Single pilot driving mode task-function allocation verification method

Abstract

A single pilot driving mode task-function allocation verification method is provided, which constructs a flight scenario model, a system organizational architecture model, a system operation status model and a task organization management model for a two-pilot driving mode and a single pilot driving mode, respectively extracts the task-function matrices of the two-pilot driving mode and the single pilot driving mode from the model, then merges them and uses the TsCluster algorithm to perform differentially used task-function double clustering and low-used task-function double clustering mining and coverage calculation on the merged matrices, thereby completing the verification and analysis of whether the single-pilot driving function decomposition can meet the two-pilot driving task requirements. The present invention combines the constructed system model with the double clustering mining algorithm, and based on the mining results, calculates and verifies the coverage of the single pilot driving mode covering the two-pilot driving requirements.

Description

Technical Field

[0001] The present invention relates to a technology in the field of flight control, in particular to a commercial aircraft single pilot driving mode task-function allocation verification method based on double clustering mining. Background Art

[0002] The single-pilot driving mode is to reduce the number of crew members from two to one, while covering all flight phases, flight environments, meteorological conditions, airport requirements, and system capabilities and states, and maintaining the driving decision-making ability, efficiency, and effectiveness of the two-pilot driving mode. In the single-pilot driving mode, there is only one pilot on board, and the current independent cognition, behavior, and ability of a single pilot in the air do not meet the safety, driving load, and disability replacement requirements proposed by the International Civil Aviation Organization. Therefore, it is necessary to establish a new flight mission-function allocation method and verify whether it meets the application requirements of the two-pilot driving mode. Summary of the invention

[0003] Aiming at the verification problem of whether the existing task-function allocation method in the single-pilot driving mode can meet the application requirements of the two-person driving mode, the present invention proposes a single-pilot driving mode task-function allocation verification method, combines the constructed system model with the double clustering mining algorithm, and based on the mining results, calculates and verifies the coverage rate of the single-pilot driving mode covering the two-person driving requirements.

[0004] The present invention is achieved through the following technical solutions:

[0005] The present invention relates to a commercial aircraft single pilot driving mode task-function allocation verification method based on double cluster mining. The method comprises the following steps: constructing a flight scenario model of a two-pilot driving mode and a single pilot driving mode, a system organizational architecture model, a system operating state model and a task organization management model, extracting the task-function matrices of the two-pilot driving mode and the single pilot driving mode from the models, merging the two matrices, and performing differentially used task-function double cluster mining and low-used task-function double cluster mining and coverage calculation on the merged matrices using a TsCluster algorithm, thereby completing verification and analysis on whether the single pilot function decomposition can meet the two-pilot task requirements.

[0006] The present invention relates to a system for implementing the above method, comprising: a system model construction unit, a model relationship extraction unit and a verification analysis unit, wherein: the system model construction unit constructs a flight scenario model, a system organizational architecture model, a system operation status model and a task organization management model for a two-person driving mode and a single pilot driving mode; the model relationship extraction unit extracts the task-function matrix of the two-person driving mode and the single pilot driving mode respectively according to the constructed system model, and merges the two matrices; the verification analysis unit uses a TsCluster algorithm to perform differentially used task-function double clustering and low-used task-function double clustering mining and coverage calculation on the merged matrix, and analyzes the mining results to obtain a conclusion on whether the single-person driving function decomposition can meet the two-person driving task requirements.

[0007] Technical Effects

[0008] The present invention combines the constructed system model with the bicluster mining algorithm, calculates and verifies the coverage of the single-pilot driving mode covering the two-person driving requirements for typical flight scenarios, and solves the verification problem of whether the task-function allocation method under the single-pilot driving mode can meet the application requirements of the two-person driving mode, and proposes an efficient bicluster mining algorithm. The present invention combines the constructed system model with the bicluster mining algorithm, and based on the mining results, calculates and verifies the coverage of the single-pilot driving mode covering the two-person driving requirements. BRIEF DESCRIPTION OF THE DRAWINGS

[0009] Figure 1 It is the overall flow chart of the method of the present invention;

[0010] Figure 2 It is the flow chart of TsCluster algorithm of the present invention;

[0011] Figure 3 This is the result diagram of the TsCluster algorithm test run;

[0012] Figure 4 This is a system organization architecture model diagram under single pilot driving mode;

[0013] Figure 5 It is a model diagram of the system operation status in single pilot driving mode;

[0014] Figure 6 This is the mining result of TsCluster algorithm. DETAILED DESCRIPTION

[0015] like Figure 1 As shown, this embodiment relates to a commercial aircraft single pilot driving mode task-function allocation verification method based on double clustering mining, which includes:

[0016] The first step is to construct a hierarchical and mutually coupled flight scenario model, system organizational architecture model, system operation status model and task organization and management model for the two-pilot mode and the single-pilot mode, among which: the flight scenario model describes the dynamic relationship between the crew, the aircraft and the environment and covers the key flight environment, flight weather, aircraft factors, flight operations related to the crew's tasks and all basic functions and factors related to flight operations; the system organizational architecture model describes the organizational cross-linking relationship between systems; the system operation status model describes the changes in the operating status within the system under specific flight scenarios; the task organization and management model describes the calling relationship between tasks, functions and resources.

[0017] The critical flight environment includes: flight planning, crew tasks, air traffic management, air traffic approval and route permissions.

[0018] The flight weather includes: visibility, turbulence, icing, thunderstorms, wind shear, crosswind, etc.

[0019] The aircraft factors include: weight, balance, minimum equipment list, malfunctions, abnormal events, etc.

[0020] The organizational cross-linking relationship includes: interface configuration and interaction relationship between systems.

[0021] The calling relationship includes: decomposition of top-level flight tasks, configuration of functions and calling of resources.

[0022] In the second step, the task-function matrix of the two-pilot mode and the single-pilot driving mode are extracted from the flight scenario models, system organization architecture model, system operation status model and task organization and management model of the two-pilot mode and the single-pilot driving mode respectively. Specifically, taking the Mgaic Systems of Systems of Architect construction model as an example, the task organization and management model is generally presented through an activity diagram model, and the extraction of the task-function call relationship can be realized through the opaque action module in the activity diagram.

[0023] The task-function matrix is ​​a two-dimensional real number matrix M=F×T, where: the row F is the set of functions, the column T is the set of tasks, and the element M of the matrix M is ij is the usage rate of task j for function i. To facilitate mining, the original values ​​in the task-function matrix are usually discretized into 1, -1 and 0. Among them: 1 means that the task execution strongly uses a certain function, -1 means that a certain function is weakly used, and 0 means that the function is not used.

[0024] The functions F in the task-function matrix include:

[0025] ① The same functions used in both dual-pilot mode and single-pilot mode, such as the primary pilot function F1;

[0026] ② Functions used only in two-person driving mode, such as PM (pilot monitoring) monitoring function F2;

[0027] ③ New functions designed in the single pilot driving mode to make up for the lack of independent cognition, behavior and ability of the single pilot in the air, such as the newly designed onboard automatic system monitoring function F2 in the two-pilot mode 2-a , Ground Operator Monitoring Function F 2-b .

[0028] The third step is to transform the task-function matrix of the two-pilot mode and the single-pilot mode into a matrix by 1) removing the same functions used by both the two-pilot mode and the single-pilot mode, and 2) adding the two-pilot function F2 required for mission execution to the new function F in the single-pilot mode. 2-a and F 2-b The usage of merged together rules are merged.

[0029] In the fourth step, the TsCluster algorithm is used to perform differentially used task-function double clustering and low-used task-function double clustering mining and coverage calculation on the merged matrix to verify and analyze whether the single-person driving function decomposition can meet the requirements of the two-person driving task.

[0030] The task-function double clustering means that when the discretized usage value of function F1 for tasks T1 and T2 is M 11 and M 12 1) When M 11 =1,M 12 = -1, then the difference is satisfied, recorded as "F1"; 2) When M 11 =-1, M 12 =1, then the difference is satisfied, recorded as "*F1"; 3) When M 11 =-1, M 12 =-1, it satisfies low usability and is recorded as "-F1"; other cases are not recorded.

[0031] like Figure 2 As shown, the TsCluster algorithm includes:

[0032] Step 1) Scan the merged function-resource usage matrix to generate a task-task weight graph, which is stored in a cross-linked list structure.

[0033] The task-task weight graph definition means: the task-task weight graph can be represented by the set L = {S, P, V}, each vertex in the vertex set P in the weight graph represents a task; when there is an edge between a pair of vertices, the functions of the two tasks represented by this pair of vertices meet the task-function double clustering definition, and the set of edges is S; the weight on each edge is the set of functions that meet the task-function double clustering definition under the two tasks connected to this edge, and V is the set of weights.

[0034] Step 2) Use column expansion and do not retain candidate sets to mine all the largest task-function double clusters, and based on the pruning lemma, design three pruning strategies to improve the mining efficiency of the algorithm, including:

[0035] The pruning lemma 1 means: Assume that E currently expands the differential utilization rate double cluster, C is the candidate task set of E, and P is the predecessor task set of E. When for the candidate task C m (C m ∈C) any function F j For example, when it is in the form of "F j ”, when there is a predecessor task P n (P n ∈P), in P n There is also a function F j , then C m Function F in j The predecessor task P n Expanded.

[0036] The pruning strategy 1 is: when the candidate task C m To be pruned, the following conditions must be met: (1) EC m .Task is EP n .Task subset; (2)EC m .Function is EP n .Function subset; (3) For candidate task C m (C m ∈C) any function F j , which is in the form of "F j ”, there is a predecessor task P n (P n ∈P), in P n There is also a function "F j ”.

[0037] The pruning lemma 2 means: Assume that E currently expands the differential utilization rate double cluster, C is the candidate task set of E, and P is the predecessor task set of E. When for the candidate task C m (Cm ∈C) any function F j For example, when it is in the form of "*F j ”, when there is a predecessor task P n (P n ∈P), in P n There is also a function F j And the function is in the form of "-F j ”, then C m Function F in j The predecessor task P n Expanded.

[0038] The pruning strategy 2 is: when the candidate task C m To be pruned, the following conditions must be met: (1) EC m .Task is EP n .Task subset; (2)EC m .Function is EP n .Function subset; (3) For candidate task C m (C m ∈C) any function F j , which is in the form of "*F j ”, there is a predecessor task P n (P n ∈P), in P n There is also a function F j , and its expression is "-F j ”.

[0039] The pruning lemma 3 means: Assume that E currently expands the differential utilization rate double cluster, C is the candidate task set of E, and P is the predecessor task set of E. When for the candidate task C m (C m ∈C) any function F j For example, when it is in the form of "-F j ”, when there is a predecessor task P n (P n ∈P), in P n There is also a function F j And the function is in the form of "-F j ”, then C m Function F in j The predecessor task P n Expanded.

[0040] The pruning strategy 3 is: when the candidate task C m To be pruned, the following conditions must be met: (1) ECm .Task is EP n .Task subset; (2)EC m .Function is EP n .Function subset; (3) For candidate task C m (C m ∈C) any function F j , which is in the form of "-F j ”, there is a predecessor task P n (P n ∈P), in P n There is also a function F j , and its expression is "-F j ”.

[0041] Step 3) Based on the mining results, calculate the coverage parameters, specifically including: when R is the output task-function bicluster, T is the task set in the bicluster R, F is the function set in the bicluster R, α is the number of output biclusters R, β is the number of output biclusters that complete full coverage, then coverage = β / α.

[0042] The criteria for full coverage include:

[0043] 1) For a function called by a task executed by two-person driving, there must be at least one alternative function in single-person driving; for example, if task T1 calls function F2 in two-person driving, then task T1 must call function F2 in single-person driving. 21 Or function F 22 , the calling degree can be different;

[0044] 2) For the functions called by a set of tasks executed by two-person driving, there must also be a set of alternative functions in the single-person mode; for example, when tasks T1 and T2 are executed together in two-person driving, functions F1 and F2 need to be called at the same time, and the usage relationship must meet the difference constraint; in the single-person mode, when tasks T1 and T2 are executed together, there needs to be a set of functions {F 11 ,F 21} or {F 12 ,F 22} or other functional combinations must also satisfy the difference usage constraints.

[0045] In this embodiment, the algorithm efficiency of the TsCluster algorithm is compared with that of the TsCluster algorithm (retaining pruning strategy 1), the TsCluster algorithm (retaining pruning strategy 2), the TsCluster algorithm (retaining pruning strategy 3), and the TsCluster algorithm (removing pruning strategy) in different data sets to verify the effectiveness of the designed pruning strategy.

[0046] The experimental environment for the test is: Intel(R)Core(TM)CPU, 8G memory, Microsoft Visual C++6.0 version, and the data set used for the test is the public data set AGEMAP.

[0047] Results comparison:

[0048]

[0049] Combination Figure 3 It can be found that the pruning strategy of this method can effectively improve the mining efficiency of the algorithm, and the effect is more significant in the larger the data set.

[0050] Taking the land-based cruise flight phase as an example, the typical flight scenario is designed as follows: After the aircraft enters the land-based cruise phase, the pilot determines the flight route according to the flight plan, calculates the flight track through the flight management system, and makes collaborative decisions with the ground operator to propose a flight track request preferred by the user. The air traffic control establishes a route flight order based on waypoints, provides flight track analysis and conflict prediction, and feedbacks flight constraints. The aircraft analyzes the flight environment, such as weather radar and airspace weather data provided by air traffic control, and finds thunderstorm weather on the route ahead. According to the weather image display, the analysis concludes that there is no interruption gap in the thunderstorm weather, which will affect the aircraft's flight. Therefore, with the assistance of the onboard automatic system and collaborative decision-making with the ground operator, the pilot independently constructs a route to avoid hazardous weather. At the same time, airspace traffic information is obtained through automatic dependent surveillance (ADS-B) and traffic information system (TIS-B), and the flight status and intention of other aircraft in the airspace are analyzed through the onboard flight management system (FMS), and conflict monitoring that violates minimum isolation is established. When a threat is detected on an aircraft flying at the same altitude, the flight management system automatically generates an alarm and provides the pilot with several different flight path modification modes to resolve potential conflicts. The pilot on board makes collaborative decisions with the ground operator to resolve the conflict.

[0051] Based on the typical flight scenarios of the above-mentioned land-based cruise phase, the main flight tasks include user preferred track T1, hazardous weather avoidance T2 and airspace conflict avoidance T3. First, Magic Systems of Systems Architect is used to build the system organization architecture model, system operation status model and task organization management model for the two-person onboard mode and the single pilot driving mode. The system organization architecture model for the single pilot driving mode is as follows: Figure 4As shown in Figure 1, it involves ADS-B, TIS-B, FMS, CDTI, airborne automatic system, onboard pilot, air-ground data link, ground operator, ground automatic system, ground network transmission, and interaction between air traffic control and airlines. Taking airspace conflict avoidance T3 as an example, the system operation state model of the single pilot driving mode is as follows: Figure 5 shown.

[0052] Based on the constructed system model, the task-function matrix of dual-pilot and single-pilot can be extracted.

[0053] The task-function matrix of two-person driving is as follows:

[0054]

[0055] The task-function matrix for single-pilot driving is as follows:

[0056]

[0057]

[0058] According to the previous merging rules, the above two matrices are merged to obtain the merged task-function matrix as follows:

[0059] T1 T2 T3 F2 1 1 -1 F2-a 1 1 -1 F2-b -1 -1 -1 F5 -1 -1 1 F5-a -1 -1 1 F5-b -1 -1 1 F11 -1 -1 1 F11-a -1 -1 1 F11-b -1 -1 -1 F11-c -1 -1 -1

[0060] The merged task-function matrix is ​​mined using the TsCluster algorithm. The mining time is 0.005s. The mining results are as follows: Figure 6 As shown, the coverage rate is 100%. It can be seen that in the typical flight scenario of the land-based cruise phase, the task-function allocation of single-pilot can meet the task requirements of two-pilot.

[0061] Compared with the prior art, the present invention solves the verification problem of whether the task-function allocation method in the single-pilot driving mode can meet the application requirements of the two-person driving mode, and proposes an efficient double clustering mining algorithm. The present invention combines the constructed system model with the double clustering mining algorithm, and based on the mining results, calculates and verifies the coverage rate of the single-pilot driving mode covering the two-person driving requirements.

[0062] The above-mentioned specific implementation can be partially adjusted in different ways by those skilled in the art without departing from the principle and purpose of the present invention. The protection scope of the present invention shall be based on the claims and shall not be limited by the above-mentioned specific implementation. Each implementation scheme within its scope shall be subject to the constraints of the present invention.

Claims

1. A commercial aircraft single pilot driving mode task-function allocation verification method based on double clustering mining, characterized in that: By constructing the flight scenario model, system organizational architecture model, system operation status model and task organization and management model of the two-person driving mode and the single-pilot driving mode, and extracting the task-function matrix of the two-person driving mode and the single-pilot driving mode from the model, the two-person driving mode and the single-pilot driving mode are merged and the TsCluster algorithm is used to mine the differentially used task-function double clusters and the low-used task-function double clusters and calculate the coverage rate of the merged matrix, the verification and analysis of whether the single-person driving function decomposition meets the two-person driving task requirements are completed, including: The first step is to build a flight scenario model, system organizational structure model, system operation status model and task organization and management model for the two-pilot mode and the single-pilot mode, which are progressive and mutually coupled. The flight scenario model describes the dynamic relationship between the crew, the aircraft and the environment and covers the key flight environment, flight weather, aircraft factors, flight operations related to the crew's tasks, and all basic functions and factors related to flight operations; the system organizational structure model describes the organizational cross-linking relationship between systems; the system operation status model describes the changes in the internal operation status of the system under specific flight scenarios; the task organization and management model describes the calling relationship between tasks, functions and resources; The second step is to extract the task-function matrix of the two-pilot mode and the single-pilot mode from the flight scenario model, system organizational structure model, system operation status model and task organization and management model of the two-pilot mode and the single-pilot mode respectively; The third step is to transform the task-function matrix of the two-pilot mode and the single-pilot mode into a matrix by 1) removing the same functions used by both the two-pilot mode and the single-pilot mode, and 2) adding the two-pilot function F2 required for mission execution to the new function F in the single-pilot mode. 2-a and F 2-b Merge the rules that combine usage of The fourth step is to use the TsCluster algorithm to mine the differentially used task-function double clusters and the low-used task-function double clusters of the merged matrix and calculate the coverage rate to verify and analyze whether the single-driver function decomposition meets the requirements of the two-driver task. The critical flight environment includes: flight planning, crew tasks, air traffic management, air traffic approval and route clearance; The flight weather conditions include: visibility, turbulence, ice accumulation, thunderstorms, wind shear, and crosswind; The aircraft factors described include: weight, balance, minimum equipment list, malfunctions, and unusual events; The organizational cross-linking relationship includes: interface configuration and interaction relationship between systems; The functions F in the task-function matrix include: ① The same functions used in both two-pilot mode and single-pilot mode; ② Functions used only in two-person driving mode; ③ New functions designed in the single-pilot driving mode to compensate for the lack of independent cognition, behavior and ability of the single pilot in the air.

2. The method for verifying the task-function allocation of a single pilot driving mode of a commercial aircraft based on double clustering mining according to claim 1 is characterized in that: The second step is specifically as follows: constructing a model for Magic Systems of Systems of Architect, presenting the task organization management model through an activity diagram model, and extracting the task-function call relationship through the opaqueaction module in the activity diagram; The calling relationship includes: decomposition of top-level flight tasks, configuration of functions and calling of resources.

3. The commercial aircraft single pilot driving mode task-function allocation verification method based on double clustering mining according to claim 1 is characterized in that: The task-function matrix is ​​a two-dimensional real number matrix M=F×T, where: the row F is the set of functions, the column T is the set of tasks, and the element M of the matrix M is ij is the utilization rate of function i by task j; The original values ​​in the task-function matrix are discretized into 1, -1 and 0, where 1 means that the task execution strongly uses a certain function, -1 means that the function is weakly used, and 0 means that the function is not used.

4. The method for verifying the task-function allocation of a commercial aircraft single pilot driving mode based on double clustering mining according to claim 1 is characterized in that: The task-function double clustering means that when the discretized usage value of function F1 for tasks T1 and T2 is M 11 and M 12 ; 1) When M 11 =1,M 12 = -1, then the difference is satisfied, recorded as "F1"; 2) When M 11 =-1, M 12 =1, then the difference is satisfied, recorded as "*F1"; 3) When M 11 =-1, M 12 =-1, it satisfies low usability and is recorded as "-F1"; other cases are not recorded.

5. The commercial aircraft single pilot driving mode task-function allocation verification method based on double clustering mining according to claim 1 is characterized in that: The TsCluster algorithm includes: Step 1) Scan the merged function-resource usage matrix to generate a task-task weight graph, which is stored in a cross-linked list structure; The task-task weight graph definition means: the task-task weight graph is represented by a set L = {S, P, V}, each vertex in the vertex set P in the weight graph represents a task; when there is an edge between a pair of vertices, the functions of the two tasks represented by the pair of vertices meet the task-function double clustering definition, and the edge set is S; the weight on each edge is the function set of the two tasks connected to this edge that meet the task-function double clustering definition, and V is the set of weights; Step 2) All the largest task-function biclusters are mined by column expansion without retaining candidate sets, and three pruning strategies are designed based on the pruning lemma to improve the mining efficiency of the algorithm; Step 3) Based on the mining results, calculate the coverage parameters, specifically including: when R is the output task-function bicluster, T is the task set in the bicluster R, F is the function set in the bicluster R, α is the number of output biclusters R, β is the number of output biclusters that complete full coverage, then coverage = β / α.

6. The method for verifying the task-function allocation of a commercial aircraft single pilot driving mode based on double clustering mining according to claim 5 is characterized in that: The pruning lemma 1 means: Assume that E currently expands the differential utilization rate double cluster, C is the candidate task set of E, and P is the predecessor task set of E; when for candidate task C m (C m ∈C) any function F j For example, when it is in the form of "F j ”, when there is a predecessor task P n (P n ∈P), in P n There is also a function F j , then C m Function F in j By the predecessor task P n Expanded; the pruning strategy 1 means: when the candidate task C m To be pruned, the following conditions must be met: (1) EC m .Task is EP n .Task subset; (2)EC m .Function is EP n .Subset of Function; (3) For candidate task C m (C m ∈C) any function F j , which is in the form of "F j ”, there is a predecessor task P n (P n ∈P), in P n There is also a function "F j ”; The pruning lemma 2 means: Assume that E currently expands the differential utilization rate double cluster, C is the candidate task set of E, and P is the predecessor task set of E; when for candidate task C m (C m ∈C) any function F j For example, when it is in the form of "*F j ”, when there is a predecessor task P n (P n ∈P), in P n There is also a function F j And the function is in the form of "-F j ”, then C m Function F in j By the predecessor task P n Expanded; the pruning strategy 2 refers to: when the candidate task C m To be pruned, the following conditions must be met: (1) EC m .Task is EP n .Task subset; (2)EC m .Function is EP n .Subset of Function; (3) For candidate task C m (C m ∈C) any function F j , which is in the form of "*F j ”, there is a predecessor task P n (P n ∈P), in P n There is also a function F j , and its expression is "-F j ”; The pruning lemma 3 means: Assume that E currently expands the differential utilization rate double cluster, C is the candidate task set of E, and P is the predecessor task set of E; when for candidate task C m (C m ∈C) any function F j For example, when it is in the form of "-F j ”, when there is a predecessor task P n (P n ∈P), in P n There is also a function F j And the function is in the form of "-F j ”, then C m Function F in j By the predecessor task P n Expanded; the pruning strategy 3 refers to: when the candidate task C m To be pruned, the following conditions must be met: (1) EC m .Task is EP n .Task subset; (2)EC m .Function is EP n .Subset of Function; (3) For candidate task C m (C m ∈C) any function F j , which is in the form of "-F j ”, there is a predecessor task P n (P n ∈P), in P n There is also a function F j , and its expression is "-F j ”.

7. The method for verifying the task-function allocation of a commercial aircraft single pilot driving mode based on double clustering mining according to claim 5 is characterized in that: The criteria for full coverage include: 1) For a function called by a two-person driver to perform a certain task, there is at least one alternative function under single-person driving; 2) The functions called by the two-person pilot to perform a set of tasks, in single-person mode, there is a set of alternative functions.

8. A system for implementing the commercial aircraft single pilot driving mode task-function allocation verification method based on double clustering mining as described in any one of claims 1 to 7, characterized in that: include: The system model construction unit, the model relationship extraction unit and the verification and analysis unit, wherein: the system model construction unit constructs the flight scenario model of the two-person driving mode and the single-pilot driving mode, the system organizational structure model, the system operation status model and the task organization and management model; the model relationship extraction unit extracts the task-function matrix of the two-person driving mode and the single-pilot driving mode respectively according to the constructed system model, and merges the two matrices; the verification and analysis unit uses the TsCluster algorithm to perform differentially used task-function double clustering and low-used task-function double clustering mining and coverage calculation on the merged matrix, and analyzes the mining results to draw a conclusion on whether the single-person driving function decomposition meets the two-person driving task requirements.

Images