Smart park vehicle navigation planning method and system based on digital twinning

Abstract

The invention relates to the technical field of navigation, in particular to a smart park vehicle navigation planning method and system based on digital twinning, and the method comprises the steps: constructing a monitoring area in a smart park, collecting dynamic data and static data in the monitoring area to generate processing data, and obtaining a digital twinning model; analyzing and processing the data, and comparing the monitoring area with the digital twin model to update obstacles in the monitoring area; generating a road topological graph based on the monitoring area, abstracting a video image in the monitoring area into nodes and edges, obtaining a monitoring range proportion of the edges, and judging whether a monitoring blind area exists in the monitoring area; respectively calculating congestion coefficients according to whether monitoring blind areas exist or not, updating a digital twin model through the congestion coefficients in combination with obstacles, and adjusting a navigation path of the vehicle; according to the method, the vehicle deviation condition caused by a monitoring blind area is prevented, physical entities in a dynamic environment are positioned, the conditions such as congestion and obstacles are quickly responded, and the precision and the real-time performance of vehicle navigation are improved.

Description

Technical Field

[0001] The present invention relates to the technical field of navigation, and particularly to a method and system for vehicle navigation planning in a smart park based on digital twin. Background Art

[0002] With the rapid development of smart cities and digital twin technologies, as an important part of urban intelligent management, smart parks have gradually introduced digital twin technologies to achieve efficient management of equipment, environment, and personnel within the park. Especially in the field of vehicle scheduling and navigation, by constructing a digital twin model, it is possible to real-time map the roads, traffic flow, and environmental changes within the park, improving the operation efficiency and safety of vehicles within the park. However, the performance of traditional digital twin navigation systems still has certain limitations in complex environments. Especially when there are dynamic obstacles or monitoring blind spots within the park, existing digital twin navigation systems are difficult to react in a timely manner, affecting the intelligence and reliability of the overall traffic. Moreover, in the context of smart parks, the application of digital twin technologies is not limited to static environmental mapping. More importantly, it is necessary to achieve dynamic and real-time management and control.

[0003] When traditional digital twin systems perform vehicle navigation planning, they mainly rely on data collected by fixedly deployed monitoring systems and sensors. In a smart park, the monitoring system cannot achieve comprehensive coverage of all areas. Especially in some monitoring blind spots or signal occlusion areas, when obstacles or road congestion occur, the system cannot obtain relevant information in real time, resulting in navigation planning deviations or even misleading the navigation. In addition, the types of obstacles within the park are complex and diverse, which may include temporarily parked vehicles, moving personnel, or emergencies. Traditional systems lack the ability to perceive such dynamic information and cannot quickly update the path planning, thus reducing the accuracy and safety of navigation. Summary of the Invention

[0004] In order to solve the technical problem that the existing digital twin navigation model cannot react in real time to obstacles that appear in the monitoring blind spot, resulting in navigation planning deviations, the purpose of the present invention is to provide a method for vehicle navigation planning in a smart park based on digital twin. The specific technical solutions adopted are as follows: Construct a monitoring area within the smart park, collect dynamic data and static data within the monitoring area to generate processed data, and obtain a digital twin model; Analyze the processed data, compare the monitoring area with the digital twin model to update the obstacles within the monitoring area; Generate a road topology map based on the monitoring area, abstract the video images within the monitoring area into nodes and edges, obtain the proportion of the monitoring range of the edges, and determine whether there are monitoring blind spots within the monitoring area; Calculate the congestion coefficient according to the existence of monitoring blind spots, update the digital twin model by combining the congestion coefficient with obstacles, and adjust the navigation path of the vehicle.

[0005] Preferably, construct a monitoring area in the smart park, collect dynamic and static data in the monitoring area to generate processed data, and obtain a digital twin model, including: Deploy camera devices and sensors in the smart park to construct a monitoring area. Obtain the dynamic and static data in the monitoring area through the data acquisition module from the camera devices and sensors, and transmit them to the digital twin module to generate processed data. Combine with 3D modeling technology to construct a virtual model of the physical environment and real-time operating status in the smart park, and obtain a digital twin model through data-driven and simulation analysis.

[0006] Preferably, analyze the processed data and compare the monitoring area with the digital twin model to update the obstacles in the monitoring area, including: Analyze the processed data through the lane line detection model to obtain lane line information, compare the lane line information with the digital twin model, obtain the matching relationship between the monitoring area and the digital twin model based on the sift image registration algorithm, and update the obstacles in the monitoring area.

[0007] Preferably, generate a road topology map based on the monitoring area, abstract the video images in the monitoring area into nodes and edges, obtain the proportion of the monitoring range of the edge, and determine whether there is a monitoring blind spot in the monitoring area, including: Obtain the monitoring range of the monitoring area in the corresponding digital twin model according to the processed data, generate a road topology map based on the monitoring area, abstract the video images in the monitoring area into nodes and edges, and define the edge as lane line information; Analyze the edge and combine the monitoring range to obtain the proportion of the monitoring range of any edge; Set a judgment threshold. If the proportion of the monitoring range is greater than the judgment threshold, there is no monitoring blind spot; if the proportion of the monitoring range is less than the judgment threshold, there is a monitoring blind spot.

[0008] Preferably, if there is no monitoring blind spot, calculate the congestion coefficient, including: Perform target monitoring through the YOLO model, and count the number of vehicles passing through the edge within the historical time based on the current analysis time to obtain the average speed of any vehicle passing through any edge; Calculate the time difference based on the time when any vehicle enters any edge and the current analysis time and perform normalization processing to obtain the timeliness of the corresponding vehicle, and calculate the congestion coefficient.

[0009] Preferably, calculate the time difference based on the time when any vehicle enters any edge and the current analysis time and perform normalization processing to obtain the timeliness of the corresponding vehicle. The corresponding calculation formula is: The corresponding calculation formula is as follows: ; Wherein, represents the timeliness of the r-th vehicle on the i-th edge; represents the time when the r-th vehicle enters the i-th edge; t represents the current analysis time; norm represents the normalization function; To calculate the congestion coefficient, the corresponding calculation formula is as follows: ; Wherein, represents the congestion coefficient of the i-th edge at the current analysis time t; represents the number of vehicles passing through the i-th edge within the historical time based on the current analysis time t; r represents the number of vehicles; represents the timeliness of the r-th vehicle on the i-th edge; represents the average speed of the r-th vehicle passing through the i-th edge; represents the number of vehicles already existing on the i-th edge when the r-th vehicle enters the i-th edge; h represents the length of the edge; l represents the number of edges, denoted as the number of lanes.

[0010] Preferably, if there is a monitoring blind area, calculating the congestion coefficient includes: analyzing the adjacent edges of the edge with the monitoring blind area, obtaining the congestion coefficient and the monitoring range ratio of the adjacent edges, and obtaining an estimated value of the congestion coefficient of the edge with the monitoring blind area. The corresponding calculation formula is as follows: ; Wherein, represents the estimated value of the congestion coefficient of the j-th edge at the current analysis time t; represents the number of edges adjacent to the j-th edge; z represents the edge adjacent to the monitoring blind area; represents the monitoring range ratio of the z-th edge; represents the congestion coefficient of the z-th edge at the current analysis time t; norm represents the normalization function.

[0011] Preferably, updating the digital twin model by combining the congestion coefficient with obstacles and adjusting the navigation path of the vehicle includes: Obtaining the initial edge weight and the initial update time interval of any edge through the digital twin model, and combining the congestion coefficient with the initial edge weight and the initial update time interval to respectively obtain the updated edge weight and the update time interval based on the current analysis time; Updating the digital twin model according to the updated edge weight and the update time interval, avoiding obstacles, and adjusting the navigation path of the vehicle.

[0012] Preferably, the updated edge weights based on the current analysis time are obtained, and the corresponding calculation formula is: Calculate the updated edge weights of the edges without monitoring blind spots: ; wherein, represents the updated edge weight of the i-th edge at the current analysis time t; represents the congestion coefficient of the i-th edge at the current analysis time t; represents the initial edge weight of the i-th edge; norm represents the normalization function; Similarly, calculate the updated edge weights of the edges with monitoring blind spots; The updated update time interval based on the current analysis time is obtained, and the corresponding calculation formula is: Calculate the updated time interval of the edges without monitoring blind spots: ; wherein, represents the updated update time interval of the i-th edge; represents the congestion coefficient of the i-th edge at the current analysis time t; represents the initial update time interval; norm represents the normalization function; Similarly, calculate the updated update time intervals of the edges with monitoring blind spots.

[0013] To solve the above problems, the present application also provides a digital-twin-based intelligent park vehicle navigation planning system. The system stores program data. When the program data is executed, it combines a camera device, a sensor, a data acquisition module, a digital-twin module, a vehicle terminal, a navigation decision module, a road information database, and a road detection module to implement the digital-twin-based intelligent park vehicle navigation planning method as described in any one of the foregoing.

[0014] The present invention has the following beneficial effects: 1. The present application generates processed data from the static data and dynamic data obtained in real time by the camera device and the sensor, and combines the digital-twin model to dynamically update the road information in the park, quickly respond to congestion, obstacles and other conditions, and improve the real-time performance of navigation; By matching the monitoring area with the digital-twin model, it judges whether there is a monitoring blind spot, and obtains the corresponding congestion coefficients respectively, effectively reducing the navigation error that may be caused in each monitoring blind spot; In addition, the lane line detection and YOLO object detection models using deep learning are used to monitor different positions of the twin model in real time, and combined with the vehicle deviation of the planned path, the obstacle conditions of different vehicle navigations are updated to achieve accurate identification and positioning of vehicles and obstacles in the dynamic environment, further improving the navigation accuracy, quickly responding to congestion, obstacles and other conditions, and improving the real-time performance of navigation; In addition, the present application also considers the needs of different types of vehicles and provides personalized services such as fixed parking space navigation and free parking space search, comprehensively improving the user experience.

[0015] 2. The intelligent park vehicle navigation planning system based on digital twin provided by the present invention has the same beneficial effects as the intelligent park vehicle navigation planning method based on digital twin provided by the present invention, which will not be elaborated here. BRIEF DESCRIPTION OF THE DRAWINGS

[0016] In order to more clearly illustrate the technical solutions and advantages in the embodiments of the present invention or the prior art, the following will briefly introduce the drawings required to be used in the description of the embodiments or the prior art. Obviously, the drawings in the following description are only some embodiments of the present invention. For those of ordinary skill in the art, other drawings can be obtained based on these drawings without creative efforts.

[0017] Figure 1 It is a flowchart of the steps of an intelligent park vehicle navigation planning method based on digital twin provided by an embodiment of the present invention; Figure 2 It is a schematic diagram of nodes and edges of an intelligent park vehicle navigation planning method based on digital twin provided by an embodiment of the present invention; Figure 3 It is a schematic diagram of the structure of an intelligent park vehicle navigation planning system based on digital twin provided by an embodiment of the present invention. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0018] In order to further elaborate on the technical means and effects adopted by the present invention to achieve the intended invention purpose, the following will, in combination with the accompanying drawings and preferred embodiments, elaborate in detail on the specific implementation manners, structures, features and effects of an intelligent park vehicle navigation planning method and system based on digital twin proposed according to the present invention. In the following description, different "one embodiment" or "another embodiment" do not necessarily refer to the same embodiment. In addition, the specific features, structures or characteristics in one or more embodiments can be combined in any suitable form.

[0019] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by those of ordinary skill in the technical field to which the present invention belongs.

[0020] The following will specifically describe the specific solutions of an intelligent park vehicle navigation planning method and system based on digital twin provided by the present invention in combination with the accompanying drawings.

[0021] The prior art usually makes navigation plans based on existing situations and cannot obtain obstacles or congestion phenomena in real time, resulting in the inability to update navigation plans according to emergencies, reducing the safety of navigation. In one embodiment of the present invention, a method for vehicle navigation planning in a smart park based on digital twins is provided. By real-time monitoring of different positions of the twin model and combining the vehicle deviation of the planned path, the obstacle situation of different vehicle navigations is updated. To implement a method for vehicle navigation planning in a smart park based on digital twins, a vehicle navigation planning system based on digital twins is provided. This system is essentially a software system composed of modules that implement corresponding functions. Now, the specific steps in this method will be introduced in detail.

[0022] Please refer to Figure 1 , which shows a flowchart of the steps of a method for vehicle navigation planning in a smart park based on digital twins provided by an embodiment of the present invention. The method includes: Step S1: Construct a monitoring area in the smart park, collect dynamic data and static data in the monitoring area to generate processed data, and obtain a digital twin model; Step S2: Analyze the processed data, compare the monitoring area with the digital twin model, and update the obstacles in the monitoring area; Step S3: Generate a road topology map based on the monitoring area, abstract the video images in the monitoring area into nodes and edges, obtain the proportion of the monitoring range of the edges, and determine whether there is a monitoring blind area in the monitoring area; Step S4: Calculate the congestion coefficient according to whether there is a monitoring blind area, update the digital twin model by combining the congestion coefficient with the obstacles, and adjust the navigation path of the vehicle.

[0023] For better illustration, digital twin refers to creating a virtual copy of a physical entity to simulate, analyze, and predict various situations and behaviors in the real world. It is to use means such as computer simulation, data analysis, and artificial intelligence to operate and manage this virtual copy to achieve the purpose of simulating, analyzing, and predicting various complex situations and behaviors in the real world. A smart park is an innovative concept that integrates modern information technology and traditional park management. By using advanced technologies such as the Internet of Things, big data analysis, and cloud computing, it realizes the efficient allocation and management of various resources in the park. Therefore, vehicle navigation planning in a smart park based on digital twins aims to create a virtual smart park model, obtain the driving path of vehicles in the park, analyze traffic flow, real-time update congestion situations, obstacles, and possible emergencies on the road, update the initial digital twin model, and formulate the optimal navigation route to help vehicles in the smart park drive more efficiently and safely, reduce waiting time, and improve the overall traffic efficiency.

[0024] Preferably, in this embodiment, obstacles refer to objects that should not appear on the lane lines, which may interfere with or endanger the normal driving of the vehicle. They may be stationary, such as construction waste left on the road, fallen cargo, and accident debris that have not been cleared in time, or they may be active, such as animals, pedestrians, or even other out-of-control vehicles that suddenly cross the road. The presence of obstacles will not only hinder the smooth flow of traffic, but may also cause traffic accidents and pose a threat to the safety of drivers and pedestrians. Therefore, in order to better obtain the navigation plan of the vehicle, the obstacles in the smart park also need to be updated in real time.

[0025] Furthermore, step S1 includes: Cameras and sensors are deployed in the smart park to build a monitoring area. The data collected by the cameras and sensors is obtained through the data acquisition module to obtain dynamic data and static data in the monitoring area, and the data is transmitted to the digital twin module to generate and process data. Combined with 3D modeling technology, a virtual model of the physical environment and real-time operating status in the smart park is constructed, and a digital twin model is obtained through data-driven and simulation analysis.

[0026] As an optional implementation manner, in this embodiment, the imaging device refers to a camera.

[0027] It can be explained that the sensor is used to obtain static data of environmental information, that is, the sensor detects parameters such as temperature, humidity, light intensity, air quality, etc. in the environment through various sensing technologies to provide real-time environmental information parameters.

[0028] Specifically, cameras and sensors are deployed at major road nodes, key intersections and parking areas to build monitoring areas, collect real-time video images to obtain dynamic data such as vehicles, pedestrians and obstacles, and static data such as roads, buildings, parking spaces and traffic facilities, and obtain static and dynamic data from cameras and sensors through the data acquisition module for integration to obtain environmental perception information within the smart park, which is then transmitted to the digital twin module to generate processed data. Combined with three-dimensional modeling technology, a virtual model of the physical environment within the smart park is created using a computer to reflect the real-time operating status. Through data-driven and simulation analysis, that is, using processed data to identify patterns and phenomena, hidden relationships in the data are discovered, and systems and processes in the real world are simulated to obtain a digital twin model for subsequent vehicle navigation planning.

[0029] Furthermore, step S2 includes: The lane line information is obtained by analyzing and processing data through the lane line detection model, and the lane line information is compared with the digital twin model. The matching relationship between the monitoring area and the digital twin model is obtained based on the SIFT image registration algorithm, and the obstacles in the monitoring area are updated.

[0030] For better illustration, through a lane line detection model based on deep learning, the LaneNet (Lane Detection Network) is sampled to accurately obtain lane line information. That is, through a convolutional neural network architecture, it can handle various complex road scenarios, including curved, intersecting, and multi-lane situations. LaneNet not only improves the detection accuracy but also reduces the false alarm rate, ensuring the safety of vehicles during driving.

[0031] The sift (Scale-Invariant Feature Transform) image registration algorithm is an algorithm for image feature extraction and matching. By detecting the key parts in the image and calculating the directions, scales, and descriptions of these key parts, it generates feature descriptions with scale invariance and rotation invariance, enabling accurate matching even in complex environments such as changes in perspective, lighting conditions, or noise interference.

[0032] Specifically, based on the lane line detection model to obtain lane line information, the road area within the monitoring area is obtained. That is, an image video is obtained through a camera device, and an RGB (Red, Green, Blue) color image is input. That is, an image of other colors is generated by combining lights of three colors, red, green, and blue, with different intensities. A binary mask of the same size as the input image is output, which is used to represent the pixel area of the lane line, and each pixel value is 0 or 1, where 0 represents the background and 1 represents the lane line.

[0033] The obtained lane line is compared with the digital twin model. Based on the sift image registration algorithm, representative feature points on the lane line are selected, that is, areas with invariance that can be accurately identified and matched even under different perspectives and lighting conditions. It is registered with the top-view projection result of the digital twin model to determine the specific position and direction of the lane line in the monitoring area, obtaining the matching relationship between the monitoring area and the digital twin model, ensuring that the digital twin model can accurately reflect the road conditions in the smart campus to assist in realizing the real-time update of obstacles, improving the accuracy of vehicle navigation in the smart campus, and the efficiency and reliability of the intelligent transportation system.

[0034] It can be understood that through the matching relationship between the monitoring area and the digital twin model, the correlation characteristics between multiple monitoring areas in the smart campus can be obtained, which is conducive to subsequent analysis of whether there are monitoring blind spots, that is, potential monitoring loopholes, in each road section.

[0035] Please refer to Figure 2, which shows a schematic diagram of nodes and edges of a digital-twin-based intelligent park vehicle navigation planning method provided by an embodiment of the present invention. Among them, V0, V1, V2, and V3 represent nodes, 1, 2, 3, 4, and 5 represent edges, and the arrows represent the allowed traffic directions.

[0036] Further, in step S3, it includes: Step S31: Obtain the monitoring range of the monitoring area in the corresponding digital-twin model according to the processed data, generate a road topology map based on the monitoring area, abstract the video images in the monitoring area into nodes and edges, and define the edges as lane line information.

[0037] An explanation is made that the road topology map is a graphical display method for representing the road network structure, which represents the intersections and road segments of the road through nodes and edges; it can intuitively depict the intersections of the road and the connection relationships between various road segments; among them, the nodes represent the key positions of the road, such as traffic points like intersections, parking area entrances, and turning points, etc.; the edges are used to represent the road segments between two nodes, and the edges are lane line information, including length, width, road surface type, and allowed traffic directions; when an edge is two-way traffic, since the road obstacles and congestion conditions are different when the vehicle travels in different directions, this road section is divided into two edges in different directions.

[0038] Step S32: Analyze the edges and combine with the monitoring range to obtain the proportion of the monitoring range of any edge.

[0039] Specifically, obtain the monitoring range of the monitoring area in the corresponding digital-twin model according to the processed data, that is, manually calibrate the monitoring range of any edge; the proportion of the monitoring range refers to the proportion of the monitoring range of any edge in the total length of the corresponding edge, denoted as, representing the proportion of the monitoring range of the th edge.

[0040] Step S33: Set a judgment threshold. If the proportion of the monitoring range is greater than the judgment threshold, there is no monitoring blind area; if the proportion of the monitoring range is less than the judgment threshold, there is a monitoring blind area.

[0041] As an optional implementation manner, in this embodiment, the judgment threshold is.

[0042] Specifically, when, it means that the monitoring range of the th edge in the road topology map is large and there is basically no monitoring blind area. At this time, path planning can be carried out with confidence; on the contrary, when, it means that the monitoring range of the th edge in the road topology map is small and there may be a monitoring blind area. Therefore, when navigating the vehicle, it is necessary to focus on the blind area range in the corresponding road section and take corresponding measures to identify and avoid potential monitoring blind areas; that is, distinguish each road section in the intelligent park through the size of the proportion of the monitoring range to more accurately navigate the vehicle effectively.

[0043] Understandably, there is no monitoring blind spot, indicating that the proportion of the monitoring range is greater than the judgment threshold, which means that the monitoring coverage of the side corresponding to the proportion of the monitoring range is relatively high. At this time, the congestion coefficient can be obtained according to the real-time monitoring.

[0044] Furthermore, in step S4, if there is no monitoring blind spot, calculating the congestion coefficient includes: Step S411: Perform target monitoring through the YOLO model, and based on the current analysis time, count the number of passing vehicles on the side within the historical time to obtain the average speed of any vehicle passing any side.

[0045] For better illustration, the YOLO (You Only Look Once) model is a real-time object detection algorithm. It transforms the object detection task into a single regression problem and performs object detection by directly predicting the bounding box and class probability in the image, that is, dividing the image into grids, and each grid is responsible for predicting the object whose center point falls within the grid to improve the detection speed.

[0046] As an optional implementation manner, in this embodiment, the historical time is within the past 10 minutes based on the current analysis time.

[0047] Specifically, for the road section with no monitoring blind spot or a small monitoring blind spot, through the obtained monitoring video, use the YOLO model to perform object detection on each frame of the monitoring video, and extract the position and quantity of vehicles; among them, the output result of YOLO includes the bounding box coordinates of each vehicle, the confidence of each bounding box, and the class label; then count the number of passing vehicles on the side within the historical time to obtain the average speed of any vehicle passing any side, and record the average speed of the kth vehicle passing the jth section as.

[0048] Step S412: Calculate the time difference according to the moment when any vehicle enters any side in combination with the current analysis time and perform normalization processing to obtain the timeliness of the corresponding vehicle, and calculate the congestion coefficient.

[0049] Furthermore, in step S412, calculating the time difference according to the moment when any vehicle enters any side in combination with the current analysis time and performing normalization processing to obtain the timeliness of the corresponding vehicle, the corresponding calculation formula is: ; where represents the timeliness of the rth vehicle on the ith side; Denote the moment when the $r$-th vehicle enters the $i$-th edge; $t$ represents the current analysis moment; norm represents the normalization function. It should be noted that timeliness refers to the efficiency and speed of a vehicle to complete a task or reach a destination within a specific time period, which reflects the running state of the vehicle on the road. By analyzing timeliness, traffic flow, road congestion, and vehicle scheduling can be optimized to improve the operation efficiency of the entire digital twin system.

[0050] Calculate the congestion coefficient, and the corresponding calculation formula is: ; Wherein, Denote the congestion coefficient of the $i$-th edge at the current analysis moment $t$; Denote the number of vehicles passing through the $i$-th edge within the historical time based on the current analysis moment $t$; $r$ represents the number of vehicles; Denote the timeliness of the $r$-th vehicle on the $i$-th edge; Denote the average speed of the $r$-th vehicle passing through the $i$-th edge; Denote the number of vehicles already existing on the $i$-th edge when the $r$-th vehicle enters the $i$-th edge; $h$ represents the length of the edge; l Denote the number of edges, denoted as the number of lanes. It should be noted that to ensure the significance of the calculation results, in the fractional operation of this embodiment of the present invention, when the denominator is 0, a tuning factor greater than 0 needs to be added to the denominator to prevent the denominator from being 0. The value of the tuning factor is set by the implementer according to the actual situation, and this application does not make special restrictions.

[0051] It should be noted that the congestion coefficient is an indicator to measure the degree of road congestion, comprehensively considering factors such as vehicle timeliness, the number of passing vehicles, and traffic flow; by analyzing the congestion coefficient, the real-time traffic conditions of the road can be better understood, laying a foundation for better planning of vehicle paths subsequently.

[0052] It can be understood that if there is a monitoring blind area on the edge and the monitoring blind area is relatively large, obtain the relevant connection between the currently analyzed edge and its adjacent edges, estimate the congestion coefficient of the edge with the monitoring blind area to obtain an estimated value of the congestion coefficient, that is, expand the monitoring area of the monitoring blind area to make up for the information loss in the monitoring blind area, and then indirectly infer the congestion coefficient of the edge with the monitoring blind area to assist the digital twin system in making a new navigation plan.

[0053] Furthermore, in step S4, if there is a monitoring blind area, calculating the congestion coefficient includes: Analyze the adjacent edges of the edge with the monitoring blind area, obtain the congestion coefficient and the proportion of the monitoring range of the adjacent edges, and obtain an estimated value of the congestion coefficient of the edge with the monitoring blind area. The corresponding calculation formula is: ; Among them, represents the estimated value of the congestion coefficient of the j-th edge at the current analysis time t; represents the number of edges adjacent to the j-th edge; z represents the edge adjacent to the area with monitoring blind spots; represents the proportion of the monitoring range of the z-th edge; represents the congestion coefficient of the z-th edge at the current analysis time t; norm represents the normalization function.

[0054] Specifically, in this embodiment, the j-th edge is defined as the road section with monitoring blind spots. First, the starting point and the ending point of the j-th edge are obtained, and the edges with the starting point of the j-th edge as the ending point and the edges with the ending point of the j-th edge as the starting point are selected. The boundaries that meet such conditions are defined as the edges adjacent to the area with monitoring blind spots; for better illustration, in Figure 2 , if the 4th edge is analyzed, according to the representation of the direction, the associated edges are the 3rd and 5th edges; then, the estimated value of the congestion coefficient of each edge is obtained in combination with the adjacent edges.

[0055] It can be understood that during the process of vehicle navigation in a traditional smart park, the navigation selects the optimal path based on the road topology map generated by the digital twin system. If the traditional digital twin model is not updated in a timely manner, it will directly affect the final vehicle navigation result; therefore, according to the obtained congestion coefficient or the estimated value of the congestion coefficient of each edge, the obstacles of any road section are updated in real time, and then the congestion coefficient or the estimated value of the congestion coefficient is analyzed to complete the update of the traditional digital twin model.

[0056] Furthermore, in step S4, the digital twin model is updated by combining the congestion coefficient with the obstacles, and the navigation path of the vehicle is adjusted, including: Step S421: Obtain the initial edge weight and the initial update time interval of any edge through the digital twin model, and combine the congestion coefficient with the initial edge weight and the initial update time interval to obtain the updated edge weight and the update time interval based on the current analysis time respectively.

[0057] It can be explained that in this embodiment, the initial edge weight is obtained according to the length of the edge. Among them, the edge weight refers to the weight value of the edge, and the initial edge weight is obtained by first obtaining the actual length of the edge, and then converting the actual length of the edge into an edge weight value to reflect the actual length of the edge in the algorithm, and further reflecting factors such as traffic congestion and road conditions; the initial update time interval refers to the time interval between data synchronization and update operations when the digital twin system starts and begins to run, and is used to illustrate the initial state of navigation.

[0058] Optionally, in this embodiment, the initial update time interval is 30 min.

[0059] Step S422: Update the digital twin model according to the updated edge weights and update time intervals, avoid obstacles, and adjust the navigation path of the vehicle.

[0060] Further, in step S421, the updated edge weights obtained based on the current analysis time are calculated using the following formula: Calculate the updated edge weights for edges without monitoring blind spots: ; where represents the updated edge weight of the i-th edge at the current analysis time t; represents the congestion coefficient of the i-th edge at the current analysis time t; represents the initial edge weight of the i-th edge; norm represents the normalization function.

[0061] Similarly, calculate the updated edge weights for edges with monitoring blind spots. Specifically, the corresponding formula is:

[0062] where represents the updated edge weight of the j-th edge at the current analysis time t; represents the estimated congestion coefficient of the j-th edge at the current analysis time t; represents the initial edge weight of the j-th edge.

[0063] Further, the updated update time interval obtained based on the current analysis time is calculated using the following formula: Calculate the updated time interval for edges without monitoring blind spots: ; where represents the updated update time interval of the i-th edge; represents the congestion coefficient of the i-th edge at the current analysis time t; represents the initial update time interval; norm represents the normalization function.

[0064] Similarly, calculate the updated update time interval for edges with monitoring blind spots. Specifically, the corresponding formula is: ; where represents the updated update time interval of the j-th edge; represents the estimated congestion coefficient of the j-th edge at the current analysis time t.

[0065] Specifically, based on the foregoing, obtain the updated edge weights and update time intervals of each edge in the smart park, and then update the digital twin model to adjust the optimal path planning for the vehicle to reach the destination from the current position in real time. That is, adjust the edge weights in the road topology map based on the updated edge weights and update time intervals, combine with the Dijkstra algorithm, and avoid obstacles to obtain the updated navigation path of the vehicle in real time to adjust the vehicle heading in real time. Among them, the Dijkstra algorithm is an algorithm used to find the optimal path in the road topology map. It starts from the source point, gradually expands the shortest path tree, selects an unvisited node that is closest to the current node each time, adds it to the shortest path tree, and updates the distances from other nodes to the source point. This process continues until all nodes have been visited or the target node is reached.

[0066] Preferably, when different vehicles enter the smart park, according to the vehicle recognition system in the monitoring area, obtain the license plate information, and analyze whether there is a fixed vehicle parking position for the current vehicle inside the smart park. If there is, perform intelligent navigation on the updated digital twin model in the smart park to determine the starting point and the target position point through the foregoing vehicle planning method to ensure that the vehicle can reach the predetermined position efficiently and accurately. On the contrary, if there is no fixed parking space in the smart park, obtain the available parking spaces based on the digital twin model, and then navigate to the available parking spaces to enable the vehicle to park smoothly while maintaining the order and efficiency inside the smart park.

[0067] It can be understood that the present application generates processed data from the static data and dynamic data obtained in real time by the camera device and the sensor, combines with the digital twin model, can dynamically update the road information in the park, quickly respond to situations such as congestion and obstacles, and improve the real-time performance of navigation. By matching the monitoring area with the digital twin model, judge whether there is a monitoring blind area, and obtain the corresponding congestion coefficients respectively, effectively reducing the navigation error that may be caused in each monitoring blind area. In addition, adopt the lane line detection of deep learning and the YOLO target detection model, monitor different positions of the twin model in real time, combine with the deviation of the vehicle from the planned path, update the obstacle situation of different vehicle navigations, realize the accurate identification and positioning of vehicles and obstacles in the dynamic environment, further improve the navigation accuracy, quickly respond to situations such as congestion and obstacles, and improve the real-time performance of navigation. In addition, the present application also considers the needs of different types of vehicles and provides personalized services such as fixed parking space navigation and available parking space search to comprehensively improve the user experience.

[0068] Please refer to Figure 3 , which shows a schematic structural diagram of a vehicle navigation planning system for a smart park based on digital twins provided by an embodiment of the present invention.

[0069] An embodiment of the present invention proposes a digital-twin-based intelligent park vehicle navigation planning system. The system stores program data. When the program data is executed, in combination with a camera device, sensors, a data acquisition module, a digital-twin module, a vehicle terminal, a navigation decision module, a road information database, and a road detection module, it realizes a digital-twin-based intelligent park vehicle navigation planning method as described in the previous embodiment. This system has the same beneficial effects as the previously provided digital-twin-based intelligent park vehicle navigation planning method and will not be elaborated here.

[0070] For better illustration, the camera device can capture a continuous sequence of images. After digital processing, the images are converted into electrical signals to obtain real-time video stream data. The sensors can sense and respond to specific types of input signals and convert these input signals into electrical signals to monitor environmental changes or device states in real time and generate sensing data. The sensing data is transmitted to the data acquisition module, and after analysis, processed data is obtained. The data acquisition module is responsible for collecting the required information from various different data sources to ensure the accuracy and integrity of the data. The processed data is transmitted to the digital-twin platform to obtain an initial digital-twin model, that is, through the digital-twin platform, the processed data is synchronized and simulated in real time to realize the monitoring, analysis, and optimization of physical entities in the intelligent park.

[0071] Connected to the digital-twin platform are respectively a vehicle terminal, a road information database, and a road detection module. Among them, the vehicle terminal is a navigation terminal, including a GPS (Global Positioning System Module) module, a display screen, a communication module, etc., which is used to indicate the position of the vehicle and receive navigation instructions sent by the digital-twin model and to feedback the vehicle state in real time. The road information database refers to the central server, that is, the cloud computing platform, which is used to store historical data, road information, and dynamic environment information in the intelligent park, and also assists in running the core models and algorithms of the digital-twin module. The storage of dynamic environment information is to respond to various emergencies in a timely manner, avoid potential problems in the intelligent park, and prevent dangers caused by the untimely response of the digital-twin model. The road detection module can detect road congestion and vehicle driving conditions in the intelligent park in real time, predict and analyze obstacles in the monitored area, which is conducive to updating obstacles in the intelligent park and avoiding them during navigation.

[0072] The navigation decision-making module is connected between the vehicle terminal and the road detection module. The congestion coefficient or the estimated value of the congestion coefficient obtained in the road detection module is transmitted to the navigation decision-making module. In the navigation decision-making module, the updated edge weight and the updated time interval are used to update the initial digital twin model, and a new navigation path is sent to the vehicle terminal. Among them, the navigation decision-making module can find the optimal navigation path based on the updated digital twin model and the obstacles updated in real time, in combination with the path planning algorithm and the driving state of the current vehicle.

[0073] Preferably, in this embodiment, the digital twin-based intelligent park vehicle navigation planning system further includes an edge computing device, a wireless communication device, and a user interaction and display module. Among them, the edge computing module is used to process the data of the camera device and the sensor in real time to reduce the data transmission delay. The wireless communication device mainly uses 5G or Wi-Fi 6 communication technology to ensure real-time communication between the vehicles in the intelligent park and the digital twin platform. 5G, that is, the fifth-generation mobile communication technology; Wi-Fi 6, also known as 802.11ax, provides a higher data transmission rate and more efficient network capacity than the previous generation of Wi-Fi technology. The user interaction and display module can display the navigation information, road conditions, and vehicle positions in the current intelligent park to the user, enabling the user to quickly understand through a graphical interface, with strong flexibility and improved user experience.

[0074] It can be understood that when the modules of a digital twin-based intelligent park vehicle navigation planning system are operating, they need to use a digital twin-based intelligent park vehicle navigation planning method provided in the foregoing embodiment. Therefore, whether the method and program data are integrated or different hardware is configured to produce functions similar to the effects achieved by the present invention, they all fall within the protection scope of the present invention.

[0075] It should be noted that the above sequence of the embodiments of the present invention is only for description and does not represent the superiority or inferiority of the embodiments. The processes depicted in the drawings do not necessarily require the specific order or continuous order shown to achieve the desired results. In some embodiments, multitasking and parallel processing are also possible or may be advantageous.

[0076] Each embodiment in this specification is described in a progressive manner. The same or similar parts among the embodiments can be referred to each other, and the key points of each embodiment are the differences from other embodiments.

Claims

1. A vehicle navigation planning method for a smart park based on digital twins, characterized in that: The method comprises: Construct a monitoring area within the smart park, collect dynamic and static data within the monitoring area, generate and process data, and obtain a digital twin model; Analyze and process data, compare the monitored area with the digital twin model, and update obstacles in the monitored area; Generate a road topology map based on the monitoring area, abstract the video images in the monitoring area into nodes and edges, obtain the monitoring range ratio of the edges, and determine whether there is a monitoring blind spot in the monitoring area; The congestion coefficient is calculated according to whether there is a monitoring blind spot. The digital twin model is updated based on the congestion coefficient and obstacles to adjust the vehicle's navigation path.

2. A method for smart park vehicle navigation planning based on digital twins as claimed in claim 1, characterized in that: Construct a monitoring area within the smart park, collect dynamic data and static data within the monitoring area to generate and process data, and obtain a digital twin model, including: Cameras and sensors are deployed in the smart park to build a monitoring area. The data collected by the cameras and sensors is obtained through the data acquisition module to obtain dynamic data and static data in the monitoring area, and the data is transmitted to the digital twin module to generate and process data. Combined with 3D modeling technology, a virtual model of the physical environment and real-time operating status in the smart park is constructed, and a digital twin model is obtained through data-driven and simulation analysis.

3. A method for smart park vehicle navigation planning based on digital twins as claimed in claim 1, characterized in that: Analyze and process data, compare the monitored area with the digital twin model, and update obstacles in the monitored area, including: The lane line information is obtained by analyzing and processing data through the lane line detection model, and the lane line information is compared with the digital twin model. The matching relationship between the monitoring area and the digital twin model is obtained based on the SIFT image registration algorithm, and the obstacles in the monitoring area are updated.

4. A method for smart park vehicle navigation planning based on digital twins as claimed in claim 1, characterized in that: Generate a road topology map based on the monitoring area, abstract the video images in the monitoring area into nodes and edges, obtain the monitoring range ratio of the edges, and determine whether there is a monitoring blind spot in the monitoring area, including: The monitoring range of the monitoring area in the corresponding digital twin model is obtained based on the processed data, and a road topology map is generated based on the monitoring area. The video images in the monitoring area are abstracted into nodes and edges, and the edges are defined as lane line information; Analyze the edge and combine it with the monitoring range to get the monitoring range ratio of any edge; Set a judgment threshold. If the monitoring range accounts for a larger proportion than the judgment threshold, there is no monitoring blind spot. If the monitoring range accounts for a smaller proportion than the judgment threshold, there is a monitoring blind spot.

5. The method for smart park vehicle navigation planning based on digital twins according to claim 1, characterized in that: If there is no monitoring blind spot, calculate the congestion coefficient, including: The YOLO model is used to monitor the target. Based on the current analysis time, the number of vehicles passing through the edge in the historical time is counted to obtain the average speed of any vehicle passing through any edge. According to the time when any vehicle enters any side and the current analysis time, the time difference is calculated and normalized to obtain the timeliness of the corresponding vehicle, and the congestion coefficient is calculated.

6. A method for smart park vehicle navigation planning based on digital twins as claimed in claim 5, characterized in that: According to the time when any vehicle enters any side and the current analysis time, the time difference is calculated and normalized to obtain the timeliness of the corresponding vehicle. The corresponding calculation formula is: ； in, represents the timeliness of the r-th vehicle on the i-th edge; represents the time when the rth vehicle enters the ith edge; t represents the current analysis time; norm represents the normalization function; Calculate the congestion coefficient, the corresponding calculation formula is: ； in, represents the congestion coefficient of the ith edge at the current analysis time t; represents the number of vehicles passing through the i-th edge in the historical time based on the current analysis time t; r represents the number of vehicles; represents the timeliness of the r-th vehicle on the i-th edge; represents the average speed of the r-th vehicle passing through the i-th edge; represents the number of vehicles already on the ith edge when the rth vehicle enters the ith edge; h represents the length of the edge; l Represents the number of edges, recorded as the number of lanes.

7. A method for smart park vehicle navigation planning based on digital twins as claimed in claim 5, characterized in that: If there is a monitoring blind spot, the congestion coefficient is calculated, including: analyzing the adjacent edges of the monitoring blind spot, obtaining the congestion coefficient and monitoring range ratio of the adjacent edges, and obtaining an estimated value of the congestion coefficient of the edge with the monitoring blind spot. The corresponding calculation formula is: ； in, represents the estimated value of the congestion coefficient of the jth edge at the current analysis time t; represents the number of edges adjacent to the jth edge; z represents the edge adjacent to the monitoring blind spot; Indicates the monitoring range proportion of the zth edge; represents the congestion coefficient of the zth edge at the current analysis time t; norm represents the normalization function.

8. A method for smart park vehicle navigation planning based on digital twins as claimed in claim 7, characterized in that: The digital twin model is updated by combining the congestion coefficient with obstacles to adjust the vehicle’s navigation path, including: The initial edge weight and initial update time interval of any edge are obtained through the digital twin model, and the congestion coefficient is combined with the initial edge weight and initial update time interval to obtain the updated edge weight and update time interval based on the current analysis time. The digital twin model is updated according to the updated edge weights and update time intervals to avoid obstacles and adjust the vehicle’s navigation path.

9. The method for smart park vehicle navigation planning based on digital twins as claimed in claim 7, characterized in that: Get the updated edge weight based on the current analysis time. The corresponding calculation formula is: Calculate the updated edge weight of the edge without monitoring blind spot: ； in, represents the updated edge weight of the i-th edge at the current analysis time t; represents the congestion coefficient of the ith edge at the current analysis time t; represents the initial edge weight of the ith edge; norm represents the normalization function; similarly, the updated edge weight of the edge with a monitoring blind spot is calculated; Get the updated time interval based on the current analysis time, and the corresponding calculation formula is: Calculate the updated time interval without monitoring blind spot edge: ； in, represents the update time interval after the i-th edge is updated; represents the congestion coefficient of the ith edge at the current analysis time t; Represents the initial update time interval; norm represents the normalization function; similarly, the update time interval after the update of the edge with monitoring blind area is calculated.

10. A smart park vehicle navigation planning system based on digital twins, characterized in that: The system stores program data. When the program data is executed, it combines with camera equipment, sensors, data acquisition modules, digital twin modules, vehicle terminals, navigation decision modules, road information databases and road detection modules to implement a smart park vehicle navigation planning method based on digital twins as described in any one of claims 1 to 9.

Images