Enclosed scene map identification dynamic path planning method and system

Abstract

The application provides a closed scene map identification dynamic path planning method and system, the method comprises the following steps: constructing a dynamic map, the dynamic map comprises a static identification layer and a dynamic identification layer, the static identification layer is used for storing fixed geographic information and attribute data in a closed scene, and the dynamic identification layer is embedded in the static identification layer and is configured to be used for marking dynamic information in the scene in real time. Deploy a multi-source sensing device, collect multi-source sensing data in the closed scene in real time, and update the dynamic identification layer based on the multi-source sensing data. In response to a path planning request, an initial path candidate set is generated based on the static identification layer and the updated dynamic identification layer. Each candidate path in the initial path candidate set is evaluated based on a time cost model, and the path with the lowest total cost is selected as the optimal path. Real-time monitoring of the change of the dynamic identification layer is performed, and when dynamic information update is detected, a re-planning process is triggered, a new optimal path is generated, and the display is updated.

Description

Technical Field

[0001] This invention relates to the fields of data processing and map creation technology, and in particular to a method and system for dynamic path planning for map markers in closed scenes. Background Technology

[0002] In the field of intelligent navigation and route planning technology, closed scenarios such as large shopping malls, underground parking lots, factory workshops, and hospital buildings have increasingly prominent needs for efficient and accurate route planning due to their complex internal structures and frequent flow of people and equipment.

[0003] Currently, most path planning systems in closed environments still rely on static maps for path calculation. These static maps display fixed information, such as passageway locations and obstacle distribution, and cannot reflect dynamic changes within the environment in real time, such as temporary obstacles, crowd congestion, and equipment malfunctions. For example, in large shopping malls, static map-based navigation systems may guide customers through congested areas caused by temporary promotions, impacting the navigation experience; in factory workshops, when equipment malfunctions and blocks passageways, intelligent inspection robots cannot plan detours in time, leading to task delays; and in underground parking lots, vehicle navigation systems cannot update parking space occupancy in real time, often guiding drivers to areas with no available spaces, wasting time.

[0004] Furthermore, while some existing technologies attempt to incorporate dynamic factors, shortcomings remain in map labeling. For example, map labeling information may not be detailed enough to accurately distinguish different types of areas and objects; or the integration of map labeling with path planning algorithms may be insufficient, resulting in low accuracy and efficiency in dynamic path planning. Summary of the Invention

[0005] In order to overcome the shortcomings of the prior art, the present invention provides a method and system for dynamic path planning with map marking in closed scenes.

[0006] To achieve the above objectives, a first aspect of the present invention provides a dynamic path planning method for closed scene map identification, comprising: A dynamic map is constructed, which includes a static label layer and a dynamic label layer. The static label layer is used to store fixed geographic information and attribute data within a closed scene, while the dynamic label layer is embedded in the static label layer and is configured to annotate dynamic information within the scene in real time. Deploy multi-source sensing devices to collect multi-source sensing data in closed scenes in real time, and update the dynamic identification layer based on the multi-source sensing data; In response to the route planning request, an initial set of route candidates is generated based on the static identifier layer and the updated dynamic identifier layer; Each candidate path in the initial path candidate set is evaluated based on a time cost model, which includes travel time parameters, congestion level parameters, path length parameters, and path risk coefficients. The total cost of each candidate path is calculated through weighted averages, and the path with the lowest total cost is selected as the optimal path. The system monitors changes in the dynamic identifier layer in real time. When a dynamic information update is detected, it triggers a replanning process to generate a new optimal path and update the display.

[0007] According to an embodiment of the first aspect of the present invention, updating the dynamic identifier layer includes: The collected multi-source sensing data are integrated to form dynamic environmental information. When conflicts occur in the multi-source sensing data, a strategy combining priority adjudication and confidence fusion is adopted to determine the dynamic environmental information. Based on dynamic environmental information, the dynamic identifier layer is updated using a combination of event-driven and timed updates. Event-driven updates are triggered when the perceived data exceeds a preset threshold or a preset event is detected, while timed updates refresh the dynamic identifier layer at fixed intervals.

[0008] According to an embodiment of the first aspect of the present invention, the travel time parameter is calculated based on the length of each segment of the path and the predicted travel speed, wherein the predicted travel speed is determined based on the maximum permissible speed of the segment and the real-time congestion level; the congestion level parameter is the weighted average congestion level of the entire path; and the path risk coefficient is based on the dynamic information marked by the dynamic identification layer.

[0009] According to an embodiment of the first aspect of the present invention, the fixed geographic information in the static identification layer includes channel connectivity relationships. In response to a route planning request, an initial route candidate set is generated based on the starting point and ending point determined by the user, combined with the channel connectivity relationships in the static identification information.

[0010] According to an embodiment of the first aspect of the present invention, each fixed element is identified and assigned a unique identification code within a static identification layer.

[0011] According to an embodiment of the first aspect of the present invention, while responding to path planning, the type of the planning task is obtained, and the weights of each parameter in the time cost model are dynamically set according to the task type.

[0012] A second aspect of this invention provides a dynamic path planning system for closed-scene map labeling, comprising a dynamic map construction module, a multi-source sensing module, a path planning module, a path selection module, and an update and display module. The dynamic map construction module constructs a dynamic map, which includes a static label layer and a dynamic label layer. The static label layer stores fixed geographic information and attribute data within the closed scene, while the dynamic label layer is embedded in the static label layer and configured to annotate dynamic information within the scene in real time. The multi-source sensing module collects multi-source sensing data within the closed scene in real time and transmits it to the dynamic map construction module to update the dynamic label layer. In response to a path planning request, the path planning module generates an initial path candidate set based on the static label layer and the updated dynamic label layer. The path selection module evaluates each candidate path in the initial path candidate set based on a time cost model, which includes travel time parameters, congestion level parameters, path length parameters, and path risk coefficients. The total cost of each candidate path is calculated through weighted averages, and the path with the lowest total cost is selected as the optimal path. The update display module monitors changes in the dynamic identifier layer in real time. When a dynamic information update is detected, it triggers a replanning process, generates a new optimal path, and updates the display.

[0013] According to an embodiment of the second aspect of the present invention, the multi-source sensing module and the dynamic map construction module communicate asynchronously through a message queue or RESTful API, and transmit data using a unified data format.

[0014] According to an embodiment of the second aspect of the present invention, the dynamic map construction module is further configured to update the dynamic identification layer using a combination of event-driven and timed updates, and to determine the dynamic identification information using a priority adjudication and confidence fusion strategy when there is a conflict between multi-source sensing data; the priority adjudication and confidence fusion strategy includes: pre-setting the priority order of different data sources, and when there is a data conflict, performing weighted fusion based on the confidence weight of each data source to obtain the fused dynamic identification information.

[0015] According to an embodiment of the second aspect of the present invention, the dynamic map construction module updates the dynamic identifier layer in real time and triggers the path specification module to generate a new optimal path and send it to the terminal for display.

[0016] In summary, the closed-scene map-marked dynamic path planning method provided by this invention, by setting a dynamic marker layer, enables the closed-scene path planning system to update from static to dynamic, and from passive to active. The dynamic marker layer marks dynamic information such as temporary obstacles, congestion levels, and parking space occupancy in real time. It employs a combination of event-driven and timed updates, interacting in real time with multi-source sensing devices to ensure that map information accurately reflects the current state of the scene. This significantly improves the response speed of path planning while allowing path planning to consider factors such as congestion and risk factors within the closed scene in real time, saving considerable time for navigation and communication within the closed scene. Furthermore, the dynamic marker layer update, which supports multi-source data sensing fusion, not only provides precise sensing of dynamic layer information but also greatly enhances the reliability of path planning through the fusion of multi-source data.

[0017] To make the above and other objects, features and advantages of the present invention more apparent and understandable, preferred embodiments are described below in detail with reference to the accompanying drawings. Attached Figure Description

[0018] Figure 1 The diagram shown is a flowchart of a dynamic path planning method for closed scene map identification provided in an embodiment of the present invention.

[0019] Figure 2 The diagram shown is a principle block diagram of a closed scene map identification dynamic path planning system provided in an embodiment of the present invention. Detailed Implementation

[0020] Existing navigation systems based on static maps cannot reflect dynamic changes in the scene in real time, resulting in unreasonable path planning, excessive path costs, and even robot task failure in intelligent scenarios.

[0021] In view of this, such as Figure 1As shown, the closed-scene map-based dynamic path planning method provided in this embodiment includes: constructing a dynamic map, which includes a static label layer and a dynamic label layer. The static label layer is used to store fixed geographic information and attribute data within the closed scene, and the dynamic label layer is embedded in the static label layer and configured to annotate dynamic information within the scene in real time (step S10). Deploying multi-source sensing devices to collect multi-source sensing data within the closed scene in real time, and updating the dynamic label layer based on the multi-source sensing data (step S20). Responding to a path planning request, generating an initial path candidate set based on the static label layer and the updated dynamic label layer (step S30). Evaluating each candidate path in the initial path candidate set based on a time cost model, which includes travel time parameters, congestion level parameters, path length parameters, and path risk coefficients; calculating the total cost of each candidate path through weighted averages, and selecting the path with the lowest total cost as the optimal path (step S40). Monitoring changes in the dynamic label layer in real time, and triggering a replanning process when dynamic information updates are detected, generating a new optimal path and updating its display (step S50).

[0022] This embodiment uses the application of a dynamic path planning method based on a closed scene map to a large underground parking lot as an example. The parking lot has 5 floors, contains more than 1,000 parking spaces, has complex internal passages, frequent vehicle flow, and often experiences temporary vehicle congestion and rapid changes in parking space occupancy.

[0023] Step S10 involves constructing a dynamic map labeling system that includes a static labeling layer and a dynamic labeling layer. This step includes: Construct a static identification layer: Pre-build a detailed electronic map of each floor of the parking lot and mark it with fixed geographical information. Fixed geographical information includes fixed elements such as passage distribution, load-bearing wall locations, entrances and exits, elevator shafts, and fixed obstacles (such as pillars). Assign a unique identification code to each fixed element and store its attribute data, such as passage width, height limit, turning radius, maximum permissible speed, and direction of travel.

[0024] Construct a dynamic labeling layer: Embed dynamic elements within each floor of the static labeling layer to label dynamic information within the scene in real time. Dynamic elements mainly include: parking space occupancy status (vacant or occupied, represented by 0 or 1); passageway congestion level (usually represented by a continuous value between 0 and 1, where 0 is completely unobstructed and 1 is completely congested); the location, impact range, and expected duration of temporary obstacles (such as disabled vehicles, cleaning areas, and construction cones), which can be quantified by associating temporary fault absence markers with static elements (such as passageway locations) and embedding time parameters; and equipment operating status (such as whether elevators are operating normally, whether roller shutters are closed, etc., which can also be indicated using 0 or 1).

[0025] Next, in step S20, multi-source sensing devices are deployed inside the parking lot to acquire dynamic environmental information in real time. Specifically, high-definition cameras can be installed at entrances, intersections, and key areas to monitor vehicle density, speed, and temporary obstacles using image recognition algorithms; geomagnetic sensors or ultrasonic detectors can be deployed above each parking space to detect its occupancy status in real time; infrared sensors or millimeter-wave radar can be deployed in congested areas to detect lane occupancy; and sensors can be installed in intelligent devices (such as inspection robots and automated forklifts) to report their location, operating status, and encountered obstacles. Furthermore, user terminals (such as mobile apps) allow users to actively report congestion information or temporary obstacles as an auxiliary data source. All sensing devices continuously collect data and transmit it to the dynamic map building module via wired or wireless networks to update the dynamic identification layer's status in real time.

[0026] Specifically, the closed scene map labeling dynamic path planning method provided in this embodiment processes each piece of sensing data after obtaining multi-source sensing data before transmitting it to the dynamic map construction module. The data processing includes: Data alignment and cleaning: First, timestamp alignment (based on the system master clock, with an allowable error within ±50ms) and spatial coordinate unification (converting all data to a unified parking lot plane coordinate system) are performed on data from different devices; then, outliers are removed using the 3σ principle to ensure data quality.

[0027] Image data processing: Background modeling (e.g., Gaussian mixture model) is performed on the camera video stream to separate moving objects in the foreground. Then, a lightweight object detection algorithm (such as YOLOv5s) is used to identify obstacle types (people, vehicles, cones, etc.) and estimate the area density, which is then converted into a congestion level. Simultaneously, the pixel coordinates of the detected objects are mapped to the map coordinate system using a homography matrix to obtain their precise locations.

[0028] Sensor data filtering: Kalman filtering is used to smooth the distance and angle data returned by infrared sensors, millimeter-wave radar, etc., to remove measurement noise and obtain a stable target motion trajectory.

[0029] Data fusion and distribution: The processed multi-source data is encapsulated in a unified format and transmitted to the dynamic map construction module via asynchronous communication methods such as message queues (e.g., MQTT) or RESTful APIs to achieve real-time updates of the dynamic identifier layer. However, this invention does not impose any limitations on this.

[0030] In this embodiment, updating the dynamic identifier layer in step S20 includes: Step S201: The collected multi-source sensing data is fused to form dynamic environmental information. When conflicts occur in the multi-source sensing data, a strategy combining priority adjudication and confidence fusion is used to determine the dynamic environmental information. When conflicts arise in the multi-source sensing data (e.g., camera analysis indicates a channel is congested, while infrared sensors show it is unobstructed), the dynamic map construction module uses a priority adjudication and confidence fusion strategy to determine the final dynamic identification information. Specifically, based on a pre-set priority order (priority from high to low: fixed sensors (e.g., geomagnetic sensors, infrared detectors) > visual recognition data > device status data > user terminal feedback), data from high-priority sensing devices (i.e., infrared detectors) is initially adopted. Based on this, the historical confidence level of this priority is comprehensively considered. If its confidence level exceeds a set confidence threshold, its data is ultimately determined as dynamic environmental information; otherwise, data from the next lower priority sensing device is considered. However, this invention does not impose any limitations on this. In other embodiments, priorities can also be quantified into weights (e.g., fixed sensor weight 0.5, visual data weight 0.3, device data weight 0.15, user feedback weight 0.05), and then combined with confidence level to calculate a comprehensive weight and perform a weighted average.

[0031] Step S202: Based on the determined dynamic environmental information, the dynamic labeling layer is updated using a combination of event-driven and timed updates. Event-driven updates are triggered when perceived data exceeds a preset threshold (e.g., congestion exceeds threshold 0.7, parking space status changes, obstacles appear) or a preset event is detected. Timed updates refresh the dynamic labeling layer at fixed intervals.

[0032] In the closed scene map marker dynamic path planning method provided in this embodiment, the dynamic marker layer interacts with multi-source sensing devices in real time through a combination of event-driven and timed updates, ensuring that the map marker information can accurately reflect the current state of the scene and provide a dynamic data foundation for subsequent path planning.

[0033] In response to the user's input request for route planning, step S30 generates an initial set of candidate routes based on the static identifier layer and the updated dynamic identifier layer. Possible implementations of this step include: Step 301: Receive the path planning request submitted by the user, which includes the coordinates of the starting point and the ending point. Step 302: Obtain current map identifier information: Obtain the latest static and dynamic identifier layer data from the dynamic map to form the current environment map; Step 303: Generate an initial candidate set of paths: The fixed geographic information within the static identifier layer includes channel connectivity relationships. In response to the path planning request, based on the user-defined start and end points and the channel connectivity relationships in the static identifier information, multiple feasible paths from the start point to the end point are generated using a classic path search algorithm (such as the A* algorithm), forming a candidate set.

[0034] Next, step S40 will invoke the time cost model to evaluate multiple feasible paths within the candidate set to select the optimal path. The time cost model considers the following factors: Predicted travel time T: Divide each feasible path into segments based on the channel, with each segment having a length of... l j Divide by the predicted traffic speed of that section v j The total time is obtained by summing the results. The predicted travel speed is based on the maximum permissible speed for that segment. v max and real-time congestion levels cong j calculate:

[0035] in, j The first one within the feasible path j Each segment The congestion impact parameter can be set to 0.8.

[0036] Congestion level D: The weighted average of the congestion levels of each segment within the feasible path, with the weights equal to the length of each segment. l j :

[0037]

[0038] in, L This represents the total length of the feasible paths. m This represents the total number of segments within a feasible path.

[0039] Path risk coefficient R: The risk level of each segment is calculated based on dynamic elements such as temporary obstacles, construction areas, and equipment failures marked by the dynamic identification layer. r j (0 represents no risk, 1 represents completely impassable), and then combined into the risk coefficient of the entire path:

[0040] If the risk coefficient of a certain segment is close to 1, then that path is directly eliminated.

[0041] The above four indicators are each multiplied by a preset weighting coefficient ( α, β , c , d The total path cost C is obtained by weighted summation.

[0042] in, α + β + c + d =1.

[0043] Furthermore, the four weight parameters mentioned above can be dynamically preset based on the task type or user preferences selected by the user when making a route planning request. Taking task type as an example, the task type for this route planning is obtained simultaneously, and the aforementioned weight parameters are dynamically adjusted based on the task type. For instance, for a normal navigation task, the following settings can be configured: α =0.4, β =0.3, c =0.2, d =0.1; however, for urgent tasks, the passage time T can be set to have a higher weight, such as α =0.7, β =0.1, c =0.1, d =0.1; For routine inspection tasks of robots, to reduce the risk of obstacles, obstacle-free priority can be selected, that is, matching a higher weight to the path risk coefficient R, such as α =0.2, β =0.2, c =0.2, d =0.4. However, the present invention does not limit this in any way. In other embodiments, the above weighting parameter can also be dynamically set according to user preferences.

[0044] After obtaining the total path cost C through weighted summation, step S40 selects the path with the lowest total cost as the optimal path by sorting and returns the result to the terminal display module to guide the user or intelligent agent to navigate along the selected optimal path. Furthermore, during navigation, step S50 monitors changes in the dynamic identification layer in real time. When a dynamic information update is detected, a replanning process is triggered (i.e., steps S30 and S40 are re-executed), generating a new optimal path and updating the display to ensure that the user or device can obtain the latest path guidance in a timely manner.

[0045] Through practical application testing, based on the closed scene map identification dynamic path planning method provided in this embodiment, the average time for car owners to find the target parking space or exit in the parking lot has been reduced from 15 minutes to 5 minutes, and the vehicle congestion rate in the parking lot has been reduced by 40%, which greatly improves the operational efficiency of the parking lot and the parking experience of car owners, fully demonstrating the practicality and superiority of the present invention.

[0046] Corresponding to the aforementioned closed-scene map-based dynamic path planning method, this embodiment also provides a planning system, which includes a dynamic map construction module 10, a multi-source sensing module 20, a path planning module 30, a path selection module 40, and an update and display module 50. The dynamic map construction module 10 constructs a dynamic map, which includes a static label layer and a dynamic label layer. The static label layer stores fixed geographic information and attribute data within the closed scene, while the dynamic label layer is embedded in the static label layer and configured to annotate dynamic information within the scene in real time. The multi-source sensing module 20 collects multi-source sensing data within the closed scene in real time and transmits it to the dynamic map construction module to update the dynamic label layer. In response to a path planning request, the path planning module 30 generates an initial path candidate set based on the static label layer and the updated dynamic label layer. The path selection module 40 evaluates each candidate path in the initial path candidate set based on a time cost model, which includes travel time parameters, congestion level parameters, path length parameters, and path risk coefficients. The total cost of each candidate path is calculated through weighted averages, and the path with the lowest total cost is selected as the optimal path. The update display module 50 monitors changes in the dynamic identifier layer in real time. When a dynamic information update is detected, it triggers a replanning process, generates a new optimal path, and updates the display.

[0047] Because the closed scene map marking dynamic path planning system provided in this embodiment is different from the one described in the above embodiments... Figure 1 The system corresponding to the closed scene map identification dynamic path planning method in the above embodiment shares the same concept in the execution steps of each module. Therefore, steps S10-50 above can be referred to, and will not be elaborated here.

[0048] In one possible implementation, the multi-source sensing module 20 and the dynamic map building module 10 communicate asynchronously through a message queue or RESTful API, and use a unified data format for data transmission.

[0049] In one possible implementation, the dynamic map construction module 10 is further used to update the dynamic identification layer using a combination of event-driven and timed updates, and to determine the dynamic identification information using a priority adjudication and confidence fusion strategy when multi-source sensing data conflicts. The priority adjudication and confidence fusion strategy includes: pre-setting the priority order of different data sources, and when data conflicts occur, performing weighted fusion based on the confidence weight of each data source to obtain the fused dynamic identification information; as described in step S20 above.

[0050] In one possible implementation, during navigation, the dynamic map building module 10 updates the dynamic identifier layer in real time and triggers the path specification module 30 to generate a new optimal path and send it to the terminal for display.

[0051] The various modules in the closed scene map identification dynamic path planning system can be implemented in whole or in part through software, hardware, or a combination thereof. These modules can be embedded in the terminal so that the processor in the terminal can call and execute the corresponding operations of these modules.

[0052] In summary, the closed-scene map-marked dynamic path planning method provided by this invention, by setting a dynamic marker layer, enables the closed-scene path planning system to update from static to dynamic, and from passive to active. The dynamic marker layer marks dynamic information such as temporary obstacles, congestion levels, and parking space occupancy in real time. It employs a combination of event-driven and timed updates, interacting in real time with multi-source sensing devices to ensure that map information accurately reflects the current state of the scene. This significantly improves the response speed of path planning while allowing path planning to consider factors such as congestion and risk factors within the closed scene in real time, saving considerable time for navigation and communication within the closed scene. Furthermore, the dynamic marker layer update, which supports multi-source data sensing fusion, not only provides precise sensing of dynamic layer information but also greatly enhances the reliability of path planning through the fusion of multi-source data.

[0053] Although the present invention has been disclosed above by way of preferred embodiments, it is not intended to limit the present invention. Any person skilled in the art may make some modifications and refinements without departing from the spirit and scope of the present invention. Therefore, the scope of protection of the present invention shall be determined by the scope of protection claimed in the claims.

Claims

1. A method for dynamic path planning using map markers in a closed scene, characterized in that, include: A dynamic map is constructed, comprising a static labeling layer and a dynamic labeling layer. The static labeling layer is used to store fixed geographic information and attribute data within a closed scene, and the dynamic labeling layer is embedded in the static labeling layer and configured to annotate dynamic information within the scene in real time. Deploy multi-source sensing devices to collect multi-source sensing data in a closed scene in real time, and update the dynamic identification layer based on the multi-source sensing data; In response to a path planning request, an initial path candidate set is generated based on the static identifier layer and the updated dynamic identifier layer. Each candidate path in the initial path candidate set is evaluated based on a time cost model, which includes travel time parameters, congestion level parameters, path length parameters, and path risk coefficients. The total cost of each candidate path is obtained through weighted calculation, and the path with the lowest total cost is selected as the optimal path. The system monitors changes in the dynamic identifier layer in real time. When a dynamic information update is detected, it triggers a replanning process to generate a new optimal path and update the display.

2. The method for dynamic path planning with map markers in a closed scene according to claim 1, characterized in that, Updating the dynamic identifier layer includes: The collected multi-source sensing data are integrated to form dynamic environmental information. When conflicts occur in the multi-source sensing data, a strategy combining priority adjudication and confidence fusion is adopted to determine the dynamic environmental information. Based on the dynamic environment information, the dynamic identifier layer is updated using a combination of event-driven and timed updates. Event-driven updates are triggered when the perceived data exceeds a preset threshold or a preset event is detected, while timed updates refresh the dynamic identifier layer at fixed intervals.

3. The method for dynamic path planning with map markers in a closed scene according to claim 1, characterized in that, The travel time parameter is calculated based on the length of each segment of the path and the predicted travel speed. The predicted travel speed is determined based on the maximum permissible speed of the segment and the real-time congestion level. The congestion level parameter is the weighted average congestion level of the entire path. The path risk coefficient is based on the dynamic information marked by the dynamic identification layer.

4. The method for dynamic path planning with map markers in a closed scene according to claim 1, characterized in that, The fixed geographic information within the static identifier layer includes channel connectivity relationships. In response to a route planning request, an initial route candidate set is generated based on the user-determined starting point and destination, combined with the channel connectivity relationships in the static identifier information.

5. The method for dynamic path planning with map markers in a closed scene according to claim 1, characterized in that, Within the static identification layer, each fixed element is identified and assigned a unique identification code.

6. The method for dynamic path planning with map markers in a closed scene according to claim 1, characterized in that, While responding to path planning, the type of the planning task is obtained, and the weights of each parameter in the time cost model are dynamically set according to the task type.

7. A dynamic path planning system for map marking in a closed scene, characterized in that, include: A dynamic map construction module constructs a dynamic map, which includes a static labeling layer and a dynamic labeling layer. The static labeling layer is used to store fixed geographic information and attribute data within a closed scene. The dynamic labeling layer is embedded in the static labeling layer and is configured to annotate dynamic information within the scene in real time. The multi-source perception module collects multi-source perception data in a closed scene in real time and transmits it to the dynamic map construction module to update the dynamic identification layer. The path planning module, in response to the path planning request, generates an initial path candidate set based on the static identifier layer and the updated dynamic identifier layer; The route selection module evaluates each candidate path in the initial candidate path set based on a time cost model, which includes travel time parameters, congestion level parameters, path length parameters, and path risk coefficients. The total cost of each candidate path is calculated through weighted averages, and the path with the lowest total cost is selected as the optimal path. The update display module monitors changes in the dynamic identifier layer in real time. When a dynamic information update is detected, it triggers a replanning process to generate a new optimal path and update the display.

8. The closed scene map identification dynamic path planning system according to claim 7, characterized in that, The multi-source perception module and the dynamic map construction module communicate asynchronously through message queues or RESTful APIs, and use a unified data format for data transmission.

9. The closed scene map identification dynamic path planning system according to claim 7, characterized in that, The dynamic map construction module is also used to update the dynamic identification layer by combining event-driven and timed updates, and to determine the dynamic identification information by using a priority adjudication and confidence fusion strategy when there is a conflict in multi-source perception data. The priority adjudication and confidence fusion strategy includes: pre-setting the priority order of different data sources; when data conflicts occur, weighted fusion is performed based on the confidence weight of each data source to obtain the fused dynamic identification information.

10. The closed scene map identification dynamic path planning system according to claim 7, characterized in that, The dynamic map building module updates the dynamic identifier layer in real time and triggers the path specification module to generate a new optimal path and send it to the terminal for display.

Images