A method for compiling a station yard map with integrated driving prompts

Abstract

The present invention discloses a method for compiling a station yard map with integrated driving instructions, comprising the following steps: S1, generating common control combinations and association patterns to form a control usage behavior model; S2, inputting the control usage behavior model into a graph neural network to generate intelligent layout suggestions; S3, automatically laying out the controls using a force-directed algorithm based on the intelligent layout suggestions to generate a preliminary station yard map layout; S4, constructing a finite state machine for the controls based on the preliminary station yard map layout to enable the controls to achieve dynamic behavior response; S5, generating a station yard map with dynamic response capabilities based on the dynamic control behavior; S6, using simulated annealing tree search to calculate the optimal route path, and combining it with a constraint propagation algorithm to perform path conflict detection and adjustment; S7, performing image processing and feature calibration in an augmented reality environment to provide an intuitive path and station yard view. The present invention utilizes graph neural networks and simulated annealing tree search, etc., to achieve path optimization and station yard map compilation.

Description

Technical Field

[0001] The present invention relates to the technical field of scene graph compilation, and in particular to a method for compiling a station yard map for integrating driving prompts. Background Art

[0002] In modern rail transit dispatching and management, the compilation and maintenance of station yard maps is a crucial component in ensuring train scheduling, operational safety, and station management. Traditional methods for compiling station yard maps typically rely on manual design and static layouts. While this approach can meet basic needs on a small scale or in a fixed mode, it exhibits significant shortcomings in complex dispatching scenarios and in transportation systems that change in real time. Existing station yard map compilation methods often lack intelligence, dynamic responsiveness, and user interactivity, making them difficult to meet the high real-time and complexity requirements of modern rail transit.

[0003] Existing station map compilation technology primarily relies on static graphics software and manual adjustments, requiring designers to gradually complete layout and control settings based on the actual station conditions. This approach has significant limitations. For example, when scheduling conditions or station layouts change, recompiling or modifying station maps requires extensive manual intervention, which is time-consuming and labor-intensive, making it difficult to complete adaptive adjustments in a short period of time. Furthermore, static station maps lack the ability to adapt to real-time data and dynamic environments, resulting in delayed response and untimely information updates when handling emergencies or complex scheduling tasks.

[0004] In terms of data analysis and automatic layout, existing technologies have introduced some simple algorithms to assist in layout design, such as rule-based automatic layout algorithms or some heuristic methods. However, these methods can generally only handle layout problems of a limited scale and cannot dynamically adapt to changes in control state or complex user interactions. Furthermore, these algorithms lack sufficient intelligence for layout and path planning, and are unable to effectively integrate historical data and user behavior patterns to provide more targeted layout suggestions or path optimization strategies. Existing layout technologies lack complexity and flexibility, and are unable to achieve efficient layout and real-time optimization of multiple controls in large-scale scenarios.

[0005] When it comes to path planning, existing technologies typically rely on basic path search algorithms, such as Dijkstra's algorithm or the A* algorithm. These algorithms are capable of path search and optimization in static, simple station maps, but exhibit significant limitations in dynamic environments. These traditional algorithms struggle to provide fast and effective path optimization under real-time data synchronization and complex constraints, especially when faced with frequently changing station layouts and multi-dimensional path conflicts. Furthermore, these algorithms lack flexibility and adaptability, and are unable to synchronize and update path plans in multi-user environments, resulting in performance bottlenecks in real-time applications.

[0006] When it comes to user interaction and real-time feedback, existing technologies often rely on one-way data flows or simple event monitoring mechanisms, lacking efficient real-time data synchronization. Data transmission between servers and clients often suffers from delays and inconsistencies, which impacts the user experience and system responsiveness. While some systems may employ simple data synchronization and monitoring mechanisms, the performance and stability of these systems often fall short of meeting actual requirements when handling complex interactions and multi-user real-time operations.

[0007] In terms of image processing and augmented reality integration, traditional station map compilation systems primarily focus on the flat display of two-dimensional graphics, lacking integration with the physical scene and the ability to present three-dimensional space. Even existing augmented reality applications are limited to simple path overlay displays, struggling to support complex virtual control operations and real-time multi-view adjustments. The lack of support for feature calibration and real-time device posture estimation results in insufficient display accuracy between virtual controls and the physical scene, hindering the intuitiveness and effectiveness of operations.

[0008] Therefore, how to provide a method for compiling a station yard map with integrated driving prompts is a problem that those skilled in the art urgently need to solve. Summary of the Invention

[0009] One purpose of the present invention is to propose a method for compiling a station yard map that integrates driving prompts. The present invention comprehensively uses data mining, graph neural networks, long short-term memory networks and augmented reality technologies to achieve intelligent station yard map compilation and path optimization. Path calculation and conflict detection are performed through intelligent layout algorithms and simulated annealing tree search combined with constraint propagation algorithms, which effectively improves the rationality of layout and the efficiency of path planning. Real-time data synchronization technology ensures rapid response between the client and the server, and enhances the interactivity and stability of the system. Combined with augmented reality technology, virtual controls and paths are superimposed and displayed in the physical station yard, providing an intuitive and dynamic operation view to meet the needs of modern rail transit scheduling.

[0010] A method for compiling a station yard map with integrated driving prompts according to an embodiment of the present invention includes the following steps:

[0011] S1. Collect historical station yard design data and user operation records, and use association rule mining algorithms to analyze them, generate common control combinations and association patterns, and form a control usage behavior model;

[0012] S2. Input the control usage behavior model into the graph neural network for route configuration analysis, and combine it with the long short-term memory network to predict the time series of user operation behavior and generate intelligent layout suggestions;

[0013] S3. Based on the intelligent layout suggestions, the force-directed algorithm is used to automatically layout the controls. Combined with the Adam optimizer, the controls are reasonably distributed in the station map, and a preliminary station map layout is generated.

[0014] S4. Build a finite state machine for the control based on the preliminary station layout, define different states and behavior transitions of the control, and process user interaction events through an event-driven model to enable the control to achieve dynamic behavior response in the station layout;

[0015] S5. Based on dynamic control behavior, use server-client real-time data synchronization technology to achieve real-time data updates between the client and the server, forming a station map with dynamic response capabilities;

[0016] S6. Using real-time synchronized data, simulated annealing tree search is used to calculate the optimal route path, and constraint propagation algorithm is used to detect and adjust the path conflict;

[0017] S7. Combine the dynamically responsive station map with the optimal approach path, perform image processing and feature calibration in an augmented reality environment, and overlay the virtual controls on the physical station to provide an intuitive path and station view.

[0018] Optionally, the S2 specifically includes:

[0019] S21. Perform preliminary data preprocessing on the control usage behavior model, and use normalization and feature scaling techniques to normalize the model data;

[0020] S22. Input the preprocessed control usage behavior model into the graph neural network, establish a multidimensional feature vector for each control in the station field diagram through a node embedding method, and determine the association weights and path relationships between the controls based on the learning results of the control usage behavior model;

[0021] S23. In the graph neural network, the graph convolution layer is used for information transmission and aggregation, and the control features and weight information are propagated layer by layer. The node feature vector is calculated and updated through the adjacency matrix, and the control spatial features in the station map are gradually extracted.

[0022] S24. Input the control embedding result output by the graph neural network into the long short-term memory network for analyzing the time series data of the control usage behavior model. The long short-term memory network is used to extract the short-term dynamic characteristics and long-term trend characteristics of the control operation behavior to predict the user's layout adjustment preference.

[0023] S25. In the long short-term memory network, the past control operation history and the current control usage behavior model output are combined to comprehensively evaluate the association pattern between controls and the user's potential layout behavior, generate preliminary output of intelligent layout suggestions, and perform iterative adjustments to further optimize the layout suggestions.

[0024] S26. Fusion the control space features generated by the graph neural network with the layout suggestions output by the long short-term memory network to form the final intelligent layout suggestions.

[0025] Optionally, the S3 specifically includes:

[0026] S31. Receive the control feature vector and layout weight information extracted from the intelligent layout suggestion, initialize the position, connection relationship and physical parameters of each control node in the force-directed algorithm, and build a basic graph structure for the control layout;

[0027] S32. Adopt an improved dynamic force adjustment mechanism. In the force-directed algorithm, the attractive and repulsive force parameters are adjusted in real time according to the node density and layout complexity, dynamically adapting to the distance between nodes and the neighborhood distribution state.

[0028] S33. During the iterative layout process, a multi-threaded parallel computing strategy is used to process the control nodes in parallel by partition, and the forces and position changes between the controls are calculated. Each iteration includes node position adjustment and conflict detection, gradually approaching the global optimal layout state.

[0029] S34. Introduce a hierarchical layout strategy, prioritize the positions of high-priority controls, and then adjust the positions of low-priority controls, to improve the sense of hierarchy between controls and the rationality of functional distribution of the layout;

[0030] S35. Input the iterative layout result into the enhanced Adam optimizer. The Adam optimizer introduces a learning rate adjustment module and a gradient change tracking mechanism to achieve detailed adjustment of the control node position.

[0031] S36. Combined with the adjustment results output by the Adam optimizer, the entire layout is finally integrated and verified to form a preliminary station map layout.

[0032] Optionally, the S4 specifically includes:

[0033] S41. Receive the generated preliminary station layout, define the initial state and properties of the controls, including idle, occupied, and locked, and assign a state identifier and initial parameters to each control;

[0034] S42. Based on the preliminary station diagram layout, construct the finite state machine of the control, define the state set Q and event set E, and set the state transition function:

[0035] δ(q,e)=q′,δ:Q×E→Q;

[0036] Where δ represents the state transition function, q represents the current control state, e represents the event that triggers the state change, and q' represents the target control state;

[0037] S43. Configure an event-driven model to capture user interaction events and system-triggered events through a real-time event monitoring mechanism, and bind them to the state transition rules of the control's finite state machine so that events trigger state changes of the control, ensuring real-time response of the control to user operations.

[0038] S44. Define the dynamic state update formula of the control, and update the state parameters in combination with event triggering and time step:

[0039]

[0040] Among them, P ( t+1 ) Represents the state parameter of the control at time t+1, P ( t ) represents the state parameter of the control at time t, α represents the adjustment coefficient, w e represents the event weight, e(t) represents the event function, γ represents the stability constant, Δt represents the time step, and β represents the adjustment factor;

[0041] S45. Synchronize the control status update result to the station layout, and through data binding and real-time visualization mechanism, make the control status change immediately reflected on the graphical interface;

[0042] S46. Conduct simulation tests on control behavior responses to verify the response speed and stability of controls under various events and operating conditions, optimize the state transition logic and event-driven model of the finite state machine, and enable controls to achieve dynamic behavior responses in the station layout.

[0043] Optionally, the S5 specifically includes:

[0044] S51, receiving the generated dynamic control behavior data, initializing the real-time data synchronization module between the server and the client, and establishing a two-way communication channel;

[0045] S52. Define a data transmission protocol and construct a control status data packet structure D = {ID, S, T}, where ID represents the unique identifier of the control, S represents the current status information of the control, and T represents the timestamp;

[0046] S53, configure the event monitoring mechanism on the server side to capture the change event of the control state in real time and trigger the data transmission function T s (d i ):

[0047]

[0048] Among them, d i Indicates the control status data packet, ΔS ( t ) represents the rate of change of the control state at time t, ∈ represents the smoothing term, δ represents the smoothing parameter for adjusting state change detection, ζ represents the threshold of the state change rate, θ represents the attenuation factor for controlling transmission delay, T0 represents the packet generation time, τ represents the maximum allowed delay time, and ω represents the adjustment frequency factor;

[0049] S54, the client receives and parses the status data packet D transmitted from the server, decodes and verifies it through the parsing module, and the parsed control status is applied to the station layout of the client in real time;

[0050] S55. Use data binding technology to synchronize the parsed control state with the client site map. The data synchronization process dynamically adjusts the display priority and layout hierarchy of the control.

[0051] S56. Configure real-time monitoring and error correction mechanisms in the synchronization process to monitor packet loss, delays, or potential errors in data transmission. When an anomaly is detected, trigger the adaptive backtracking and retransmission mechanism to adjust the transmission strategy and form a site map with dynamic response capabilities.

[0052] Optionally, the S6 specifically includes:

[0053] S61. Acquire real-time synchronous data, including control status, station layout, and user interaction information, initialize the path search module, set the starting node and target node of the path search, and configure the path search parameters, including the initial temperature, cooling rate, and path evaluation criteria;

[0054] S62. Initiate a simulated annealing tree search algorithm, expand the current path as the initial path, and randomly select adjacent nodes in the search tree structure to generate candidate paths; the initial path is random and is used to explore diverse solutions in the path space; for each candidate path, calculate a path weight as a measure of path quality; the path weight is determined by a combination of path length, path node priorities, and specific constraints;

[0055] S63. After the path evaluation, the acceptance probability function of the simulated annealing algorithm is used to decide whether to accept the new path. The acceptance probability is controlled by the current temperature. The higher the temperature, the greater the acceptance probability. As the iteration proceeds, the temperature gradually decreases and the acceptance probability decreases, causing the search process to gradually transition from extensive exploration to local fine search.

[0056] S64. During the path generation and path selection phases, the constraint propagation algorithm is used to perform conflict detection on the path. During the detection process, all nodes and edges in the path are scanned to determine whether there is a conflict with fixed controls, preset path restrictions, or dynamic elements in the station map. If a conflict is detected, the constraint propagation algorithm identifies and marks the neighboring areas of the conflicting nodes.

[0057] S65: After the path adjustment mechanism is activated, an alternative node adjacent to the conflicting node is reselected to repair the path. The adjustment process is iterated in conjunction with the simulated annealing algorithm. By continuously updating and recalculating the path quality, the entire path is gradually optimized to achieve a conflict-free optimal solution.

[0058] S66. After the path optimization and adjustment are completed, the final optimal path is integrated into the station map layout, the station map is updated in real time, and the path planning results are displayed.

[0059] Optionally, the S7 specifically includes:

[0060] S71, receiving a station map with dynamic response capabilities and a generated optimal approach path, and loading them into an augmented reality module to provide a basis for virtual overlay display of the path and controls;

[0061] S72. Apply image processing technology to pre-process the station map and use image enhancement and edge detection algorithms to optimize image clarity;

[0062] S73. Use feature calibration technology to identify key locations and reference points in the physical station field, and use a scale-invariant feature transformation algorithm to calibrate and match the features in the physical station field;

[0063] S74. Use the Perspective-n-Point algorithm to perform real-time pose estimation of the device, combining the calibration features of the physical station and the parameters of the camera equipment to determine the real-time position and perspective of the device, ensuring the correct alignment of the virtual path and controls at different perspectives;

[0064] S75. Projecting the optimal route path and controls into the augmented reality environment through a rendering engine, achieving superimposed display of the virtual path and controls on the physical station. The rendering process includes dynamic adjustment of the path lines and real-time response of the controls.

[0065] S76. Test and verify the stability and responsiveness of paths and controls in the augmented reality environment, adjust display parameters and rendering logic, and provide intuitive path and station views.

[0066] The beneficial effects of the present invention are:

[0067] First, by incorporating data mining and machine learning technologies, this invention enables in-depth analysis of historical station design data and user operation records, automatically generating control usage behavior models and intelligent layout recommendations. This eliminates the need for manual compilation of station maps, significantly improving design efficiency and accuracy. Furthermore, by incorporating graph neural networks and long-short-term memory networks, control configuration and user behavior prediction become more precise, enhancing the intelligent level of layout.

[0068] Secondly, in terms of layout algorithms, a combination of a force-directed algorithm and the Adam optimizer achieves a rational distribution of controls within the station map. Compared to traditional manual layout or simple automatic layout methods, the automatic layout of this invention dynamically adapts to the spatial requirements of different controls, optimizing the layout and making the visualization and functional distribution of the station map clearer and more reasonable.

[0069] Furthermore, the application of a simulated annealing tree search algorithm and a constraint propagation algorithm addresses challenges in path planning and conflict detection, making path calculation more efficient and flexible, and adaptable to ever-changing dispatch environments. Compared to existing path planning techniques, this approach maintains efficient and accurate path search even under complex, multi-constraint conditions, ensuring the practicality and reliability of station maps in dynamic environments.

[0070] In terms of real-time data synchronization, the present invention realizes the rapid synchronization and update of control status and path information through server-client two-way communication and event-driven mechanism. This not only improves the response speed of the system, but also ensures the user's interactive experience and data consistency in a multi-user operating environment. Users can view changes in control status and path updates in real time on the client, avoiding the problems of data delay and synchronization instability in traditional systems. Finally, combined with augmented reality technology, virtual controls and paths are superimposed on the physical station for display, allowing dispatchers to intuitively view and manage driving prompts and path planning in the actual operating environment. This intuitive augmented reality display method enhances the user's understanding of the station map and the convenience of operation, effectively improving the efficiency and accuracy of dispatching work. BRIEF DESCRIPTION OF THE DRAWINGS

[0071] The accompanying drawings are used to provide a further understanding of the present invention and constitute a part of the specification. Together with the embodiments of the present invention, they are used to explain the present invention and do not constitute a limitation of the present invention. In the accompanying drawings:

[0072] Figure 1 This is a flow chart of a method for compiling a station yard map with integrated driving prompts proposed by the present invention;

[0073] Figure 2 This is a structural diagram of simulated annealing tree search path optimization and path conflict detection and adjustment for a method for compiling a station yard map with integrated driving prompts proposed by the present invention. DETAILED DESCRIPTION

[0074] The present invention will now be described in further detail with reference to the accompanying drawings, which are simplified schematic diagrams that illustrate the basic structure of the present invention in a schematic manner.

[0075] refer to Figure 1 and Figure 2 A method for compiling a station yard map with integrated driving prompts includes the following steps:

[0076] S1. Collect historical station yard design data and user operation records, and use association rule mining algorithms to analyze them, generate common control combinations and association patterns, and form a control usage behavior model;

[0077] S2. Input the control usage behavior model into the graph neural network for route configuration analysis, and combine it with the long short-term memory network to predict the time series of user operation behavior and generate intelligent layout suggestions;

[0078] S3. Based on the intelligent layout suggestions, the force-directed algorithm is used to automatically layout the controls. Combined with the Adam optimizer, the controls are reasonably distributed in the station map, and a preliminary station map layout is generated.

[0079] S4. Build a finite state machine for the control based on the preliminary station layout, define different states and behavior transitions of the control, and process user interaction events through an event-driven model to enable the control to achieve dynamic behavior response in the station layout;

[0080] S5. Based on dynamic control behavior, use server-client real-time data synchronization technology to achieve real-time data updates between the client and the server, forming a station map with dynamic response capabilities;

[0081] S6. Using real-time synchronized data, simulated annealing tree search is used to calculate the optimal route path, and constraint propagation algorithm is used to detect and adjust the path conflict;

[0082] S7. Combine the dynamically responsive station map with the optimal approach path, perform image processing and feature calibration in an augmented reality environment, and overlay the virtual controls on the physical station to provide an intuitive path and station view.

[0083] In this embodiment, S2 specifically includes:

[0084] S21. Perform preliminary data preprocessing on the control usage behavior model, and use normalization and feature scaling techniques to normalize the model data;

[0085] S22. Input the preprocessed control usage behavior model into the graph neural network, establish a multidimensional feature vector for each control in the station field diagram through a node embedding method, and determine the association weights and path relationships between the controls based on the learning results of the control usage behavior model;

[0086] S23. In the graph neural network, the graph convolution layer is used for information transmission and aggregation, and the control features and weight information are propagated layer by layer. The node feature vector is calculated and updated through the adjacency matrix, and the control spatial features in the station map are gradually extracted.

[0087] S24. Input the control embedding result output by the graph neural network into the long short-term memory network for analyzing the time series data of the control usage behavior model. The long short-term memory network is used to extract the short-term dynamic characteristics and long-term trend characteristics of the control operation behavior to predict the user's layout adjustment preference.

[0088] S25. In the long short-term memory network, the past control operation history and the current control usage behavior model output are combined to comprehensively evaluate the association pattern between controls and the user's potential layout behavior, generate preliminary output of intelligent layout suggestions, and perform iterative adjustments to further optimize the layout suggestions.

[0089] S26. Fusion the control space features generated by the graph neural network with the layout suggestions output by the long short-term memory network to form the final intelligent layout suggestions.

[0090] In this embodiment, S3 specifically includes:

[0091] S31. Receive the control feature vector and layout weight information extracted from the intelligent layout suggestion, initialize the position, connection relationship and physical parameters of each control node in the force-directed algorithm, and build a basic graph structure for the control layout;

[0092] S32. Adopt an improved dynamic force adjustment mechanism. In the force-directed algorithm, the attractive and repulsive force parameters are adjusted in real time according to the node density and layout complexity, dynamically adapting to the distance between nodes and the neighborhood distribution state.

[0093] S33. During the iterative layout process, a multi-threaded parallel computing strategy is used to process the control nodes in parallel by partition, and the forces and position changes between the controls are calculated. Each iteration includes node position adjustment and conflict detection, gradually approaching the global optimal layout state.

[0094] S34. Introduce a hierarchical layout strategy, prioritize the positions of high-priority controls, and then adjust the positions of low-priority controls, to improve the sense of hierarchy between controls and the rationality of functional distribution of the layout;

[0095] S35. Input the iterative layout result into the enhanced Adam optimizer. The Adam optimizer introduces a learning rate adjustment module and a gradient change tracking mechanism to achieve detailed adjustment of the control node position.

[0096] S36. Combined with the adjustment results output by the Adam optimizer, the entire layout is finally integrated and verified to form a preliminary station map layout.

[0097] In this embodiment, the S4 specifically includes:

[0098] S41. Receive the generated preliminary station layout, define the initial state and properties of the controls, including idle, occupied, and locked, and assign a state identifier and initial parameters to each control;

[0099] S42. Based on the preliminary station diagram layout, construct the finite state machine of the control, define the state set Q and event set E, and set the state transition function:

[0100] δ(q,e)=q′,δ:Q×E→Q;

[0101] Where δ represents the state transition function, q represents the current control state, e represents the event that triggers the state change, and q' represents the target control state;

[0102] S43. Configure an event-driven model to capture user interaction events and system-triggered events through a real-time event monitoring mechanism, and bind them to the state transition rules of the control's finite state machine so that events trigger state changes of the control, ensuring real-time response of the control to user operations.

[0103] S44. Define the dynamic state update formula of the control, and update the state parameters in combination with event triggering and time step:

[0104]

[0105] Among them, P( t+1 ) Represents the state parameter of the control at time t+1, P ( t ) represents the state parameter of the control at time t, α represents the adjustment coefficient, w e represents the event weight, e ( t ) represents the event function, γ represents the stability constant, Δt represents the time step, and β represents the adjustment factor;

[0106] S45. Synchronize the control status update result to the station layout, and through data binding and real-time visualization mechanism, make the control status change immediately reflected on the graphical interface;

[0107] S46. Conduct simulation tests on control behavior responses to verify the response speed and stability of controls under various events and operating conditions, optimize the state transition logic and event-driven model of the finite state machine, and enable controls to achieve dynamic behavior responses in the station layout.

[0108] In this embodiment, the S5 specifically includes:

[0109] S51, receiving the generated dynamic control behavior data, initializing the real-time data synchronization module between the server and the client, and establishing a two-way communication channel;

[0110] S52. Define the data transmission protocol and construct the control status data packet structure D= { ID,S,T } , where ID represents the unique identifier of the control, S represents the current state information of the control, and T represents the timestamp;

[0111] S53, configure the event monitoring mechanism on the server side to capture the change event of the control state in real time and trigger the data transmission function T s( d i) :

[0112]

[0113] Among them, d i Indicates the control status data packet, ΔS ( t ) represents the rate of change of the control state at time t, ∈ represents the smoothing term, δ represents the smoothing parameter for adjusting state change detection, ζ represents the threshold of the state change rate, θ represents the attenuation factor for controlling transmission delay, T0 represents the packet generation time, τ represents the maximum allowed delay time, and ω represents the adjustment frequency factor;

[0114] S54, the client receives and parses the status data packet D transmitted from the server, decodes and verifies it through the parsing module, and the parsed control status is applied to the station layout of the client in real time;

[0115] S55. Use data binding technology to synchronize the parsed control state with the client site map. The data synchronization process dynamically adjusts the display priority and layout hierarchy of the control.

[0116] S56. Configure real-time monitoring and error correction mechanisms in the synchronization process to monitor packet loss, delays, or potential errors in data transmission. When an anomaly is detected, trigger the adaptive backtracking and retransmission mechanism to adjust the transmission strategy and form a site map with dynamic response capabilities.

[0117] In this embodiment, S6 specifically includes:

[0118] S61. Acquire real-time synchronous data, including control status, station layout, and user interaction information, initialize the path search module, set the starting node and target node of the path search, and configure the path search parameters, including the initial temperature, cooling rate, and path evaluation criteria;

[0119] S62. Initiate a simulated annealing tree search algorithm, expand the current path as the initial path, and randomly select adjacent nodes in the search tree structure to generate candidate paths; the initial path is random and is used to explore diverse solutions in the path space; for each candidate path, calculate a path weight as a measure of path quality; the path weight is determined by a combination of path length, path node priorities, and specific constraints;

[0120] S63. After the path evaluation, the acceptance probability function of the simulated annealing algorithm is used to decide whether to accept the new path. The acceptance probability is controlled by the current temperature. The higher the temperature, the greater the acceptance probability. As the iteration proceeds, the temperature gradually decreases and the acceptance probability decreases, causing the search process to gradually transition from extensive exploration to local fine search.

[0121] S64. During the path generation and path selection phases, the constraint propagation algorithm is used to perform conflict detection on the path. During the detection process, all nodes and edges in the path are scanned to determine whether there is a conflict with fixed controls, preset path restrictions, or dynamic elements in the station map. If a conflict is detected, the constraint propagation algorithm identifies and marks the neighboring areas of the conflicting nodes.

[0122] S65: After the path adjustment mechanism is activated, an alternative node adjacent to the conflicting node is reselected to repair the path. The adjustment process is iterated in conjunction with the simulated annealing algorithm. By continuously updating and recalculating the path quality, the entire path is gradually optimized to achieve a conflict-free optimal solution.

[0123] S66. After the path optimization and adjustment are completed, the final optimal path is integrated into the station map layout, the station map is updated in real time, and the path planning results are displayed.

[0124] In this embodiment, the S7 specifically includes:

[0125] S71, receiving a station map with dynamic response capabilities and a generated optimal approach path, and loading them into an augmented reality module to provide a basis for virtual overlay display of the path and controls;

[0126] S72. Apply image processing technology to pre-process the station map and use image enhancement and edge detection algorithms to optimize image clarity;

[0127] S73. Use feature calibration technology to identify key locations and reference points in the physical station field, and use a scale-invariant feature transformation algorithm to calibrate and match the features in the physical station field;

[0128] S74. Use the Perspective-n-Point algorithm to perform real-time pose estimation of the device, combining the calibration features of the physical station and the parameters of the camera equipment to determine the real-time position and perspective of the device, ensuring the correct alignment of the virtual path and controls at different perspectives;

[0129] S75. Projecting the optimal route path and controls into the augmented reality environment through a rendering engine, achieving superimposed display of the virtual path and controls on the physical station. The rendering process includes dynamic adjustment of the path lines and real-time response of the controls.

[0130] S76. Test and verify the stability and responsiveness of paths and controls in the augmented reality environment, adjust display parameters and rendering logic, and provide intuitive path and station views.

[0131] Example 1:

[0132] In order to verify the feasibility of the present invention in implementation, the present invention is applied to the main railway hub of a modern city. Dispatchers are faced with the task of quickly dispatching a large number of trains during peak hours. This complex environment requires dispatchers to be able to quickly formulate and adjust driving plans, and at the same time respond to emergencies and dynamic changes in a very short time. However, the existing station yard map compilation method relies on manual operation and static path planning, which makes it difficult to achieve real-time response and adjustment to complex scheduling scenarios, resulting in problems such as path conflicts, resource waste and scheduling delays during peak hours. To this end, we apply the station yard map compilation method of the present invention to this embodiment to verify its effectiveness and improvement effect in practical applications.

[0133] During peak hours, dispatchers need to compile and update station yard maps within 5 minutes to accommodate train arrivals and departures. First, the system accesses real-time yard data, including current track status, train locations, and dispatch requests. This data is then imported into the system's path search module. The system uses historical data and association rule mining algorithms to generate a control usage behavior model, predicting common control combinations and user layout preferences. This control information is then analyzed for route configuration using a graph neural network. This information is then combined with a long-short-term memory network to predict the dispatcher's potential operational behavior and generate intelligent layout recommendations.

[0134] During application, the system utilizes a force-directed algorithm combined with the Adam optimizer to automatically layout controls, ensuring a reasonable distribution of controls within the station map and rapidly completing preliminary graphics compilation. To ensure the system's real-time response to complex path planning requests, a simulated annealing tree search algorithm is used for path optimization. The system automatically identifies path nodes and their conflicts, dynamically iterating to select the optimal path and avoid conflicts between paths. A constraint propagation algorithm is further used for path conflict detection and adjustment, ensuring the resulting conflict-free path.

[0135] This system enables dispatchers to view the generated optimal path and the dynamic response of controls in real time. By integrating augmented reality technology, virtual controls are overlaid onto the physical station, allowing dispatchers to directly see the dynamic display of the path within the actual scene. This approach enabled dispatchers to complete the compilation and multiple adjustments of the station map in just 5 minutes, reducing the time by 60% compared to traditional methods. The system's real-time data synchronization function ensures the consistency of dispatch information between the server and client, thereby achieving stable and accurate operation.

[0136] During one month of actual application, we selected data from the morning peak hours (7:00-9:00) and evening peak hours (17:00-19:00) every day for verification. The data showed that after implementing the method of the present invention, the time for dispatchers to compile and adjust station maps during peak hours was reduced to an average of 3 minutes, which is a significant improvement over the 8 minutes of the traditional manual compilation method. The incidence of path conflicts dropped from the original 15% to 2%. At the same time, due to path optimization and dynamic layout, the average dispatch delay of trains was reduced from 7 minutes to 3 minutes. Combined with user feedback, dispatchers unanimously stated that the system has made significant improvements in operational intuitiveness and response speed, and the augmented reality real-time path display makes decision-making more efficient.

[0137] Table 1 Comparison of station map preparation and adjustment time during peak hours between the traditional invention and the present invention

[0138]

[0139]

[0140] Table 2 Comparison of path conflict incidence between traditional invention and the present invention

[0141]

[0142] Table 1 above shows a comparison of the time for compiling and adjusting station yard maps between the traditional method and the method of the present invention during peak hours. The data shows that after using the method of the present invention, the average time for dispatchers to complete the compilation of station yard maps is significantly reduced during peak hours in the morning and evening. For example, during the morning peak hour (7:00-9:00) on November 1, 2024, the traditional method takes about 8 minutes, while the method of the present invention only takes 3 minutes, saving about 62.5% of the time. Similar trends have been verified on other dates and time periods, indicating that the present invention greatly improves the compilation efficiency when responding to complex dispatching tasks in real time. On average, the traditional method takes about 8 minutes, while the average time taken by the method of the present invention is about 3 to 3.5 minutes, showing its significant advantage in efficiency.

[0143] Table 2 above shows a comparison of the incidence rates of path conflicts, reflecting the superiority of the method of the present invention in path optimization and conflict detection. The data show that the path conflict rate of the traditional method is relatively high during peak hours, reaching about 15%, while the method of the present invention significantly reduces this ratio. For example, during the morning peak period on November 1, 2024, the traditional path conflict rate was 15%, while after using the method of the present invention, it was only 2%, and the conflict incidence rate was reduced by about 87%. Data from other dates also showed similar improvements, proving that the present invention exhibits better stability and reliability when dealing with complex path planning and real-time data synchronization. Through simulated annealing tree search and constraint propagation algorithms, the present invention effectively reduces the occurrence of path conflicts, enabling dispatchers to dispatch and manage more smoothly, thereby improving overall dispatching efficiency and system security.

[0144] The above description is only a preferred specific embodiment of the present invention, but the scope of protection of the present invention is not limited thereto. Any technician familiar with the technical field, within the technical scope disclosed by the present invention, who makes equivalent replacements or changes based on the technical solution and inventive concept of the present invention, should be covered by the scope of protection of the present invention.

Claims

1. A method for compiling a station yard map with integrated driving prompts, characterized in that: The steps include: S1. Collect historical station yard design data and user operation records, and use association rule mining algorithms to analyze them, generate common control combinations and association patterns, and form a control usage behavior model; S2. Input the control usage behavior model into the graph neural network for route configuration analysis, and combine it with the long short-term memory network to predict the time series of user operation behavior and generate intelligent layout suggestions; S3. Based on the intelligent layout suggestions, the force-directed algorithm is used to automatically layout the controls. Combined with the Adam optimizer, the controls are reasonably distributed in the station map, and a preliminary station map layout is generated. S4. Build a finite state machine for the control based on the preliminary station layout, define different states and behavior transitions of the control, and process user interaction events through an event-driven model to enable the control to achieve dynamic behavior response in the station layout; S5. Based on dynamic control behavior, use server-client real-time data synchronization technology to achieve real-time data updates between the client and the server, forming a station map with dynamic response capabilities; S6. Using real-time synchronized data, simulated annealing tree search is used to calculate the optimal route path, and constraint propagation algorithm is used to detect and adjust the path conflict; S7. Combine the dynamically responsive station map with the optimal approach path, perform image processing and feature calibration in an augmented reality environment, and overlay the virtual controls on the physical station to provide an intuitive path and station view.

2. The method for compiling a station yard map for integrating driving prompts according to claim 1 is characterized in that: The S2 specifically includes: S21. Perform preliminary data preprocessing on the control usage behavior model, and use normalization and feature scaling techniques to normalize the model data; S22. Input the preprocessed control usage behavior model into the graph neural network, establish a multidimensional feature vector for each control in the station field diagram through a node embedding method, and determine the association weights and path relationships between the controls based on the learning results of the control usage behavior model; S23. In the graph neural network, the graph convolution layer is used for information transmission and aggregation, and the control features and weight information are propagated layer by layer. The node feature vector is calculated and updated through the adjacency matrix, and the control spatial features in the station map are gradually extracted. S24. Input the control embedding result output by the graph neural network into the long short-term memory network for analyzing the time series data of the control usage behavior model. The long short-term memory network is used to extract the short-term dynamic characteristics and long-term trend characteristics of the control operation behavior to predict the user's layout adjustment preference. S25. In the long short-term memory network, the past control operation history and the current control usage behavior model output are combined to comprehensively evaluate the association pattern between controls and the user's potential layout behavior, generate preliminary output of intelligent layout suggestions, and perform iterative adjustments to further optimize the layout suggestions. S26. Fusion the control space features generated by the graph neural network with the layout suggestions output by the long short-term memory network to form the final intelligent layout suggestions.

3. The method for compiling a station yard map for integrating driving prompts according to claim 1 is characterized in that: The S3 specifically includes: S31. Receive the control feature vector and layout weight information extracted from the intelligent layout suggestion, initialize the position, connection relationship and physical parameters of each control node in the force-directed algorithm, and build a basic graph structure for the control layout; S32. Adopt an improved dynamic force adjustment mechanism. In the force-directed algorithm, the attractive and repulsive force parameters are adjusted in real time according to the node density and layout complexity, dynamically adapting to the distance between nodes and the neighborhood distribution state. S33. During the iterative layout process, a multi-threaded parallel computing strategy is used to process the control nodes in parallel by partition, and the forces and position changes between the controls are calculated. Each iteration includes node position adjustment and conflict detection, gradually approaching the global optimal layout state. S34. Introduce a hierarchical layout strategy, prioritize the positions of high-priority controls, and then adjust the positions of low-priority controls, to improve the sense of hierarchy between controls and the rationality of functional distribution of the layout; S35. Input the iterative layout result into the enhanced Adam optimizer. The Adam optimizer introduces a learning rate adjustment module and a gradient change tracking mechanism to achieve detailed adjustment of the control node position. S36. Combined with the adjustment results output by the Adam optimizer, the entire layout is finally integrated and verified to form a preliminary station map layout.

4. The method for compiling a station yard map for integrating driving prompts according to claim 1 is characterized in that: The S4 specifically includes: S41. Receive the generated preliminary station layout, define the initial state and properties of the controls, including idle, occupied, and locked, and assign a state identifier and initial parameters to each control; S42. Based on the preliminary station layout, build the finite state machine of the control and define the state set and event collections , and set the state transition function: ； in, represents the state transition function, Indicates the current control status. Indicates the event that triggers the state change, Indicates the target control state; S43. Configure an event-driven model to capture user interaction events and system-triggered events through a real-time event monitoring mechanism, and bind them to the state transition rules of the control's finite state machine so that events trigger state changes of the control, ensuring real-time response of the control to user operations. S44, combining event triggering and time step to update the control state; S45. Synchronize the control status update result to the station layout, and make the control status change instantly reflected on the graphical interface through data binding and real-time visualization mechanism; S46. Conduct simulation tests on control behavior responses to verify the response speed and stability of controls under various events and operating conditions, optimize the state transition logic and event-driven model of the finite state machine, and enable controls to achieve dynamic behavior responses in the station layout.

5. The method for compiling a station yard map for integrating driving prompts according to claim 1 is characterized in that: The S5 specifically includes: S51, receiving the generated dynamic control behavior data, initializing the real-time data synchronization module between the server and the client, and establishing a two-way communication channel; S52. Define data transmission protocol and build control status data packet structure ,in Represents the unique identifier of the control. Represents the current state information of the control. Indicates a timestamp; S53. Configure the event monitoring mechanism on the server side to capture the change event of the control state in real time and trigger the data transmission function : ； in, Represents the control status data packet, Indicates the state of the control at time The rate of change, represents the smoothing term, represents the smoothing parameter that regulates the detection of state changes, The threshold value indicating the rate of state change, represents the attenuation factor that controls the transmission delay, Indicates the time when the data packet is generated. Indicates the maximum allowed delay time. represents the adjustment frequency factor; S54. The client receives and parses the status data packet transmitted from the server , decoded and verified by the parsing module, the parsed control status will be applied to the client's site layout in real time; S55. Use data binding technology to synchronize the parsed control state with the client site map. The data synchronization process dynamically adjusts the display priority and layout hierarchy of the control. S56. Configure real-time monitoring and error correction mechanisms in the synchronization process to monitor packet loss, delays, or potential errors in data transmission. When an anomaly is detected, trigger the adaptive backtracking and retransmission mechanism to adjust the transmission strategy and form a site map with dynamic response capabilities.

6. The method for compiling a station yard map for integrating driving prompts according to claim 1 is characterized in that: The S6 specifically includes: S61. Acquire real-time synchronous data, including control status, station layout, and user interaction information, initialize the path search module, set the starting node and target node of the path search, and configure the path search parameters, including the initial temperature, cooling rate, and path evaluation criteria; S62. Initiate a simulated annealing tree search algorithm, expand the current path as the initial path, and randomly select adjacent nodes in the search tree structure to generate candidate paths; the initial path is random and is used to explore diverse solutions in the path space; for each candidate path, calculate a path weight as a measure of path quality; the path weight is determined by a combination of path length, path node priorities, and specific constraints; S63. After the path evaluation, the acceptance probability function of the simulated annealing algorithm is used to decide whether to accept the new path. The acceptance probability is controlled by the current temperature. The higher the temperature, the greater the acceptance probability. As the iteration proceeds, the temperature gradually decreases and the acceptance probability decreases, causing the search process to gradually transition from extensive exploration to local fine search. S64. During the path generation and path selection phases, the constraint propagation algorithm is used to perform conflict detection on the path. During the detection process, all nodes and edges in the path are scanned to determine whether there is a conflict with fixed controls, preset path restrictions, or dynamic elements in the station map. If a conflict is detected, the constraint propagation algorithm identifies and marks the neighboring areas of the conflicting nodes. S65. After the path adjustment mechanism is activated, an alternative node adjacent to the conflicting node is reselected to repair the path. The adjustment process is iterated in conjunction with the simulated annealing algorithm. By continuously updating and recalculating the path quality, the entire path is gradually optimized to achieve a conflict-free optimal route. S66. After the path optimization and adjustment are completed, the final optimal approach path will be integrated into the station map layout, and the station map will be updated in real time to display the path planning results.

7. The method for compiling a station yard map for integrating driving prompts according to claim 1 is characterized in that: The S7 specifically includes: S71, receiving a station map with dynamic response capabilities and a generated optimal approach path, and loading them into an augmented reality module to provide a basis for virtual overlay display of the path and controls; S72. Apply image processing technology to pre-process the station map and use image enhancement and edge detection algorithms to optimize image clarity; S73. Use feature calibration technology to identify key locations and reference points in the physical station field, and use a scale-invariant feature transformation algorithm to calibrate and match the features in the physical station field; S74. Use the Perspective-n-Point algorithm to perform real-time pose estimation of the device, combining the calibration features of the physical station and the parameters of the camera equipment to determine the real-time position and perspective of the device, ensuring the correct alignment of the virtual path and controls at different perspectives; S75. Projecting the optimal route path and controls into the augmented reality environment through a rendering engine, achieving superimposed display of the virtual path and controls on the physical station. The rendering process includes dynamic adjustment of the path lines and real-time response of the controls. S76. Test and verify the stability and responsiveness of paths and controls in the augmented reality environment, adjust display parameters and rendering logic, and provide intuitive path and station views.

Images