An action sequence planning method for airport vehicle central control testing

Abstract

The present invention proposes an action sequence planning method for a central control test of an airport vehicle, and the method comprises the following steps: S1: performing image recognition on a driver's hand, collecting the driver's specific gestures and click positions through image data, constructing a graph structure through the image data collected by the driver, and abstractly extracting the edges in the image into a list; then splitting the action data into operable buttons and extracting them into edges; S2: adding all edges into a path table, and setting the path table as E_initial=V_in, E_target=V_out, V_path[V_ID], Path_len=1; S3: taking one of the edges as a starting point, and taking the page number of the page jumped to after the button of the edge is used as input data to query the existing data in the graph, and taking sel as a prefix, and all data of sel.E_initial==V_out.

Description

Technical Field

[0001] The invention belongs to the field of test sequence planning in computers, and in particular relates to an action sequence planning method for central control testing of airport vehicles. Background Art

[0002] The vehicle central control system is equivalent to the brain core of a vehicle, controlling the software and hardware in the vehicle, such as air conditioning, music, seats, windows, locks, wipers, etc. It is very necessary to test the software of the vehicle central control system to ensure the reliable operation of the vehicle equipment. For airport vehicles, a reliable central control system is even more important to ensure the orderly operation of the airport, such as more perfect testing of the central control system and completing the test by planning the action sequence of personnel.

[0003] Most of the tests on the central control system in the market use black box tests. When testers use black box tests, there will be incomplete tests for airport vehicles with complex functions, and there will be a lack of boundary condition tests. Drivers will use various functions in the process of driving the vehicle. At the same time, because there are various types of airport vehicles, there are shuttle buses, baggage conveyor belt vehicles, baggage tractors, passenger elevators, etc. The difference in each type of vehicle makes the operation of various vehicles different. Therefore, the central control test for ordinary vehicles is slightly insufficient. Therefore, a solution is needed to perform software detection on the central control system of airport vehicles. This requires the detection of the driver's actions, and the identification of different actions of different types of vehicles according to the driver's operation plan during the operation of the vehicle, and the central control test for this type of vehicle. Summary of the invention

[0004] The present invention proposes an action sequence planning method for central control testing of airport vehicles, which solves the problems of less data collection, complex data processing and fixed testing methods in central control testing of airport vehicles to a certain extent.

[0005] The technical solution of the present invention is implemented as follows: an action sequence planning method for airport vehicle central control testing, the method comprising the following steps:

[0006] S1: Perform image recognition on the driver's hand, collect the driver's specific gestures and click positions through image data, construct a graph structure with the image data collected by the driver, abstract and extract the edges in the image into a list; then split the action data into operable buttons and extract them into edges;

[0007] S2: Add all edges to the path table, and set the path table to E_initial=V_in, E_target=V_out, V_path[V_ID], Path_len= 1;

[0008] S3: Starting from one of the sides, query the existing data in the graph using the page number of the page that the button on this side jumps to after use as the input data, and query all the data with the prefix "sel" and sel.E_initial == V_out; create a new path for the data with sel.E_target!= E_initial and add it to the path table, using "new" as the prefix for the newly created path;

[0009] S4: Repeat step S3 until all the data with Path_len == 1 has been traversed. At this time, a large number of data with Path_len == 2 are generated. Query all the data with Path_len == 2, delete the duplicate data with the same E_initial and the same E_target, and only keep any one of them;

[0010] S5: If Edge.count > e^n, then continue with steps S3 and S4, traverse all the points with Path_len == 1, update the query and create new path information from the data with Path_len == n, and the new path Path_len == n + 1, until Edge.count < e^n and enter step S6:

[0011] S6: Query all the data with E_initial == 00001. If the number of queries is less than Edge.count - 1, then traverse all the reached points according to E_target, and then query the missing target points according to the point table. Excluding itself, record all the unreached target points, and use depth traversal until all the target points are queried. Keep the path with the smallest Path_len and create new data, and there are a total of (Edge.count - 1)^2 - Edge.count items of data in the final path table.

[0012] This action sequence planning method for airport vehicle central control testing has the following main differences compared with the prior art:

[0013] Image recognition and gesture analysis: This method first performs image recognition on the driver's hand to collect specific gesture and click position data. This way can capture the driver's operation intention more accurately, and is more intuitive and natural compared with traditional button or touch screen input. By constructing a graph structure and extracting edges and buttons, the operation sequence can be better abstracted and expressed, laying a foundation for subsequent path planning. Dynamic path planning: This method adopts the idea of dynamic programming, starting from the initial state, and gradually constructs possible operation paths. By continuously iterating and expanding the path length, more operation sequence combinations can be explored. This method can cover all possible operation scenarios more comprehensively and improve the integrity and effectiveness of the test.

[0014] Path optimization and deduplication: During the path generation process, this method effectively reduces data redundancy by removing duplicate paths with the same starting and ending points and retaining only one representative path. This optimization can reduce computational and storage overhead and improve algorithm efficiency.

[0015] Adaptive termination condition: This method introduces a termination condition based on an exponential function (Edge.count>e^n), which can adaptively adjust the depth of path generation according to the scale of the problem. This avoids blind exhaustive search and improves the efficiency of the algorithm while ensuring coverage.

[0016] Depth-first search completion: For target points that are not covered, this method uses a depth-first search strategy to complete them and select the shortest path. This ensures the completeness of the test case and also optimizes the path length to a certain extent.

[0017] Comprehensive path coverage: The final generated path table contains (Edge.count-1)^2 - Edge.count items. This method can produce more comprehensive and diverse test cases, which is conducive to discovering potential problems and edge cases.

[0018] Compared with the existing technology, this method has advantages in the following aspects: through image recognition and gesture analysis, this method can better simulate the real human-computer interaction process and improve the authenticity and effectiveness of the test. Secondly, the dynamic path planning and optimization strategy enable this method to effectively control the computational complexity while ensuring coverage. Furthermore, the adaptive termination condition and depth-first search completion mechanism enable this method to be flexibly adjusted according to the characteristics of the specific problem, which not only ensures the integrity of the test, but also avoids unnecessary waste of resources. Finally, by generating a large number of diverse test cases, this method can more comprehensively evaluate the performance and reliability of the central control system.

[0019] This method combines multiple technologies such as image recognition, dynamic planning, and path optimization to form a systematic and intelligent test solution. It can not only improve test efficiency and coverage, but also better simulate real operation scenarios, providing strong support for the performance evaluation and quality assurance of airport vehicle central control systems. The application of this innovative method is expected to significantly improve the reliability and user experience of airport vehicle central control systems.

[0020] As a preferred implementation, the path table in step S2 uses the page as point data in the graph, and the internal structure includes V_ID and Keys; wherein V_ID is the page ID number, which uses an int type variable, is unmodifiable and unique; Keys is the page name, which is used as the name of the point data to store the name of the page.

[0021] As a preferred implementation, in step S1, the button is used as edge data in the graph, and the internal structure includes: E_ID, V_in, V_out, Control_type, Elocal and Etarget, where E_ID is the edge ID number, V_in is the edge input page ID number, V_out is the edge output page ID number, Control_type is the operation type, Elocal is the position coordinate of the page where the button is located, and Etarget is the position coordinate information of the page that is finally reached by dragging and sliding.

[0022] As a preferred implementation, the Elocal is the position of the page where the button is located. When the button stays on the central control touch screen, the page is split into 64*64 squares and sequentially encoded into one-dimensional data to obtain 4096 coordinate points, starting from 0001; the Control_type is an operation on a non-central control touch screen, then the mechanical buttons are numbered, the first 6 bits are the number of the mechanical button, and the last 6 bits and Etaget together constitute 18 bits of position information as the position.

[0023] After adopting the above technical scheme, the beneficial effects of the present invention are as follows: Automation and intelligence: The method automatically captures the driver's gestures and operations through image recognition technology, without manual recording, which greatly improves the test efficiency and accuracy. Comprehensiveness: By constructing a graph structure and a traversal algorithm, the method can systematically cover all possible operation paths to ensure the comprehensiveness and completeness of the test. Optimization: By deleting duplicate paths and retaining the shortest path, the method can optimize the test sequence, reduce redundant operations, and improve the test efficiency. Scalability: The method can adjust parameters (such as Edge.count) according to actual needs to adapt to central control systems of different complexities. Visualization: Through the graph structure representation, the operation path can be intuitively displayed, which is convenient for analysis and optimization. Standardization: A standardized test method is provided, which helps to improve the consistency and comparability of test quality. Cost saving: The time and labor cost of manual testing are reduced, and the test efficiency is improved. Safety improvement: Through comprehensive testing, potential system defects can be better discovered and the operational safety of airport vehicles can be improved. Continuous optimization: The method can continuously accumulate test data to provide a basis for the continuous improvement of the system. Strong adaptability: This method can be applied to various types of airport vehicle central control systems and has broad application prospects. In general, this method provides an efficient, comprehensive and intelligent solution for airport vehicle central control testing by combining image recognition, graph theory and algorithm design, which helps to improve the quality and efficiency of testing and ultimately enhance the performance and safety of airport vehicles. BRIEF DESCRIPTION OF THE DRAWINGS

[0024] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative labor.

[0025] Figure 1 is a flow chart of the method of the present invention;

[0026] Figure 2 Schematic diagram of the path of the present invention. DETAILED DESCRIPTION

[0027] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0028] Example

[0029] like Figure 1 As shown, a motion sequence planning method for airport vehicle central control testing comprises the following steps:

[0030] S1: Perform image recognition on the driver's hand, collect the driver's specific gestures and click positions through image data, construct a graph structure with the image data collected by the driver, abstract and extract the edges in the image into a list; then split the action data into operable buttons and extract them into edges;

[0031] S2: Add all edges to the path table, and set the path table to E_initial=V_in, E_target=V_out, V_path[V_ID], Path_len= 1;

[0032] S3: Take one of the edges as the starting point, and use the page number of the page that jumps to after the button of the edge is used as the input data to query the existing data in the graph, with sel as the prefix, and all the data with sel.E_initial==V_out; create a new path for the data with sel.E_target!=E_initial and add it to the path table, and use new as the prefix of the newly created path;

[0033] S4: Repeat step S3 until all the data with Path_len == 1 are traversed. At this time, a large number of data with Path_len == 2 are generated. Query all the data with Path_len == 2, delete the duplicate data with the same E_initial and the same E_target, and only keep any one of them;

[0034] S5: If Edge.count > e^n, continue with step S3 and step S4. All the points with Path_len == 1 will be traversed, and new path information will be updated and created from the data with Path_len == n. The new path has Path_len == n + 1 until Edge.count < e^n and enter step S6:

[0035] S6: Query all the data with E_initial == 00001. If the query count is less than Edge.count - 1, traverse all the reached points according to E_target, then query the missing target points according to the point table, record all the unreached target points except itself, and use depth traversal until all the target points are queried. Keep the path with the smallest Path_len and create new data, and there are (Edge.count - 1)^2 - Edge.count items of data in the final path table.

[0036] This method is an action sequence planning method for airport vehicle central control testing. Its working principle and process are as follows: First, the driver's hand movements and click positions are captured by image recognition technology, and these data are constructed into a graph structure. The edges in the graph are abstracted into a list, and the operable buttons are extracted as edges. Then all edges are added to the path table, and the initial state and target state are set. Then, starting from a certain edge, query the edges connected to it, create a new path and add it to the path table. Repeat this process until all paths with a length of 1 are traversed, a large number of paths with a length of 2 are generated, and duplicate paths are deleted. If the number of edges is greater than the exponential function, continue the above process and continuously increase the path length until the number of edges is less than the exponential function. Finally, query all data with a starting point of 00001. If the number is insufficient, supplement the missing target points through deep traversal and retain the shortest path. The whole process continuously builds and improves the path table through iteration and optimization, and finally obtains a complete table containing all possible paths. Compared with the existing technology, the main innovations of this method are: using image recognition technology to obtain the actual operation data of the driver, which improves the authenticity and accuracy of the test; using a graph structure to represent the operation sequence, which is convenient for path planning and optimization; gradually building a path table by iteration, which can handle complex operation sequences; introducing an exponential function as a judgment condition, which effectively controls the computational complexity; 5) using a deep traversal method to supplement the missing paths and ensure the integrity of the path table. In general, this method integrates multiple technologies such as image recognition, graph theory, and path planning, and can efficiently generate comprehensive test cases, which helps to improve the test efficiency and quality of the airport vehicle central control system.

[0037] The present invention mainly provides a method for testing the central control of a vehicle, and the driver's action recognition is required in data collection. The data is mainly collected on the click operation of the driver's hand on the touch screen of the vehicle computer, and a graph structure is constructed according to the driver's hand movements. The vehicle-mounted central control screen is abstractly modeled, and the operation buttons near the vehicle-mounted central control screen are also abstractly constructed in a graph. The pages of the central control screen are taken as points in the graph, and the driver's operating actions are taken as edges in the graph. In data processing, the driver's actions are simply classified, including clicks, slides, long presses, double clicks, drags, etc. The sequence of the driver's actions is recorded, including the combination of actions, and an initial graph model is constructed. Using a similar method to the incremental model in software engineering, the simplest graph structure is first extracted from the driver's operation, and then the point data and edge data are separately extracted according to the action sequence and image recognition.

[0038] During the data collection phase, the driver's hand image recognition is first required to find out the driver's specific gestures and click positions. For example, when the driver extends his index finger and bends the other fingers, the area between the index fingers is used as a marking point to identify the operation, such as click, double-click, slide, drag, etc. If the driver clicks the screen with all five fingers open, the driver's hand movements are identified at this time. When it is determined that the index finger is higher than the middle finger, the position of the middle finger tip is used as the marking point, otherwise the position of the index finger tip is used as the marking point. If the driver's hand movement is grasping, it is determined that the driver is operating an operation button other than the central control screen, and the palm position is used as a marking point to record the trajectory of the driver's hand movement as data. If the driver's hand movement is pinching, the position of the fingertips together is used as a marking point, and the rotation movement and movement trajectory of the hand are recorded at the same time.

[0039] This method uses a new data structure and adopts a structure that is more convenient for searching in the code.

[0040] Data structure and encoding:

[0041] The page is regarded as the point data in the graph, and the internal structure includes: V_ID (page ID number), Keys (name of the page). V_ID is an int variable, which cannot be modified and is unique. For example, 0001, 0002 encoding. Keys is the name of the point data, which stores the name of the page, such as "Homepage", "Himalaya Radio", "NetEase Cloud Music" and other string data. The button is regarded as the edge data in the graph, and the internal structure includes: E_ID (edge ​​ID number), V_in (edge ​​entry page ID number), V_out (edge ​​exit page ID number), Control_type (operation type), Elocal (position coordinates of the page where the button is located), Etarget (position coordinate information of the page that is finally reached if dragging and sliding). E_ID is the unique identifier of the edge and is not repeatable. For example, 1001, 1002. Point data starts with 0, and edge data starts with 1.

[0042] V_in refers to the page number of the page where the button exists, such as 0001, 0002. If 0000 is used as the value of V_in, it means that it is universally applicable. For example, in an Android phone with a full screen, swiping up from the bottom is to return to the home page, which is universal for all pages of all programs, and swiping down from the top is universal for the notification bar and status bar. The V_in value of this button is set to 0000.

[0043] V_out is the page number to jump to after the button is used, for example, 0001, 0004, excluding 0000, because 0000 is the V_in value for global use, and there is no specific page, so it cannot jump.

[0044] Control_type is the operation type of the button when it is used. If it is a click-type button (click on the touch screen), the value is set to 01, left swipe, right swipe, up swipe and down swipe (such as switching pages) are set to 02, 03, 04, and 05 respectively, touch screen long press is set to 06, touch screen double click is set to 07, touch screen drag (such as dragging a file to the recycle bin) is set to 08, press (press on a mechanical case) is set to 17, pinch clockwise rotation is set to 18, pinch counterclockwise rotation is set to 19, grab and drag is set to 20, mechanical button long press is set to 21, mechanical button push (index finger press and push) is set to 22, and mechanical button pinch and slide is set to 23. This encoding design method allows the first bit of the 5-bit Control_type to determine whether it is an operation on the central control touch screen or a mechanical button, which is convenient for circuit board design in actual applications.

[0045] Elocal is the position of the page where the button is located. If it is an operation button for the central control touch screen, the page is split into 64*64 squares and sequentially encoded into one-dimensional data to obtain 4096 coordinate points, starting from 0001, which requires 12 bits. You can click on the corresponding position more accurately, while reducing unnecessary calculations that are too precise. Because when clicking a button, the size of the button is usually not much smaller than the size of a human fingernail, so splitting the interface into 64*64 is enough to imitate human finger operations. If Control_type is an operation of a non-central control touch screen, the mechanical button is numbered, the first 6 bits are the number of the mechanical button, and the last 6 bits and Etaget together form an 18-bit position information for the position. If it is a fixed position, the position is numbered. If it is a sliding operation or a rotation operation, the position and angle at this time are recorded. The first 9 bits of the 18 bits are the current position or angle, and the last 9 bits are the position or rotation angle after moving.

[0046] Etarget is the final position reached during drag and slide operations, and is the same encoding format as Elocal, with a total of 12 bits. If the operation is on the central control touch screen and the operation is dragging, this is the destination address coordinate, and Etarget is not set when sliding. If it is a mechanical operation, it is the destination position coordinate of sliding or dragging, or the angle of rotation operation. In practical applications, this method can comprehensively measure the vehicle central control, and effectively imitate and learn different operations of drivers for different models. It can reduce human labor, and the vehicle can be operated and tested by a robotic arm simulating a human arm, which is a saving of human resources. At the same time, due to the accuracy of the machine, the reliability and high efficiency of the working process can be guaranteed. There will be no special artificial circumstances that cause problems with airport vehicles and delays in aircraft takeoff and landing. At the same time, because the path is stored in a list, it is convenient and fast when searching for the path. You only need to record the action sequence in advance to complete it quickly in actual operation.

[0047] The point data structure in this application document is shown in Table 1;

[0048]

[0049] The edge data structure is shown in Table 2:

[0050]

[0051] Its path content is as follows Figure 2 As shown in Table 3:

[0052]

[0053] In the path table in step S2, the page is used as point data in the graph, and the internal structure includes V_ID and Keys; wherein V_ID is the page ID number, which uses an int type variable, cannot be modified and is unique; Keys is the page name, which is the name of the point data and stores the name of the page. The page is used as the point data in the graph, and each point is given a unique V_ID and Keys attribute, which realizes the abstract representation of the interface structure. This representation method not only simplifies the complex interface structure, but also facilitates subsequent path planning and optimization. V_ID uses an int type variable and cannot be modified, which ensures the unique identification of each page and avoids possible conflicts and confusion. This design enables the system to quickly locate and access specific pages, improving operational efficiency. Keys, as the page name, stores information that is easier for humans to understand, enhancing the readability and maintainability of the system. This dual identification method not only ensures the efficient operation of the system, but also facilitates the work of developers and testers.

[0054] This structural design provides a solid foundation for subsequent path planning. Through V_ID, the system can quickly build and traverse the graph structure; and Keys can be used to generate more readable test reports. This design also has good scalability. If you need to add a new page or modify an existing page, you only need to update the attributes of the point data without changing the overall structure. In general, this design method has obvious advantages in improving system efficiency, enhancing readability, facilitating maintenance and expansion, and provides more powerful support for airport vehicle central control testing.

[0055] In step S1, the button is used as the edge data in the graph, and the internal structure includes: E_ID, V_in, V_out, Control_type, Elocal and Etarget, where E_ID is the edge ID number, V_in is the edge entry page ID number, V_out is the edge exit page ID number, Control_type is the operation type, Elocal is the position coordinate of the page where the button is located, and Etarget is the position coordinate information of the page finally reached by dragging and sliding. Using the button as the edge data in the graph and giving it rich attribute information greatly improves the system's ability to describe user operations and accuracy. E_ID, as a unique identifier, ensures the uniqueness of each operation, which facilitates the system to quickly locate and track. The design of V_in and V_out cleverly encodes the jump relationship between pages into the edge, allowing the system to clearly represent and analyze complex interface navigation structures.

[0056] The introduction of control_type allows the system to distinguish different types of operations, such as click, slide or drag, which provides more possibilities for subsequent test case generation. The design of Elocal and Etarget is particularly innovative. They not only record the exact position of the button on the page, but also include the start and end coordinates of the slide or drag operation. This fine spatial information recording enables the system to simulate operations that are closer to real user behavior, greatly improving the authenticity and effectiveness of the test. Compared with the existing technology, the advantages of this design method are: it provides a more comprehensive and accurate description of the operation, which can capture the details of the user interface interaction; it supports complex operation sequence planning, and can generate more diverse and realistic test cases; 3) it is easy to analyze and optimize the user interface, and the interface design can be improved based on the operation data; 4) it enhances the scalability of the system and can easily adapt to new interaction methods and interface elements; 5) it provides a rich data foundation for automated testing and user behavior analysis. In general, this design not only improves the efficiency and accuracy of the test by abstracting the user interface into a graph structure with rich attributes, but also provides valuable data support for the optimization of the user interface and the study of human-computer interaction, representing an important progress in the central control test technology of airport vehicles.

[0057] The Elocal is the position of the page where the button is located. When the button stays on the central control touch screen, the page is split into 64*64 squares and sequentially encoded into one-dimensional data to obtain 4096 coordinate points, starting from 0001; the Control_type is an operation on a non-central control touch screen, then the mechanical buttons are numbered, the first 6 bits are the number of the mechanical button, and the last 6 bits and Etaget together form 18 bits of position information as the position. Dividing the touch screen page into 64*64 squares and performing one-dimensional encoding creates an accurate and easy-to-process coordinate system. This method not only improves positioning accuracy, but also simplifies the data processing process. By encoding the 4096 coordinate points starting from 0001, the position of each button can be accurately located, which helps to simulate user operation behavior more accurately. Secondly, for non-touch screen mechanical buttons, a 12-bit binary encoding method is adopted, in which the first 6 bits are used for button numbers, and the last 6 bits and Etarget together form 18 bits of position information.

[0058] This encoding method can effectively distinguish different types of input devices and provide enough information to describe the location and function of buttons. This unified and flexible encoding system enables the test system to handle the operation of touch screens and mechanical buttons at the same time, greatly enhancing the comprehensiveness and adaptability of the test. In addition, this encoding method also facilitates subsequent data analysis and processing, making it easier to identify operation modes, analyze user behavior, and optimize interface design.

[0059] This design greatly improves the accuracy and efficiency of the test through refined coordinate positioning and a unified coding system, and also provides strong data support for the optimization and improvement of the central control system. This method is not only suitable for current test needs, but also has good scalability and can adapt to new interaction methods that may appear in the future, laying a solid foundation for the long-term development of the airport vehicle central control system.

[0060] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principle of the present invention should be included in the protection scope of the present invention.

Claims

1. An action sequence planning method for airport vehicle central control testing, characterized in that: The method includes the following steps: S1: Perform image recognition on the driver's hand, collect the specific gestures and click positions of the driver through image data, construct a graph structure from the image data collected by the driver, and abstractly extract the edges in the image into a list; then split the action data into operable buttons and extract them as edges; S2: Add all the edges to the path table. The set path table is E_initial = V_in, E_target = V_out, V_path[V_ID], Path_len = 1; S3: Use one of the edges as the starting point, query the existing data in the image with the page number that the button jumps to after use as the input data, and query all the data with sel as the prefix and sel.E_initial == V_out; create a new path for the data with sel.E_target!= E_initial and add it to the path table, using new as the prefix for the newly created path; S4: Repeat step S3 until all the data with Path_len == 1 are traversed. At this time, a large number of data with Path_len == 2 are generated. Query all the data with Path_len == 2, delete the duplicate data with the same E_initial and the same E_target, and only keep any one of them; S5: If Edge.count > e^n, continue with steps S3 and S4, traverse all the points with Path_len == 1, update the query and create new path information from the data with Path_len == n. The new path has Path_len == n + 1 until Edge.count < e^n and enter step S6: S6: Query all the data with E_initial == 00001. If the query quantity is less than Edge.count - 1, traverse all the reached points according to E_target, then query the missing target points according to the point table, remove all the unreached target points except itself, and use depth traversal until all the target points are queried. Keep the path with the minimum Path_len and create new data, and there are a total of (Edge.count - 1)^2 - Edge.count items of data in the final path table.

2. The method for planning an action sequence for central control testing of airport vehicles according to claim 1, characterized in that: In the path table in step S2, the page is used as the point data in the graph, and the internal structure includes V_ID and Keys; where V_ID is the page ID number, which is an int-type variable, cannot be modified and is unique; Keys is the page name, which is used as the name of the point data and stores the name of the page.

3. The method for motion sequence planning for airport vehicle central control testing according to claim 1, characterized in that: In step S1, the button is used as the edge data in the graph, and the internal structure includes: E_ID, V_in, V_out, Control_type, Elocal and Etarget, where E_ID is the edge ID number, V_in is the edge input page ID number, V_out is the edge output page ID number, Control_type is the operation type, Elocal is the position coordinates of the page where the button is located, and Etarget is the position coordinate information of the page that is finally reached by dragging and sliding.

4. The method for planning an action sequence for central control testing of airport vehicles according to claim 3, characterized in that: The Elocal is the position of the page where the button is located. When the button stays on the central control touch screen, the page is split into 64*64 squares and sequentially encoded into one-dimensional data to obtain 4096 coordinate points, starting from 0001; the Control_type is an operation on a non-central control touch screen, then the mechanical buttons are numbered, the first 6 bits are the number of the mechanical button, and the last 6 bits and Etaget together form an 18-bit position information as the position.

Images