperatorFunctionalState of evaluating operator real-time and accurately, OFS), according to operator's functional status adjustment System function allocation strategy, make the operating performance state that people and machine reach best, thus do not need the cognitive state estimating operator.

Summary of the invention

In order to overcome the deficiencies in the prior art, the invention provides a kind of driving cabin dynamic function-allocation method based on operator's functional status, namely estimating operator's functional status according to physiological parameter measurement value, thus carrying out the dynamic conditioning of function robotization grade.

The technical solution adopted for the present invention to solve the technical problems comprises the following steps:

1) function privilege corresponding to each robotization grade of civil aircraft driving cabin system is determined;

2) heart rate of operator and EEG signals data are measured, from the extracting data gathered to the characteristic index of functional status sensitivity; Choose HRV ₂and TLI ₂as the two indices of weighing operations person's physiological parameter change, and be used as the input of fuzzy logic; The output of fuzzy logic is the robotization grade of system; Wherein, HRV ₂it is HR standard deviation sigma in a minute _hRVand average value mu _hRVratio; Task load index TLI ₂represent the ratio of the different frequency range power of different EEG electrode, namely fz, Pz represent the collection point of EEG signals in 10-20 international electrode placement standard, P _{θ, Fz}represent the energy intensity of the brain electricity θ section recorded at collection point Fz, P _{α, Pz}represent the energy intensity of the brain electricity α section recorded at collection point Pz;

3) by operator performance parameter HRV ₂, TLI ₂all be divided into five grades, robotization grade LOA be divided into four grades, represent with fuzzy language value, and provide standard fuzzy value scope; Operator performance parameter HRV ₂, TLI ₂the fuzzy subset type corresponding with robotization grade LOA is triangle;

4) by the operator performance parameter HRV of input ₂, TLI ₂set up the fuzzy controller of dual input list output with the robotization grade LOA exported, the rule base of fuzzy controller is R _l: if X is A _iand Y is B _j, then Z is C _k, wherein X, Y are the input language variable of rule, and Z is output language variable, A _iand B _jfor the fuzzy set that dual input variable is corresponding, C _kfor the fuzzy set that single output variable is corresponding, R _lrepresent l article of fuzzy rule, l=1,2 ..., m, i=1,2 ..., n, k=1,2 ..., n, m represent total number of fuzzy rule;

5) adopt max-min gravity model appoach to carry out fuzzy reasoning, draw and export fuzzy quantity C=(A × B) ο R, wherein, $R=\underset{l=1}{\overset{m}{\∪}}{R}_l,R_l=(A_i\×B_j\×C_k);$

6) subset exporting fuzzy quantity is converted into the exact value that OFS is evaluated

The invention has the beneficial effects as follows: dynamic function-allocation is carried out to the man-machine automated system of driving cabin and can overcome the shortcoming that static function in the past distributes, can accurate adjustment System function allocation strategy in real time.

Accompanying drawing explanation

Fig. 1 is civil aircraft driving cabin man-machine dynamic function-allocation adjustment schematic diagram;

Fig. 2 is the membership function schematic diagram of HRV2;

Fig. 3 is the membership function schematic diagram of LI2;

Fig. 4 is the membership function schematic diagram of LOA;

Fig. 5 is dynamic function regulation rule schematic diagram;

Fig. 6 is fuzzy reasoning input, output characteristics curved surface schematic diagram.

Embodiment

Below in conjunction with drawings and Examples, the present invention is further described, the present invention includes but be not limited only to following embodiment.

The present invention includes following steps:

1. determine the function privilege of robotization grade

Merging optimization is carried out to the robotization partition of the level method of the man-machine interactive system that Sheridan and Verplank proposes, determines the function privilege corresponding to each robotization grade of civil aircraft driving cabin system.As shown in table 1, table 2:

Table 1 robotization grade LOA (Modification)

The function privilege of the different robotization grade (LOA) of table 2

2. obscurity model building

Based on the civil aircraft driving cabin dynamic function-allocation of operator's functional status schematic diagram as shown in Figure 1:

To operator's heart rate (HeartRate, HR), brain electricity (Electroencephalogram, EEG) signal data is measured, from the extracting data gathered to the characteristic index of OFS sensitivity, select the input/output variable of fuzzy controller, determine the quantity of input/output variable.Choose HRV ₂, TLI ₂as the two indices of weighing operations person's physiological parameter change, and be used as the input of fuzzy logic; The output of fuzzy logic is the robotization grade (LOA) of system.

HRV ₂be defined as the ratio of HR standard deviation and mean value in one minute period.HRV ₂computing formula as follows:

${\mathrm{HRV}}_2=\frac{{\mathrm{\σ}}_{HRV}}{{\mathrm{\μ}}_{HRV}}---\left(1\right)$

Wherein σ _hRVand μ _hRVbe respectively standard deviation and the average of HR data in 1 minute period.

Task load index TLI2 represents the ratio of the different frequency range power of different EEG electrode.The computing formula of TLI2 is as follows:

${\mathrm{TLI}}_2=\frac{P_{\mathrm{\θ},Fz}}{P_{\mathrm{\α},Pz}}---\left(2\right)$

Wherein Fz, Pz represent the collection point of EEG signals in 10-20 international electrode placement standard, P _{θ, Fz}represent the energy intensity of brain electricity θ section (4-8Hz) recorded at collection point Fz, P _{α, Pz}represent the energy intensity of brain electricity α section (8-13Hz) recorded at collection point Pz.

3. the determination of linguistic variable and fuzzy set

To the fuzzy division of dual input list output variable domain, select form and the location parameter of fuzzy set and the fuzzy set be applicable to.

By operator performance parameter HRV ₂, TLI ₂all be divided into five grades, be expressed as with fuzzy language value: " very low ", " low ", " in ", " height ", " very high "; The robotization grade LOA of output function is sent out and is divided into four grades, be expressed as with fuzzy language value: " low ", " in ", " height ", " very high ".According to each fuzzy set be defined on basic domain, determine the subordinate function of fuzzy subset belonging to precise volume and correspondence thereof.HRV ₂, TLI ₂with the scope of LOA tri-Fuzzy Linguistic Variable, as shown in table 3.

Table 3 Fuzzy Linguistic Variable and fuzzy set thereof

Input, output variable, all adopt the triangular membership being uniformly distributed (each leg-of-mutton central point is uniformly distributed on domain), full crossover (end points of each triangle base is adjacent two leg-of-mutton central points just).Membership function respectively as shown in Figure 2, Figure 3, Figure 4.

4. fuzzy rule structure

Determine the concrete form of fuzzy rule, the fuzzy rule of generation system.The rule base setting up fuzzy controller " Ruo-, " form that dual input list exports is as follows:

R _l: if (X is A _i) and (Y is B _j), then (Z is C _k) l=1,2 ..., m

Wherein X, Y are the input language variable of rule, and Z is output language variable.A _i(i=1,2 ..., n) and B _j(j=1,2 ..., n) be the fuzzy set that dual input variable is corresponding, C _k(k=1,2 ..., n) be the fuzzy set that single output variable is corresponding, R _lrepresent l article of fuzzy rule, m represents total number of fuzzy rule.

Due to HRV ₂and TLI ₂respectively there are 5 fuzzy numbers, so have 25 fuzzy control rules.The specific rules storehouse setting up fuzzy controller " Ruo-, " form that dual input list exports is as follows:

R1: if (HRV ₂vL) and (TIL ₂vL), then (LOA is L);

R2: if (HRV ₂vL) and (TIL ₂l), then (LOA is L);

R3: if (HRV ₂vL) and (TIL ₂m), then (LOA is L);

R4: if (HRV ₂vL) and (TIL ₂h), then (LOA is M);

R5: if (HRV ₂vL) and (TIL ₂vH), then (LOA is H);

R6: if (HRV ₂l) and (TIL ₂vL), then (LOA is L);

R7: if (HRV ₂l) and (TIL ₂l), then (LOA is L);

R8: if (HRV ₂l) and (TIL ₂m), then (LOA is M);

R9: if (HRV ₂l) and (TIL ₂h), then (LOA is H);

R10: if (HRV ₂l) and (TIL ₂vH), then (LOA is VH);

R11: if (HRV ₂m) and (TIL ₂vL), then (LOA is L);

R12: if (HRV ₂m) and (TIL ₂l), then (LOA is M);

R13: if (HRV ₂m) and (TIL ₂m), then (LOA is M);

R14: if (HRV ₂m) and (TIL ₂h), then (LOA is H);

R15: if (HRV ₂m) and (TIL ₂vH), then (LOA is VH);

R16: if (HRV ₂h) and (TIL ₂vL), then (LOA is M);

R17: if (HRV ₂h) and (TIL ₂l), then (LOA is H);

R18: if (HRV ₂h) and (TIL ₂m), then (LOA is H);

R19: if (HRV ₂h) and (TIL ₂h), then (LOA is H);

R20: if (HRV ₂h) and (TIL ₂vH), then (LOA is VH);

R21: if (HRV ₂vH) and (TIL ₂vL), then (LOA is H);

R22: if (HRV ₂vH) and (TIL ₂l), then (LOA is VH);

R23: if (HRV ₂vH) and (TIL ₂m), then (LOA is VH);

R24: if (HRV ₂vH) and (TIL ₂h), then (LOA is VH);

R25: if (HRV ₂vH) and (TIL ₂vH), then (LOA is VH).

Adaptation function distribution regulation rule based on operator's functional status (OFS) can be remembered (FuzzyAssociativeMemory, FAM) by Fuzzy Correlation and be described, as shown in 5 × 5 matrixes in Fig. 5.

5. Fuzzy Logic Reasoning Algorithm

What oneself knew fuzzy controller is input as X=A and Y=B, then according to fuzzy control rule, adopt the max-min gravity model appoach of Mamdani to carry out fuzzy reasoning and can show that exporting fuzzy quantity Z is

C＝(A×B)οR(3)

$R=\underset{l=1}{\overset{m}{\∪}}{R}_l---\left(4\right)$

R _l＝(A _i×B _j×C _k)(5)

Its base attribute is set to: "AND" (and) computing between input variable adopts minimum computing; "or" (or) computing between Different Rule adopts very big computing; Fuzzy implication adopts minimum computing; It is gravity model appoach that de-fuzzy then adopts.

Wherein comprise three kinds of main fuzzy logic operations:

(1) fuzzy AND operation

Fuzzy "AND" (and) be annexation between regular conditional, adopt minimum computing;

μ _A∩B(u)＝min{μ _A(u),μ _B(u)}(6)

(2) fuzzy inclusive-OR operation

Fuzzy "or" (or) is the relation between rule, adopts very big computing:

μ _A∪B(u)＝max{μ _A(u),μ _B(u)}(7)

(3) Fuzzy implication

Fuzzy implication (implication) is the relation between regular conditional and conclusion, adopts minimum computing.

A and B is the fuzzy set be defined on X and Y respectively, and the Fuzzy implication represented by A → B is the special implication relation be defined on X × Y.

$R=A\→B=\underset{X\×Y}{\∫}{\mathrm{\μ}}_A\left(x\right)^{\mathrm{\μ}}_B\left(y\right)/(x,y)---\left(8\right)$

6. de-fuzzy

Gravity model appoach is adopted the fuzzy subset that model exports to be converted into the exact value evaluated OFS.The area that the possibility distrabtion curve of Z and abscissa axis surround is asked the center of gravity of this area, the horizontal ordinate of center of gravity is as the result of ambiguity solution.Formula is:

$Z_0=\frac{{\∫}_A^B{\mathrm{Z\μ}}_C\left(Z\right)dZ}{{\∫}_A^B{\mathrm{\μ}}_C\left(Z\right)dZ}---\left(6\right)$

Fig. 6 is according to 25 fuzzy control rules, the fuzzy inference system input utilizing MATLAB fuzzy toolbox (FuzzyLogicToolBox) to obtain, output characteristics curved surface.

Claims (1)

1., based on a driving cabin dynamic function-allocation method for operator's functional status, it is characterized in that comprising the steps:

1) function privilege corresponding to each robotization grade of civil aircraft driving cabin system is determined;

2) heart rate of operator and EEG signals data are measured, from the extracting data gathered to the characteristic index of functional status sensitivity; Choose HRV ₂and TLI ₂as the two indices of weighing operations person's physiological parameter change, and be used as the input of fuzzy logic; The output of fuzzy logic is the robotization grade of system; Wherein, HRV ₂it is HR standard deviation sigma in a minute _hRVand average value mu _hRVratio; Task load index TLI ₂represent the ratio of the different frequency range power of different EEG electrode, namely fz, Pz represent the collection point of EEG signals in 10-20 international electrode placement standard, P _{θ, Fz}represent the energy intensity of the brain electricity θ section recorded at collection point Fz, P _{α, Pz}represent the energy intensity of the brain electricity α section recorded at collection point Pz;

3) by operator performance parameter HRV ₂, TLI ₂all be divided into five grades, robotization grade LOA be divided into four grades, represent with fuzzy language value, and provide standard fuzzy value scope; Operator performance parameter HRV ₂, TLI ₂the fuzzy subset type corresponding with robotization grade LOA is triangle;

4) by the operator performance parameter HRV of input ₂, TLI ₂set up the fuzzy controller of dual input list output with the robotization grade LOA exported, the rule base of fuzzy controller is R _l: if X is A _iand Y is B _j, then Z is C _k, wherein X, Y are the input language variable of rule, and Z is output language variable, A _iand B _jfor the fuzzy set that dual input variable is corresponding, C _kfor the fuzzy set that single output variable is corresponding, R _lrepresent l article of fuzzy rule, l=1,2 ..., m, i=1,2 ..., n, k=1,2 ..., n, m represent total number of fuzzy rule;

5) adopt max-min gravity model appoach to carry out fuzzy reasoning, draw and export fuzzy quantity C=(A × B) o R, wherein,

$R=\underset{l=1}{\overset{m}{\∪}}{R}_l,R_l=(A_i\×B_j\×C_k);$

6) subset exporting fuzzy quantity is converted into the exact value that OFS is evaluated