A dynamic human-machine function allocation method for manned submersibles based on non-cooperative game

Abstract

The present invention discloses a dynamic human-machine function allocation method for a manned submersible based on non-cooperative game, which belongs to the technical field of manned submersibles. The method introduces the non-cooperative game method into the field of human-machine function allocation, transforms the multi-objective optimization model of mental load and situational awareness under the allocation strategy into a non-cooperative game model and forms a mapping relationship, with mental load and situational awareness as the two parties of the non-cooperative game; based on the fuzzy clustering algorithm, the allocable functions are assigned to the game parties, and form an allocation strategy combination with the non-allocable functions, and constructs a utility matrix in combination with the game utility function. The optimal allocation strategy combination is obtained through Nash equilibrium analysis of the utility matrix, and then the allocation strategy is dynamically adjusted according to the change of the alertness of the submariner during the work process; the method has the characteristics of high allocation rationality and can effectively improve the human reliability of the submariner when driving the manned submersible.

Description

Technical Field

[0001] The present invention relates to the technical field of manned submersibles, and in particular to a method for dynamic human-machine function allocation of manned submersibles based on non-cooperative game. Background Art

[0002] A manned submersible is a submersible capable of underwater observation and operations. It is primarily used for underwater surveys, seabed exploration, development, salvage, and rescue missions, and can serve as a base for divers' underwater activities. Manned submersibles, particularly deep-sea manned submersibles, are at the forefront and commanding heights of marine development. Their capabilities reflect a country's comprehensive scientific and technological strength in materials, control, oceanography, and other fields. They can perform a variety of complex tasks, including surveying seabed resources through video and photography, deploying underwater equipment, and inspecting submarine cables and pipelines.

[0003] As the core of the man-machine-environment system in a manned submersible, the submariner is responsible for the important tasks of driving and operating the submersible. In addition, the submariner needs to monitor the system status, formulate and adjust the diving strategy according to the seabed topography, and complete the diving mission. Under time pressure, safety threats (complex deep-sea topography) and environmental factors (abnormal temperature, vibration, acceleration, noise), the submariner needs to complete complex cognitive activities and bear a high mental workload. When the submariner is under high mental workload for a long time, under continuous mental stress or engaged in monotonous and boring work for a long time, it is easy to cause psychological fatigue, reduced situational awareness and reduced execution ability, which greatly increases the probability of the submariner making wrong decisions during the mission. Therefore, how to reasonably formulate a human-machine system function allocation strategy for manned submersibles based on the work characteristics of the submariner has become one of the key issues that need to be solved in the human-machine function allocation of manned submersibles.

[0004] Existing research primarily uses single-objective optimization based on factors such as mental workload, workload, and system performance, or multi-objective optimization through the construction of weighted evaluation functions. However, single-objective optimization alone cannot adapt to the contradictions and conflicts between multiple indicators in actual function allocation. Furthermore, characterizing the inter-coupling relationships between multiple optimization objectives through empirical weighting and the construction of evaluation functions can lead to subjectivity and instability in the optimization solution. Furthermore, compared to static function allocation, dynamic functions help maintain situational awareness and manual skills, keeping overall human cognition at a reasonable level.

[0005] Therefore, there is an urgent need to design a game analysis method that can objectively deal with problems with conflict factors to dynamically allocate human-machine functions to manned submersibles to solve the problems existing in the above-mentioned existing technologies. Summary of the Invention

[0006] In response to the above-mentioned problems, the present invention aims to provide a dynamic human-machine function allocation method for manned submersibles based on non-cooperative games. This method transforms the multi-objective optimization model of mental workload and situational awareness under the allocation strategy into a non-cooperative game model and forms a mapping relationship. Mental workload and situational awareness serve as the two parties of the non-cooperative game. Based on the fuzzy clustering algorithm, the allocable functions are assigned to the game parties, and constitute an allocation strategy combination with the non-allocable functions. The utility matrix is constructed in combination with the game utility function. The optimal allocation strategy combination is obtained through Nash equilibrium analysis of the utility matrix. The allocation strategy is then dynamically adjusted according to the changes in the alertness of the submariner during work. The method has the characteristics of high allocation rationality and can effectively improve the human reliability of the submariner when driving a manned submersible.

[0007] In order to achieve the above object, the technical solution adopted by the present invention is as follows:

[0008] A method for dynamic human-machine function allocation of a manned submersible based on non-cooperative game, comprising the steps of

[0009] S1: Conduct a preliminary functional analysis of the functions involved in manned submersibles, and divide the submersible functions into four parts: information acquisition, information analysis, decision generation, and action implementation;

[0010] S2: Use the function matrix to analyze the submersible's functions, identifying the parts that humans are good at completing, the parts that machines are good at completing, and the parts that both humans and machines are good at completing. The parts of the functions that humans and machines are good at completing are classified as assignable functions, and the opposite functions are classified as non-assignable functions.

[0011] S3: Perform a hierarchical analysis of the parts of the allocable functions that both humans and machines are good at completing, and determine the allocation strategy for the allocable functions;

[0012] S4: Perform single-objective optimization on the mental workload player and the situational awareness player under the allocation strategy to obtain their optimal allocation strategy combinations; calculate the impact factor ξ of each allocable function on the mental workload player in the optimal allocation strategy combination of the mental workload player. iW The influence factor ξ of each allocable function on the situational awareness player in the optimal allocation strategy combination of the situational awareness player is iSA ;

[0013] S5: Calculate the game utility matrix based on the initial samples obtained in step S2 and step S4;

[0014] S6: Analyze the Nash equilibrium of the game utility matrix and obtain the optimal allocation strategy;

[0015] S7: When the alertness of the pulse wave signal detection drops to the set limit, the human-machine function is triggered to perform non-cooperative game redistribution.

[0016] Preferably, in step S4, the single-objective optimization is performed on the mental workload player and the situational awareness player under the allocation strategy to obtain their optimal allocation strategy combinations, thereby obtaining the influence factor ξ of each allocable function on the mental workload player in the optimal allocation strategy combination of the mental workload player. iW The influence factor ξ of each allocable function on the situational awareness player in the optimal allocation strategy combination of the situational awareness player iSA The specific process includes

[0017] S41: Perform single-objective optimization on the mental workload player and the situational awareness player, calculate the mental workload value of the allocation strategy combination, and the minimum value corresponds to the allocation strategy combination with the optimal mental workload. Then calculate the situational awareness value of the function allocation strategy combination, and the maximum value corresponds to the allocation strategy combination with the optimal situational awareness.

[0018] S42: Calculate the mental load occupied by each allocable function during the execution of the optimal allocation strategy combination of mental load using the VACP scale, and use this as the influence factor of each allocable function on the mental load game player ξ iW ;

[0019] S43: Select experts to use the AHP method to obtain the relative impact of each assignable function on situational awareness in the optimal allocation strategy combination of situational awareness, and take the average value of the relative impact as the impact factor ξ of each assignable function on the situational awareness game party iSA ;

[0020] ζ i ={ξ iW ,ξ iSA}

[0021] Where, ζ i is the set of factors that influence the optimization objective of the assignable function i.

[0022] Preferably, in step S41, the mental workload value of the function allocation strategy combination is calculated using the VACP scale; and at the same time, the situational awareness value of the function allocation strategy combination is calculated using the situational awareness comprehensive assessment technology.

[0023] Preferably, the situational awareness utility function of the situational awareness comprehensive assessment technology is:

[0024]

[0025]

[0026]

[0027] Where n is the number of questions and K is the number of subjects.

[0028] Preferably, the process of calculating the mental workload value of the function allocation strategy combination based on the VACP scale includes:

[0029] (1) Decompose the execution process of the function allocation strategy combination of a manned submersible in a certain working phase into several subtasks;

[0030] (2) The VACP scale is then used to determine the load value of each subtask. The average mental load value within the time period is used as the measurement standard for the mental load of the functional allocation strategy combination. The mental load game utility function is obtained as shown in the following formula:

[0031]

[0032]

[0033] Where W t W is the mental workload at a certain time t during the execution of the function allocation strategy combination; ab is the mental workload of subtask a in channel b during the execution process at a certain moment; W is the average mental workload during the time period T.

[0034] Preferably, the calculation process of the game utility matrix described in step S5 includes:

[0035] (1) Perform fuzzy clustering on the initial samples obtained in step S4 to achieve a strategy set diversity of assignable functions;

[0036] (2) The allocable functions of the strategy set and the non-allocable functions described in step S2 together constitute an allocation strategy combination, and then the allocation strategy combination is used to form a game utility matrix through the calculation of the mental workload game player utility function and the situational awareness game player function.

[0037] Preferably, the step S5 of performing fuzzy clustering on the initial samples obtained in step S4 to achieve the strategy set diversity of the assignable functions includes:

[0038] S51: Utility Dimension Processing

[0039] Since the utility dimensions of the two game players are different, dimensional processing is required;

[0040]

[0041] Where,

[0042] After range transformation, it is shown as follows

[0043]

[0044] S52: Establishing fuzzy similarity matrix

[0045] Establish the fuzzy similarity matrix R=(r ij ) n×n , r ij ξ″ ik and ξ″ jk The similarity is calculated by the absolute distance method as follows:

[0046]

[0047] In the formula, M is the similarity correction coefficient. When r ij ∈[0, 1];

[0048] S53: Establishing fuzzy equivalence matrix

[0049] The transitive closure matrix R is established by using the square self-synthesis method for the fuzzy similarity matrix R * , the fuzzy equivalent matrix of R, and satisfies is a Boolean operator;

[0050] S54: Strategic Clustering

[0051] According to the clustering requirements, select the cutoff number λ to classify the matrix, satisfying In order to obtain different cut-off sets, the strategic attribution of the assignable function to the game players is finally achieved by clustering the elements of the same cut-off set.

[0052] Preferably, the process of analyzing the Nash equilibrium of the game utility matrix and obtaining the optimal allocation strategy in step S6 includes:

[0053] S61: Given the total strategy space, utility function, constraints and iteration accuracy ε,

[0054] The constraints are:

[0055] W i ∈{W′ i}

[0056] SA i ∈{SA′ i}1≤i≤n

[0057] W tV ≤7

[0058] W tA ≤7

[0059] W tC ≤7

[0060] W tP ≤7

[0061] SA min ≥|SA|

[0062] Where W and SA are the mental workload function and situational awareness function under the human-machine function allocation strategy in a certain working phase of a manned submersible, respectively; W i with SA i are the mental workload and situational awareness of function i in a certain working phase of manned submersible; W′ i with SA′ i are the mental workload and situational awareness of function i in a certain working stage of a manned submersible when a feasible allocation strategy is adopted; n is the number of functions in a certain working stage of a manned submersible; W tV , W tA W tC , W tP are the sum of the visual channel load, the auditory channel load, the cognitive channel load, and the motor channel load at a certain moment; SA min and |SA| are the minimum situational awareness value and the specified minimum safety threshold of the combination of human-machine function allocation strategy in a certain working phase of a manned submersible, respectively;

[0063] S62: After fuzzy clustering of the initial sample of assignable functions described in step S5, the strategy sets S1, S2, ..., S belonging to each player are obtained. i ,…,S m ;

[0064] S63: Game analysis initialization

[0065] In the strategy space S={S1,S2,…,S i ,…,S m Randomly generate the initial feasible strategy combination of the game

[0066] S64: Note For the feasible initial strategy combination s (0) Medium The initial feasible strategy combination adopted by all other players except i (s),…,u m (s) is the optimization goal, while maintaining Under the condition that the strategy sets S1, S2, ..., S belonging to each player remain unchanged, i ,…,S m Perform corresponding single-objective optimization in

[0067] That is, for any player i, in its corresponding strategy set S i Find the optimal strategy Game Utility And satisfy the constraints where k = 1, 2, 3, ..., q;

[0068] S65: Order Calculate the two strategy combinations s before and after (0) With s (1) Whether the distance between them satisfies the convergence criterion ||s (1) -s (0) ||≤ε, if satisfied, the game ends; if not satisfied, s (0) Replace with s (1) , go to step S3 to perform loop calculation; repeat this iteration until the accuracy requirement is met.

[0069] The beneficial effects of the present invention are as follows: the present invention discloses a method for dynamic human-machine function allocation of a manned submersible based on non-cooperative game. Compared with the prior art, the present invention has the following improvements:

[0070] The present invention proposes a dynamic human-machine function allocation method for manned submersibles based on non-cooperative game theory. This method transforms a multi-objective optimization model of mental workload and situational awareness under the allocation strategy into a non-cooperative game model and forms a mapping relationship. Mental workload and situational awareness serve as the two parties in the non-cooperative game. Based on a fuzzy clustering algorithm, the allocable functions are assigned to the game parties and form an allocation strategy combination with the non-allocable functions. A utility matrix is constructed in combination with the game utility function. The optimal allocation strategy combination is obtained through Nash equilibrium analysis of the utility matrix. The allocation strategy is then dynamically adjusted according to changes in the submersible's alertness during work. During use:

[0071] 1. This method introduces non-cooperative game theory into the field of human-machine function allocation. The two players are not affected by weights, which reduces the subjective influence of humans. It provides a feasible solution to other discrete multi-objective optimization problems in the field of function allocation.

[0072] 2. This method uses mental workload and situational awareness as the game parties and the alertness of pulse wave signal detection as the trigger mechanism. The proposed dynamic function allocation method has the advantages of high allocation rationality and can effectively improve the human reliability of submariners when driving manned submersibles. BRIEF DESCRIPTION OF THE DRAWINGS

[0073] Figure 1 A schematic flow chart of the dynamic human-machine function allocation optimization model for a manned submersible considering the trigger mechanism provided by the present invention.

[0074] Figure 2 This is a flowchart of the Nash equilibrium strategy solution provided by the present invention.

[0075] Figure 3This is a conversion relationship diagram between functions and subtasks provided by the present invention.

[0076] Figure 4 This is a diagram of the human-machine function allocation model for a manned submersible based on non-cooperative game provided by the present invention.

[0077] Figure 5 This is a mapping relationship diagram between the multi-objective optimization and non-cooperative game model provided by the present invention. DETAILED DESCRIPTION

[0078] In order to enable those skilled in the art to better understand the technical solution of the present invention, the technical solution of the present invention is further described below in conjunction with the accompanying drawings and embodiments.

[0079] Example 1: Refer to the attached Figure 1-5 The method for dynamic human-machine function allocation of a manned submersible based on non-cooperative game is shown, comprising the steps of

[0080] S1: Conduct a preliminary functional analysis of the functions involved in manned submersibles, and divide the submersible functions into four parts: information acquisition, information analysis, decision generation, and action implementation;

[0081] For example, life support functions include information acquisition, information analysis, decision generation, and action implementation; driving functions include action implementation;

[0082] S2: Use a functional matrix to further analyze the submersible's functions, identifying the parts that humans excel at, the parts that machines excel at, and the parts that both humans and machines excel at. Factors for evaluation include accuracy, speed, and reliability. If a function contains parts that both humans and machines excel at, it is an assignable function; otherwise, it is a non-assignable function.

[0083] S3: Analyze the automation level of the parts of the assignable functions that both humans and machines are good at completing, so as to determine the allocation strategy of the assignable functions. The automation level classification is shown in Table 1;

[0084] Table 1: Automatic classification

[0085]

[0086] S4: Perform single-objective optimization on the mental workload player and the situational awareness player under the allocation strategy to obtain their optimal allocation strategy combinations; calculate the impact factor ξ of each allocable function on the mental workload player in the optimal allocation strategy combination of the mental workload player. iW The influence factor ξ of each allocable function on the situational awareness player in the optimal allocation strategy combination of the situational awareness player is iSA ;

[0087] S41: Perform single-objective optimization on both the mental workload and situational awareness players. The VACP scale is used to calculate the mental workload value of the functional allocation strategy combination. The minimum value corresponds to the allocation strategy combination with the optimal mental workload. The Situational Awareness Assessment Global Technique (SAGAT) is used to calculate the situational awareness value of the functional allocation strategy combination. The maximum value corresponds to the allocation strategy combination with the optimal situational awareness.

[0088] The situational awareness utility function for calculating the situational awareness value of the function allocation strategy combination based on the situational awareness comprehensive assessment technology (SAGAT) is shown as follows:

[0089]

[0090]

[0091]

[0092] Where n is the number of questions and K is the number of subjects;

[0093] The calculation process of the mental workload value of the function allocation strategy combination based on the VACP scale includes:

[0094] (1) Decompose the execution process of the function allocation strategy combination of a manned submersible in a certain working phase into several subtasks;

[0095] (2) The VACP scale is then used to determine the load value of each subtask. The average mental load value within the time period is used as the measurement standard for the mental load of the functional allocation strategy combination. The mental load game utility function is obtained as shown in the following formula:

[0096]

[0097]

[0098] Where W t W is the mental workload at a certain time t during the execution of the function allocation strategy combination; ab is the mental workload of subtask a in channel b during the execution process at a certain moment; W is the average mental workload during the time period T;

[0099] S42: Calculate the mental load occupied by each allocable function during the execution of the optimal allocation strategy combination of mental load using the VACP scale, and use this as the influence factor of each allocable function on the mental load game player ξ iW , this step is the same as the step of calculating the mental workload value of the function allocation strategy combination in S41;

[0100] S43: Select experts to use the AHP method to obtain the relative impact of each assignable function on situational awareness in the optimal allocation strategy combination of situational awareness, and take the average value of the relative impact as the impact factor ξ of each assignable function on the situational awareness game party iSA ;

[0101] ζ i ={ξ iW ,ξ iSA}

[0102] Where, ζ i is the set of factors affecting the optimization objective of the assignable function i;

[0103] S5: Calculate the game utility matrix based on the initial samples obtained in steps S2 and S4

[0104] (1) Perform fuzzy clustering on the initial samples obtained in step S4 to achieve the strategy set diversity of the assignable functions. The specific steps are as follows:

[0105] S51: Utility Dimension Processing

[0106] Since the utility dimensions of the two game players are different, dimensional processing is required to eliminate the influence;

[0107]

[0108] Where,

[0109] After range transformation, it is shown as follows

[0110]

[0111] S52: Establishing fuzzy similarity matrix

[0112] Establish the fuzzy similarity matrix R=(r ij ) n×n , where r ij ξ″ ik and ξ″ jk The similarity is calculated by the absolute distance method as follows:

[0113]

[0114] In the formula, M is the similarity correction coefficient. When r ij ∈[0, 1];

[0115] S53: Establishing fuzzy equivalence matrix

[0116] The transitive closure matrix R is established by using the square self-synthesis method for the fuzzy similarity matrix R* , that is, the fuzzy equivalent matrix of R, and satisfies is a Boolean operator;

[0117] S54: Strategic Clustering

[0118] According to the clustering requirements, select the cutoff number λ to classify the matrix, satisfying In order to obtain different cut-off sets, the strategic attribution of the assignable function to the game players is finally achieved by clustering the elements of the same cut-off set.

[0119] (2) The allocable functions that complete the game player's attribution and the non-allocable functions described in step S2 are combined to form an allocation strategy combination. The allocation strategy combination is then used to calculate the mental workload game player utility function and the situational awareness game player function to form a game utility matrix, as shown in the following table:

[0120] Table 2: Utility matrix of non-cooperative game model

[0121]

[0122] In the table, P1 and P2 are the mental workload player and situational awareness player respectively; utility and utility The mental workload utility and situational awareness utility are respectively when player P1 adopts n combination strategies for the allocable functions, player P2 adopts m combination strategies for the allocable functions, and a fixed strategy z for the non-allocable functions.

[0123] S6: Analyze the Nash equilibrium of the game utility matrix and obtain the optimal allocation strategy. The specific steps are as follows:

[0124] S61: Given the total policy space, utility function, constraints and iteration accuracy ε, the constraints are:

[0125] W i ∈{W′ i}

[0126] SA i ∈{SA′ i}1≤i≤n

[0127] W tV ≤7

[0128] W tA ≤7

[0129] W tC ≤7

[0130] W tP ≤7

[0131] SAmin ≥|SA|

[0132] Where W and SA are the mental workload function and situational awareness function under the human-machine function allocation strategy in a certain working phase of a manned submersible, respectively; W i with SA i are the mental workload and situational awareness of function i in a certain working phase of manned submersible; W′ i with SA′ i are the mental workload and situational awareness of function i in a certain working stage of a manned submersible when a feasible allocation strategy is adopted; n is the number of functions in a certain working stage of a manned submersible; W tV ,W tA ,W tC , W tP are the sum of the visual channel load, the auditory channel load, the cognitive channel load, and the motor channel load at a certain moment; SA min and |SA| are the minimum situational awareness value of the combination of human-machine function allocation strategies in a certain working phase of a manned submersible and the prescribed minimum safety threshold, respectively. To ensure the safety of the submariner, the minimum situational awareness value of the combination of human-machine function allocation strategies in a certain working phase of a manned submersible should not be lower than the prescribed minimum safety threshold;

[0133] S62: After fuzzy clustering of the initial sample of assignable functions described in step S5, the strategy sets S1, S2, ..., S belonging to each player are obtained. i ,…,S m ;

[0134] S63: Game analysis initialization, that is, in the strategy space S = {S1, S2, ..., S i ,…,S m Randomly generate the initial feasible strategy combination of the game

[0135] S64: Note For the feasible initial strategy combination s (0) Medium The initial feasible strategy combination adopted by all other players except i (s),…,u m (s) is the optimization goal, while maintaining Under the condition that the strategy sets S1, S2, ..., S belonging to each player remain unchanged, i ,…,S m In the corresponding single-objective optimization, for any player i, in its corresponding strategy set S i Find the optimal strategy Game Utility And satisfy the constraints where k = 1, 2, 3, ..., q;

[0136] S65: Order Calculate the two strategy combinations s before and after (0) With s (1) Does the distance between them (the norm of the matrix) satisfy the convergence criterion ||s (1) -s (0) ||≤ε (ε is an arbitrarily small positive number). If satisfied, the game ends; if not satisfied, s (0) Replace with s (1) , go to step S3 to perform loop calculation; repeat this iteration until the accuracy requirement is met.

[0137] S7: When the alertness of pulse wave signal detection drops to the set limit, it will trigger a non-cooperative game redistribution of human-machine functions, so that the human-machine function allocation result is consistent with low mental workload and high situational awareness at this fatigue level.

[0138] Example 2: Different from the above examples, in order to verify the effectiveness of the allocation method described in Example 1, a case study of the navigation phase of a manned submersible in this example is designed for verification:

[0139] 1. Design and research conditions are shown in Tables 3 to 7:

[0140] Table 3: Analysis of functional allocation strategy for manned submersible during navigation phase

[0141]

[0142]

[0143] Table 4: Subtask analysis table of the navigation process

[0144]

[0145] Table 5: Mental workload of functional allocation combinations

[0146]

[0147]

[0148] Note: The numerical codes in the table represent the allocation strategy for the corresponding allocable function during information acquisition, information analysis, decision making, or action execution. 1 represents machine, 2 represents human-machine, and 3 represents human. For example, 1, 1, 1, 1 represents the allocation of life support action execution to the machine, propulsion monitoring information analysis to the machine, collision avoidance decision making to the machine, and piloting action execution to the machine. The calculation of mental workload includes unallocable functions during the navigation phase. The fixed allocation strategy for communication, navigation, and detection functions is human-machine operation.

[0149] Table 6: Questionnaire 1 for the global assessment of situational awareness

[0150]

[0151] Table 7: Situational awareness by functional allocation combination

[0152]

[0153]

[0154] Note: The calculation of situational awareness includes the non-assignable functions during the navigation phase, and the fixed allocation strategy of communication function, navigation function and detection function is for human-machine operation.

[0155] 2. Construction of Game Utility Matrix

[0156] (1) Single-objective optimization and calculation of impact factor indicators

[0157] The optimal allocation strategy combination for mental workload is 1, 1, 1, 1 in Table 5. The mental workload influencing factors of each assignable function are:

[0158] ξ W ={ξ 1W ,ξ 2w ,ξ 3W ,ξ 4W}={0.124, 0.164, 0.406, 0.036}.

[0159] The optimal allocation strategy combination for situational awareness is 1, 1, 2, 1 as shown in Table 7. Three experts involved in manned submersible design were invited to score. The situational awareness impact factor for each assignable function is:

[0160] ξ SA ={ξ 15A ξ 2SA ,ξ 3SA ,ξ 4SA}={0.139, 0.139, 0.583, 0.139}.

[0161] The set of factors affecting the distributable function on both sides of the gameζ i for:

[0162] ζ1=(ξ 1W ,ξ 1SA )=(0.124, 0.139);

[0163] ζ2=(ξ 2W ,ξ 2SA )=(0.164, 0.139);

[0164] ζ3=(ξ 3W ,ξ 3SA )=(0.406, 0.583);

[0165] ζ4=(ξ 4W ,ξ 4SA )=(0.036, 0.139);

[0166] (2) Clustering of impact factor indicators

[0167] Take M = 0.4 and calculate the fuzzy similarity matrix

[0168]

[0169] The fuzzy equivalent matrix is obtained by the transitive closure method

[0170]

[0171] Take the confidence level λ=0.6, then the fuzzy clustering matrix

[0172]

[0173] Therefore, the four assignable functions can be clustered into life support function, propulsion monitoring function, and driving function as one category, and collision avoidance function as another category. Analyzing the size and clustering of the influencing factor indicators, the strategy set S1 of the mental workload player P1 is obtained as (C1, C2, C4), which contains a total of 2×2×2 (8) strategies; the strategy set S2 of the situational awareness player P2 is obtained as (C3), which contains a total of 2 strategies;

[0174] The utility value is standardized as shown below:

[0175]

[0176]

[0177] Where W nmz with SA nmzW are the mental workload utility value and situational awareness utility value when the assignable function of the mental workload player P1 adopts strategy n, the assignable function of the situational awareness player P2 adopts strategy m, and the non-assignable function adopts fixed strategy z; min With W max are the minimum and maximum values of mental workload utility; SA min with SA max are the minimum and maximum values of situational awareness utility, respectively. The mental workload and situational awareness non-cooperative game utility matrix for different allocation strategy combinations during navigation is shown in Table 8.

[0178] Table 8: Mental workload during navigation-situational awareness utility matrix

[0179]

[0180] Note: The vertical columns of the table show the eight strategic choices of the mental workload player, and the horizontal columns show the two strategic choices of the situational awareness player.

[0181] Table 8 shows a strategic analysis of the mental workload utility and situational awareness utility of different functional allocation strategy combinations during the navigation process. Regardless of which strategy combination is selected as the initial strategy, the Nash equilibrium strategy combination of both players can be found within three game rounds.

[0182] S={(S1), (S2)}={(C1, C2, C4), (C3)}={(1, 1, 1), (2)};

[0183] Taking the mental workload weight ω1 = 0.5 and the situational awareness weight ω2 = 0.5, the optimal strategy combination obtained by the linear weighting method is S = {(S1), (S2)} = {(C1, C2, C4), (C3)} = {(1, 2, 1), (1)}; when the optimal strategy S2 = (C3) = (1) is reached, the situational awareness utility is u2 = 0.126. At this time, there is still a strategy S2 = (C3) = (2) with a utility of u2 = 0.100, which is better than the linear weighted optimal strategy in terms of situational awareness. That is, under the linear weighted optimal strategy, situational awareness still has room for improvement;

[0184] Compared with other strategies of player P1, S1={(C1,C2,C4)}={(1,1,1)} has a strict advantage in terms of mental workload utility, satisfying the formula At this time, any strategy change made by P1 will only reduce its own benefits. At the same time, strategy S2 = {(C3)} = {(2)} is the optimal value of P2 under P1's strict dominant strategy. Therefore, P2 cannot unilaterally change its strategy to protect its own benefits, and the game thus reaches a Nash equilibrium.

[0185] The basic principles, main features, and advantages of the present invention are shown and described above. Those skilled in the art should understand that the present invention is not limited to the foregoing embodiments. The foregoing embodiments and descriptions are merely illustrative of the principles of the present invention. Various changes and modifications may be made to the present invention without departing from the spirit and scope of the present invention. Such changes and modifications are intended to fall within the scope of the present invention. The scope of protection claimed in the present invention is defined by the appended claims and their equivalents.

Claims

1. A method for dynamic human-machine function allocation in a manned submersible based on non-cooperative game theory, characterized by: Includes steps S1: Conduct a preliminary functional analysis of the functions involved in manned submersibles, and divide the submersible functions into four parts: information acquisition, information analysis, decision generation, and action implementation; S2: Use the function matrix to analyze the submersible's functions, identifying the parts that humans are good at completing, the parts that machines are good at completing, and the parts that both humans and machines are good at completing. The parts of the functions that humans and machines are good at completing are classified as assignable functions, and the opposite functions are classified as non-assignable functions. S3: Perform a hierarchical analysis of the parts of the allocable functions that both humans and machines are good at completing, and determine the allocation strategy for the allocable functions; S4: Perform single-objective optimization on the mental workload player and the situational awareness player under the allocation strategy to obtain their optimal allocation strategy combinations; calculate the impact factor ξ of each allocable function on the mental workload player in the optimal allocation strategy combination of the mental workload player. iW The influence factor ξ of each allocable function on the situational awareness player in the optimal allocation strategy combination of the situational awareness player is iSA ; S5: Calculate the game utility matrix based on the initial samples obtained in step S2 and step S4; S6: Analyze the Nash equilibrium of the game utility matrix and obtain the optimal allocation strategy; S7: When the alertness of the pulse wave signal detection drops to the set limit, the human-machine function is triggered to perform non-cooperative game redistribution.

2. The method for dynamic human-machine function allocation of a manned submersible based on non-cooperative game according to claim 1, characterized in that: In step S4, the single-objective optimization is performed on the mental workload player and the situational awareness player under the allocation strategy to obtain their optimal allocation strategy combinations, thereby obtaining the influence factor ξ of each allocable function on the mental workload player in the optimal allocation strategy combination of the mental workload player. iW The influence factor ξ of each allocable function on the situational awareness player in the optimal allocation strategy combination of the situational awareness player iSA The specific process includes: S41: Perform single-objective optimization on the mental workload player and the situational awareness player, calculate the mental workload value of the allocation strategy combination, and the minimum value corresponds to the allocation strategy combination with the optimal mental workload. Then calculate the situational awareness value of the function allocation strategy combination, and the maximum value corresponds to the allocation strategy combination with the optimal situational awareness. S42: Calculate the mental load occupied by each allocable function during the execution of the optimal allocation strategy combination of mental load using the VACP scale, and use this as the influence factor of each allocable function on the mental load game player ξ iW ; S43: Select experts to use the AHP method to obtain the relative impact of each assignable function on situational awareness in the optimal allocation strategy combination of situational awareness, and take the average value of the relative impact as the impact factor ξ of each assignable function on the situational awareness game party iSA ; g i ={ξ iW ,x iSA } Where, ζ i is the set of factors that influence the optimization objective of the assignable function i.

3. The method for dynamic human-machine function allocation of a manned submersible based on non-cooperative game according to claim 2, characterized in that: In step S41, the mental workload value of the function allocation strategy combination is calculated using the VACP scale; at the same time, the situational awareness value of the function allocation strategy combination is calculated using the situational awareness comprehensive assessment technology.

4. The method for dynamic human-machine function allocation of a manned submersible based on non-cooperative game according to claim 3, characterized in that: The situational awareness utility function of the situational awareness comprehensive assessment technology is: Where n is the number of questions and K is the number of subjects.

5. The method for dynamic human-machine function allocation of a manned submersible based on non-cooperative game according to claim 3, characterized in that: The calculation process of the mental workload value of the function allocation strategy combination based on the VACP scale includes: (1) Decompose the execution process of the function allocation strategy combination of a manned submersible in a certain working phase into several subtasks; (2) The VACP scale is then used to determine the load value of each subtask. The average mental load value within the time period is used as the measurement standard for the mental load of the functional allocation strategy combination. The mental load game utility function is obtained as shown in the following formula: Where W t W is the mental workload at a certain time t during the execution of the function allocation strategy combination; ab is the mental workload of subtask a in channel b during the execution process at a certain moment; W is the average mental workload during the time period T.

6. The method for dynamic human-machine function allocation of a manned submersible based on non-cooperative game according to claim 1, characterized in that: The calculation process of the game utility matrix described in step S5 includes: (1) Perform fuzzy clustering on the initial samples obtained in step S4 to achieve a strategy set diversity of assignable functions; (2) The allocable functions of the strategy set and the non-allocable functions described in step S2 together constitute an allocation strategy combination, and then the allocation strategy combination is used to form a game utility matrix through the calculation of the mental workload game player utility function and the situational awareness game player function.

7. The method for dynamic human-machine function allocation of a manned submersible based on non-cooperative game according to claim 6, characterized in that: The step S5 of performing fuzzy clustering on the initial samples obtained in step S4 to realize the strategy set diversity of the assignable function includes: S51: Utility Dimension Processing Since the utility dimensions of the two game players are different, dimensional processing is required; Where, After range transformation, it is shown as follows S52: Establishing fuzzy similarity matrix Establish the fuzzy similarity matrix R=(r ij ) n×n , r ij ξ″ ik and ξ″ jk The similarity is calculated by the absolute distance method as follows: In the formula, M is the similarity correction coefficient. When r ij ∈[0, 1]; S53: Establishing fuzzy equivalence matrix The transitive closure matrix R is established by using the square self-synthesis method for the fuzzy similarity matrix R * , the fuzzy equivalent matrix of R, and satisfies is a Boolean operator; S54: Strategic Clustering According to the clustering requirements, select the cutoff number λ to classify the matrix, satisfying In order to obtain different cut-off sets, the strategic attribution of the assignable function to the game players is finally achieved by clustering the elements of the same cut-off set.

8. The method for dynamic human-machine function allocation of a manned submersible based on non-cooperative game according to claim 1, characterized in that: The process of analyzing the Nash equilibrium of the game utility matrix and obtaining the optimal allocation strategy in step S6 includes: S61: Given the total strategy space, utility function, constraints and iteration accuracy ε, The constraints are: IN i ∈{W′ i } IN i ∈{IN′ i }1≤i≤n W tV ≤7 W tA ≤7 W tC ≤7 W tP ≤7 IN min ≥|IN| Where W and SA are the mental workload function and situational awareness function under the human-machine function allocation strategy in a certain working phase of a manned submersible, respectively; W i with SA i are the mental workload and situational awareness of function i in a certain working phase of manned submersible; W′ i with SA′ i are the mental workload and situational awareness of function i in a certain working stage of a manned submersible when a feasible allocation strategy is adopted; n is the number of functions in a certain working stage of a manned submersible; W V ,W tA ,W tC , W tP are the sum of the visual channel load, the auditory channel load, the cognitive channel load, and the motor channel load at a certain moment; SA min and |SA| are the minimum situational awareness value and the specified minimum safety threshold of the combination of human-machine function allocation strategy in a certain working phase of a manned submersible, respectively; S62: After fuzzy clustering of the initial sample of assignable functions described in step S5, the strategy sets S1, S2, ..., S belonging to each player are obtained. i ,…,S m ; S63: Game analysis initialization In the strategy space S={S1,S2,…,S i ,…,S m Randomly generate the initial feasible strategy combination of the game S64: Note For the feasible initial strategy combination s (0) Medium The initial feasible strategy combination adopted by all other players except i (s),…,u m (s) is the optimization goal, while maintaining Under the condition that the strategy sets S1, S2, ..., S belonging to each player remain unchanged, i ,…,S m Perform corresponding single-objective optimization in That is, for any player i, in its corresponding strategy set S i Find the optimal strategy Game Utility And satisfy the constraints where k = 1, 2, 3, ..., q; S65: Order Calculate the two strategy combinations s before and after (0) With s (1) Whether the distance between them satisfies the convergence criterion ||s (1) -s (0) ||≤ε, if satisfied, the game ends; if not satisfied, s (0) Replace with s (1) , go to step S3 to perform loop calculation; repeat this iteration until the accuracy requirement is met.

Images