A design method for optimizing the layout of urban scenic spot parking facilities

Abstract

The application discloses a design method for optimizing city scenic spot parking facility layout, relates to the field of parking layout design, and comprises the following steps: acquiring real-time surrounding event data of a scenic spot, analyzing event types and geographic coordinates, generating a structured data set, wherein the structured data set comprises fence boundary coordinates and event intensity factors, generating a gradient electronic fence on a GIS platform based on the structured data set, inputting the fence boundary coordinates and the event intensity factors into a pre-constructed parking demand prediction model to perform parameter reorganization, outputting a parking resource demand prediction atlas, generating a response strategy according to the electronic fence range and the parking resource demand prediction atlas, forming a scheduling instruction set, verifying the identity of a reservation vehicle through a V2X vehicle-road cooperation mechanism, matching vehicle physical parameters in combination with available parking space resources in the scheduling instruction set, and generating a three-dimensional navigation package; and the application realizes accurate adaptation of resource supply through a spatial topology matching algorithm, and shortens an emergency response cycle.

Description

Technical Field

[0001] This invention relates to the field of parking layout design, and in particular to a design method for optimizing the layout of parking facilities in urban scenic spots. Background Technology

[0002] With the deepening of smart city construction, optimizing the layout of parking facilities at popular urban attractions has become a research hotspot in the field of intelligent transportation. Current mainstream technologies mainly focus on three directions: First, predictive models based on historical traffic flow data have been applied in scenic areas, enabling pre-allocation of parking spaces by analyzing daily fluctuations in visitor flow; second, vehicle-road cooperative technology, leveraging the DSRC / C-V2X protocol, provides route guidance for reserved vehicles, shortening the average parking space search time; and third, spatial perception, through geomagnetic / video detection of parking space status, significantly improves vehicle recognition rates.

[0003] However, the existing parking facility layout still has room for improvement. First, the lack of an emergency response mechanism leads to insufficient accuracy. Traditional prediction models rely on periodic historical data and cannot account for sudden variables such as traffic accidents and medical emergencies, resulting in longer response times and exacerbating congestion on surrounding roads. In addition, insufficient vehicle-space matching accuracy causes resource mismatch. Existing V2X navigation mostly focuses on parking lot entrance guidance and lacks lane-level route planning and parking space-level positioning capabilities, resulting in higher levels of ineffective traffic flow in the last-mile parking search. Summary of the Invention

[0004] In view of the aforementioned existing problems, the present invention is proposed.

[0005] Therefore, this invention provides a design method for optimizing the layout of parking facilities in urban scenic spots to solve the problems of insufficient accuracy caused by the lack of emergency response mechanisms and resource mismatch caused by insufficient vehicle-space matching accuracy.

[0006] To solve the above-mentioned technical problems, the present invention provides the following technical solution:

[0007] In a first aspect, the present invention provides a design method for optimizing the layout of parking facilities in urban scenic spots, which includes: acquiring real-time data on emergencies around the scenic area, analyzing the types and geographical coordinates of emergencies, and generating a structured dataset; the structured dataset includes fence boundary coordinates and event intensity factors;

[0008] Based on the structured dataset, a gradient electronic fence is generated on the GIS platform. At the same time, the fence boundary coordinates and event intensity factors are input into the pre-built parking demand prediction model for parameter recombination, and a parking resource demand prediction map is output.

[0009] Based on the electronic fence range and parking resource demand forecast map, a response strategy is generated to form a scheduling instruction set;

[0010] The identity of the reserved vehicle is verified through the V2X vehicle-road cooperative mechanism, and the vehicle's physical parameters are matched with the available parking space resources in the dispatch instruction set to generate a three-dimensional navigation package.

[0011] The system performs light strip guidance based on the 3D navigation package along the travel path, and simultaneously integrates roadside anchor points and vehicle navigation data for position correction, outputting parking coordinates.

[0012] The system locates the target vehicle based on the parking coordinates, collects the tire force waveform, identifies vehicle characteristics and calculates their proportions, triggers spatial reconstruction, and generates a real-time spatial configuration map that is fed back to the decision-making unit.

[0013] As a preferred embodiment of the design method for optimizing the layout of parking facilities in urban scenic spots according to the present invention, the generation of the structured dataset specifically includes the following steps:

[0014] The data on emergencies are organized into a raw emergency dataset. Event descriptions and category information are extracted from the raw emergency dataset. A pre-trained Chinese text classifier is used to label the event types. The classified emergency dataset is output, and the event intensity factor is determined.

[0015] Using GIS tools, fence boundary coordinates were generated based on the event intensity factor, and then integrated with the event intensity factor and the classified emergency event dataset into a structured dataset.

[0016] As a preferred embodiment of the design method for optimizing the layout of parking facilities in urban scenic spots according to the present invention, the step of outputting a parking resource demand prediction map specifically includes the following steps:

[0017] Event types and geographic coordinates are extracted from structured datasets. GIS tools are used to generate gradient electronic fences with predefined ranges based on event types, and the data is then compiled into an electronic fence dataset.

[0018] The fence boundary coordinates and event intensity factors are extracted from the electronic fence dataset to form standardized input data. This data is then input into a parking demand prediction model based on a multilayer perceptron neural network, triggering dynamic parameter reorganization. The feature weights are adjusted using a weighted average formula, and a parking resource demand prediction map is output.

[0019] As a preferred embodiment of the design method for optimizing the layout of parking facilities in urban scenic spots according to the present invention, the following steps are included: A response strategy is generated based on the electronic fence range and a parking resource demand prediction map to form a scheduling instruction set.

[0020] Event types and fence boundary coordinates are extracted from the electronic fence dataset. A three-level response strategy is formulated by combining the parking resource demand prediction map, and a scheduling scheme dataset is generated.

[0021] The parking lot opening and closing status, diversion paths, and shared resource locations are extracted from the scheduling scheme dataset, organized into a standardized scheduling instruction set stored in tabular form, and transmitted to the V2X vehicle-road cooperative unit.

[0022] As a preferred embodiment of the design method for optimizing the layout of parking facilities in urban scenic spots according to the present invention, the generation of the three-dimensional navigation package specifically includes the following steps:

[0023] Extract vehicle IDs from the dispatch instruction set, verify vehicle identity using encrypted digital signatures via roadside communication equipment, and output a verification dataset confirming vehicle legitimacy.

[0024] The vehicle information database is queried using the vehicle IDs in the validation dataset to obtain the vehicle's physical parameters. The available parking space resources in the scheduling instruction set are then matched using a geometric constraint filtering method to obtain the coordinates of the matched parking space. A path planning algorithm is then used to calculate the optimal passage path and generate a 3D navigation package, taking the vehicle's current position, the coordinates of the matched parking space, and the diversion path in the scheduling instruction set as inputs.

[0025] As a preferred embodiment of the design method for optimizing the layout of parking facilities in urban scenic spots according to the present invention, the output of parking coordinates specifically includes the following steps:

[0026] The system extracts the travel path and parking space coordinates from the 3D navigation package, deploys programmable road studs through optical guidance equipment to form a dynamic light strip, sets blue light to be constantly on for straight sections, sets blue light to flash at turning nodes, and sets red light arrows for parking spaces, and outputs a programmable road stud instruction set to guide vehicles to travel along the dedicated travel path.

[0027] Based on the travel path in the 3D navigation package, the Kalman filter algorithm is used to fuse roadside anchor point signals and vehicle inertial navigation data to correct the vehicle position and output a positioning correction dataset.

[0028] The final parking coordinates are calculated using the positioning correction dataset and parking space coordinates, and then transmitted to the geomagnetic piezoelectric monitoring network via the V2X communication interface to output the parking coordinates.

[0029] As a preferred embodiment of the design method for optimizing the layout of parking facilities in urban scenic spots as described in this invention, the following steps are included: locating target vehicles based on parking coordinates and collecting tire force waveforms, identifying vehicle type characteristics and calculating their proportions, triggering spatial reconstruction, and generating a real-time spatial configuration map that is fed back to the decision-making unit:

[0030] Based on the parking coordinates, the geomagnetic piezoelectric sensor array is activated to collect the tire force waveform, and after extracting the waveform features, it is compared with the preset vehicle model database to identify the vehicle type and output the vehicle type dataset.

[0031] The vehicle type dataset is used to calculate the proportion of different vehicle types. Spatial reconstruction is triggered based on the predefined thresholds for the proportion of buses and new energy vehicles. The parking lot layout is updated using GIS tools to generate a real-time spatial configuration map, which is then transmitted to the decision-making unit via the MQTT protocol.

[0032] As a preferred embodiment of the design method for optimizing the layout of parking facilities in urban scenic spots according to the present invention, obtaining the pre-trained Chinese text classifier specifically includes the following steps:

[0033] We collected a large-scale Chinese text dataset related to emergencies, manually labeled the event types, and divided it into training and test sets.

[0034] Download the pre-trained Chinese BERT model, add a classification layer, and output the event type probability using the Softmax function;

[0035] Set the hyperparameters of the Chinese BERT model, train the classifier of the pre-trained Chinese BERT model using the cross-entropy loss function, and evaluate the performance on the test set to obtain the pre-trained Chinese text classifier.

[0036] In a second aspect, the present invention provides a computer device including a memory and a processor, wherein the memory stores a computer program, wherein when the computer program is executed by the processor, it implements any step of the design method for optimizing the layout of parking facilities at urban attractions as described in the first aspect of the present invention.

[0037] Thirdly, the present invention provides a computer-readable storage medium having a computer program stored thereon, wherein: when the computer program is executed by a processor, it implements any step of the design method for optimizing the layout of parking facilities at urban attractions as described in the first aspect of the present invention.

[0038] The beneficial effects of this invention are as follows: Addressing the deficiency of historical data distortion caused by sudden events, this invention quantifies the degree of external impact through an event intensity factor, enabling dynamic adaptive adjustment of model parameters and establishing a closed-loop feedback mechanism between the impact of sudden events and the model response. This significantly improves prediction accuracy in highly volatile scenarios and avoids decision-making failure of static models in emergency situations. Furthermore, it generates a scheduling instruction set based on the electronic fence range and parking resource demand prediction map, overcoming the problem of traditional scheduling strategies being disconnected from the scope of event impact. Through a spatial topology matching algorithm, it achieves precise resource supply matching, shortens the emergency response cycle, and reduces the risk of secondary congestion caused by resource mismatch. Attached Figure Description

[0039] To more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings used in the following description of the embodiments will be briefly introduced. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0040] Figure 1 A flowchart illustrating the design methodology for optimizing the layout of parking facilities at urban tourist attractions.

[0041] Figure 2 This is a flowchart for parking resource prediction and scheduling decision-making.

[0042] Figure 3 A flowchart for vehicle guidance and positioning correction.

[0043] Figure 4 A flowchart for spatial self-organization and reconstruction. Detailed Implementation

[0044] To make the above-mentioned objects, features and advantages of the present invention more apparent and understandable, the specific embodiments of the present invention will be described in detail below with reference to the accompanying drawings.

[0045] Many specific details are set forth in the following description in order to provide a full understanding of the invention. However, the invention may also be practiced in other ways different from those described herein, and those skilled in the art can make similar extensions without departing from the spirit of the invention. Therefore, the invention is not limited to the specific embodiments disclosed below.

[0046] Secondly, the term "one embodiment" or "embodiment" as used herein refers to a specific feature, structure, or characteristic that may be included in at least one implementation of the present invention. The phrase "in one embodiment" appearing in different places in this specification does not necessarily refer to the same embodiment, nor is it a single or selective embodiment that is mutually exclusive with other embodiments.

[0047] Reference Figures 1-4 As one embodiment of the present invention, this embodiment provides a design method for optimizing the layout of parking facilities in urban scenic spots, including the following steps:

[0048] S1: Real-time acquisition of emergency data around the scenic area, parsing of emergency types and geographical coordinates, generating a structured dataset, which includes fence boundary coordinates and event intensity factors;

[0049] Specifically, the steps include the following:

[0050] S1.1: Real-time collection of emergency event data within a designated area surrounding the scenic spot is achieved through publicly available emergency information APIs and social media platforms. The emergency information API retrieves official records of incidents such as police reports, emergency medical services, and fire alarms, including the time, location, and type of the event. Social media platforms utilize web scraping technology to collect publicly available posts with geographic tags, filtering for content containing emergency event keywords, such as traffic accidents, fires, and emergency medical services. The collected data includes event descriptions, geographic locations (latitude, longitude, or address), posting times, and event categories, outputting a raw emergency event dataset containing event descriptions, geographic locations, times, and categories.

[0051] From the original emergency data set, event descriptions and category information are extracted to parse the emergency types. Natural language processing techniques are used to extract keywords and classify event descriptions to identify event types, such as police calls, emergency medical services, fire alarms, and traffic accidents. For structured data from emergency information APIs, such as police calls, emergency medical services, fire alarms, and timestamps, the event type field is directly read. For unstructured data from social media platforms, such as post content, a pre-trained text classification model is used for event type labeling. The output is a categorized emergency dataset containing event type, geographic location, and time.

[0052] Further explanation of the pre-trained text classification model: (1) Collect a large-scale Chinese text dataset for training the BERT-based Chinese text classifier, containing text data related to emergencies. The dataset sources include publicly available social media posts, news reports, and government data records, with text content covering various types of emergencies. Each text data is manually labeled with an event type label to form a labeled training dataset. The training dataset is divided into a training set and a test set. Each data contains text content (e.g., "a fire occurred near West Lake") and the corresponding event type label, resulting in a training dataset containing text content and event type labels. (2) Download the pre-trained Chinese BERT model from a public model library. Add a classification layer on top of the pre-trained Chinese BERT model to adapt to the event type labeling task. Extract the hidden representation of the [CLS] label output from the last layer of the pre-trained Chinese BERT model as the semantic representation of the entire input text. Input the [CLS] hidden representation into a fully connected layer and map the dimension to the number of event type labels, for example, 4 categories: police, emergency, fire, and traffic accident. The fully connected layer is followed by a Softmax activation function to output the probability distribution of each event type. The output is a complete BERT-based Chinese text classifier containing a pre-trained Chinese BERT model and a classification layer. (3) Extract text content from the training dataset and preprocess it to adapt to the input format of the BERT-based Chinese text classifier. (4) Configure hyperparameters for the training process of the BERT-based Chinese text classifier, specifically: set the learning rate to 2e-5, which is selected according to the BERT fine-tuning standard guide to avoid gradient explosion; test the range of 1e-5 to 5e-5 on the validation set through grid search, select the lowest mean square error, use AdamW as the optimizer, and set the batch size to 16 to balance computational resources and gradient stability, which is determined by the validation range of 8-32. The number of training rounds is 3 to 5, and an early stopping mechanism is used, stopping when the validation set loss does not decrease for 3 consecutive rounds. Use the cross-entropy loss function as the loss function for the classification task. (5) Using the preprocessed training dataset and training hyperparameters, train a BERT-based Chinese text classifier to obtain the trained BERT-based Chinese text classifier. Use the test set to evaluate the performance of the trained BERT-based Chinese text classifier. When the accuracy meets the requirements, obtain the BERT-based Chinese text classifier with satisfactory performance, which can be used for event type labeling.

[0053] S1.2: For the categorized emergency event dataset, extract geographic location information and parse it into standardized geographic coordinates. Specifically: For event data provided by government data lines, if latitude and longitude are included, extract them directly; if only address descriptions are provided, use geocoding technology (e.g., Gaode Map API) to convert the address into latitude and longitude coordinates. For geographic tag data from social media platforms, if it is in latitude and longitude format, extract it directly; if the location is ambiguous, combine the boundary coordinates of the designated area around the scenic spot and estimate the approximate latitude and longitude using geographic information system tools. The output is a coordinate-annotated emergency event dataset containing event type, geographic coordinates, and time.

[0054] From the coordinate-annotated dataset of emergencies, event intensity factors are determined based on event type and geographical location. These factors are predefined according to the impact range and urgency of the event type: Major events are assigned a high intensity factor (e.g., fires are major events, assigned a high intensity factor between 0.8 and 1.0; high-impact events are assigned high values, with the range set by statistically averaging historical event data (e.g., fires have an average impact of 0.85); emergency response events are assigned a medium intensity factor (medium-impact events, verified based on emergency response logs, average 0.65), for example, 0.5-0.8; local events are assigned a low intensity factor (e.g., minor traffic accidents are local events, assigned a low intensity factor between 0.2 and 0.5; low-impact events, the range determined through simulation testing, average 0.35). Combining geographical coordinates, the distance between the event and the core area of ​​the scenic area is analyzed. The closer the event, the higher the intensity factor is adjusted. For example, assuming West Lake Scenic Area is the core area, the geographical coordinates of the center point of the core area are (120.15, 30.25). The geographical coordinates of the fire emergency are (120.14, 30.24), with an initial event intensity factor of 0.8. Using GIS tools, the Euclidean distance between the emergency's geographical coordinates and the center point of the scenic area's core area is calculated based on latitude and longitude, which is 1.4 kilometers. A predefined distance threshold of 2 kilometers is used based on urban traffic characteristics (based on an average urban response time of 5-10 minutes; 2km corresponds to a walking / driving range, obtained through historical data statistics). If the distance between the emergency and the center point of the scenic area's core area is less than 2 kilometers, the event intensity factor is increased by 0.1. The increase is determined by analyzing the sensitivity of the event's impact range to parking demand prediction; for example, an increase of 0.1 reduces the prediction error by 5%. Since 1.4 kilometers < 2 kilometers, the initial event intensity factor of 0.8 is increased by 0.1, resulting in an adjusted event intensity factor of 0.9. The output is an intensity-labeled emergency dataset containing event type, geographical coordinates, time, and event intensity factor.

[0055] From the intensity-labeled emergency event dataset, geographic coordinates and event intensity factors are extracted to generate fence boundary coordinates. Specifically, using GIS tools, based on event type and event intensity factors, gradient-based boundary coordinates of the impact area are generated. Major events generate large-scale fences, such as a radius of 500 meters (determined through GIS simulation verification); emergency response events generate medium-scale fences, such as a radius of 200 meters (determined through emergency response time); and local events generate small-scale fences, such as a radius of 50 meters (determined through traffic accident impact range testing). The fence boundary coordinates are represented in the form of a polygon point set. Calculations are based on geographic coordinates and employ a buffer generation algorithm in GIS. The calculation process is as follows: Event type, geographic coordinates, time, and event intensity factor are extracted from the intensity-labeled emergency event dataset to obtain an extracted dataset containing these information. Geographic Information System (GIS) tools are used to load the geographic coordinates from this extracted dataset. The buffer analysis method is initialized, resulting in a configured buffer analysis environment. Based on the event type in the extracted dataset, a predefined fence radius is established. Combined with the configured buffer analysis environment, a buffer parameter set containing geographic coordinates and the fence radius is output. Using GIS tools, the buffer generation algorithm is executed based on this parameter set to calculate a circular region centered on the geographic coordinates and with a distance equal to the fence radius. A circular buffer geometry is output. This circular buffer geometry is then converted into a polygon point set representing the fence boundary coordinates, resulting in a polygon dataset containing the fence boundary coordinates. This dataset is then organized into a fence-labeled emergency event dataset.

[0056] From the dataset of incidents marked with fences, we organize the event type, geographic coordinates, time, event intensity factor, and fence boundary coordinates to generate a structured dataset. The structured dataset is stored in tabular form, with each row recording one incident, including the following fields: event type, geographic coordinates, time, event intensity factor, and fence boundary coordinates.

[0057] To further explain, data on emergencies around the scenic area is collected in real time through government data lines and social media platforms. This data is then combined with natural language processing technology and a pre-trained BERT-based Chinese text classifier to quickly parse event types. Furthermore, geocoding technology and Geographic Information System (GIS) tools are used to convert geographical locations into standardized geographic coordinates, ensuring the timeliness and accuracy of event information. This real-time and precise capability for event data acquisition and analysis provides a reliable data foundation for subsequent dynamic optimization of parking facility layout.

[0058] S2: Based on the structured dataset, a gradient electronic fence is generated on the GIS platform. At the same time, the fence boundary coordinates and event intensity factors are input into the pre-built parking demand prediction model for parameter recombination, and the parking resource demand prediction map is output.

[0059] Specifically, the steps include the following:

[0060] S2.1: Extract event types and geographic coordinates from the structured dataset, and use Geographic Information System (GIS) tools to generate gradient-type impact area electronic fences. Specifically: Predefine fence ranges according to event types: major events generate large-scale control zones with a fence radius of 500 meters (verified through historical event simulations to cover high-risk areas); emergency events generate medium-scale fences with a fence radius of 200 meters (determined based on emergency response needs); local events generate small-scale response zones with a fence radius of 50 meters (determined based on the impact range of traffic accidents). Based on the geographic coordinates of the emergencies, a buffer analysis method is used to calculate the fence boundary coordinates, forming an electronic fence represented by a polygon point set. The output is an electronic fence dataset containing event type, geographic coordinates, time, event intensity factor, and updated fence boundary coordinates. Extract fence boundary coordinates and event intensity factors from the electronic fence dataset. The fence boundary coordinates are polygon point sets representing the boundaries of the gradient-type impact area; the event intensity factor is a numerical value reflecting the degree of impact of the emergency. Standardize the input data format of the fence boundary coordinates and event intensity factors, and organize them into an input dataset.

[0061] S2.2: The process of pre-constructing the parking demand prediction model is as follows: (1) Collect a large-scale dataset for training the parking demand prediction model, which includes historical data related to parking demand around the scenic area. The sources of the large-scale dataset include traffic flow records, tourist flow statistics, time information and emergency records of the scenic area parking lot. Each large-scale dataset is labeled with the parking demand amount to form a labeled training dataset. The training dataset is divided into a training set and a validation set. Each data point contains feature fields and target fields to obtain a training dataset containing feature fields and parking demand amount labels. For the training dataset, feature fields are extracted, and feature engineering and preprocessing are performed to obtain a preprocessed training dataset. The preprocessed training dataset contains standardized and encoded feature fields and parking demand amount labels. (2) Select the parking demand prediction model architecture: Select a neural network and construct a multilayer perceptron, which includes an input layer, several hidden layers and an output layer. The neural network uses mean squared error as the loss function. Thus, the initialized parking demand prediction model is obtained. (3) Set the training hyperparameters of the initial parking demand prediction model: Set the learning rate to 1e-3, with a learning rate range of 1e-4 to 1e-2. Select the lowest mean squared error value through cross-validation, for example, select 1e-3 after testing, and the validation set error will drop to 0.05. The optimizer is Adam, and the batch size is between 16 and 64. To balance efficiency, select 32. The number of training rounds is 50-100 (use the early stopping mechanism to avoid overfitting, and stop when the validation set loss does not decrease for 10 consecutive rounds). (4) Train the parking demand prediction model using the preprocessed training dataset and hyperparameters. Input the preprocessed training dataset into the initial parking demand prediction model in batches and calculate the predicted parking demand value output by forward propagation. Measure the loss between the predicted value and the actual parking demand label using the mean squared error loss function. Use the Adam optimizer for backpropagation to update the neural network parameters. Due to the early stopping mechanism, stop training when the validation set loss does not decrease for 10 consecutive rounds and output the trained parking demand prediction model.

[0062] The fence boundary coordinates and event intensity factor are extracted from the input dataset and used as input features for the trained parking demand prediction model. The trained parking demand prediction model is based on a multilayer perceptron neural network, initially trained using historical traffic flow data, time features, and visitor flow characteristics; feature weights are determined during the training phase. After receiving the fence boundary coordinates and event intensity factor, a geographic information system (GIS) tool is used to map the fence boundary coordinates to the spatial grid input of the parking demand prediction model, ensuring that geographic information of the affected area is incorporated into the prediction process. Dynamic parameter reorganization is triggered, and the input feature weights are adjusted using a weighted average formula:

[0063] ;

[0064] in, This represents the new, updated feature weights. This represents the smoothing factor. Based on the weight update mechanism, it is tested in the range of 0.5-0.9 through cross-validation, and the value with the lowest mean square error is selected. For example... Reduce validation set error by 10% and pass validation set cyclic testing. Values, record errors and select the best. This represents the original feature weights determined during the training phase, which are extracted from the input layer weight matrix after training. For example, [0.6, 0.2, 0.1, 0.05, 0.05] corresponds to features such as historical traffic flow. This represents the adjustment factor.

[0065] The adjustment factor is calculated based on the event intensity factor and the fence range, and the expression is:

[0066] ;

[0067] in, Indicates the adjustment factor. The weighting coefficients for the event intensity factor, ranging from 0.6 to 0.8, are selected through sensitivity analysis to minimize the prediction error. For example... To reduce the error by 8%, test using simulated event data. The value is determined by recording the accuracy of the prediction. The event intensity factor is extracted from an intensity-labeled emergency event dataset, ranging from 0.2 to 1.0, and adjusted for distance, such as in the case of a fire. , Indicates the area of ​​the fence. Indicates the maximum fence area;

[0068] The weight adjustments are determined based on the magnitude of the event intensity factor and the fence range. The weight of historical traffic flow data is reduced, for example, from 0.6 to 0.3, to decrease reliance on conventional parking patterns. The weight of the event intensity factor is increased, for example, from 0.2 to 0.5, to highlight the impact of sudden events on parking demand. The updated weights are redistributed to the input layer of the neural network. Only the input layer weights are updated, while the hidden and output layer parameters are retained, generating a parameter-recombined parking demand prediction model. This recombined model uses enhanced input features—historical traffic flow data, temporal features, visitor flow, spatial grid features, and the event intensity factor—to calculate predicted values ​​through forward propagation, generating a parking resource demand prediction map. This map is represented by a 100m × 100m grid, with a testing range of 50m-200m. The 100m grid balances accuracy and computational efficiency. Each grid cell is labeled with its geographical range and parking demand. The parking resource demand prediction map is visualized as a heatmap using GIS tools, supporting the scheduling decision module to dynamically optimize parking facility layout and prioritize the impact of sudden events.

[0069] To further explain, a parking demand prediction model with reconfigured parameters, combined with fence boundary coordinates, event intensity factors, and other input features such as current time and visitor flow, is used to predict parking demand around the scenic area. The prediction results are represented in a two-dimensional grid, covering a defined area around the scenic area, with each grid cell labeled with parking demand. Using Geographic Information System (GIS) tools, the parking demand around the scenic area is mapped to geographic coordinates, forming a visualized parking demand prediction map. This map includes the geographic range and parking demand of each grid cell, reflecting the impact of unforeseen events on parking demand in various areas.

[0070] To further explain, by inputting the fence boundary coordinates and event intensity factor into the trained parking demand prediction model, dynamic parameter reorganization is triggered. This reduces the weight of historical traffic flow data and increases the impact coefficient of sudden events, making the reorganized parking demand prediction model more focused on the impact of current sudden events. This dynamic adjustment mechanism can quickly adapt to fluctuations in parking demand caused by sudden events, ensuring that the prediction results closely match the real-time scenario and avoiding prediction bias caused by relying solely on historical data. This improves the flexibility and accuracy of parking facility layout.

[0071] S3: Generate response strategies based on the electronic fence range and parking resource demand prediction map, forming a set of scheduling instructions;

[0072] Specifically, the steps include the following:

[0073] From the electronic fence dataset, extract the event type and fence boundary coordinates, and combine them with the parking resource demand prediction map dataset to formulate a three-level response strategy, specifically: (1) For major events (such as fire alarms, with a fence radius of 500 meters), identify the core parking lot covered by the fence boundary coordinates, mark it as closed, query the location of the backup parking lot, and generate diversion paths. (2) For emergency incidents, identify the roads within the fence boundary coordinates, open emergency lanes, and generate green wave traffic zones. The green wave traffic zones are obtained by optimizing signal timing through a traffic signal coordination algorithm. Specifically, the road IDs of the green wave traffic zones and the fence boundary coordinates of the emergency incidents are extracted. The green wave traffic zone dataset containing the road IDs and fence boundary coordinates is loaded using the traffic signal coordination algorithm. The locations and current timing parameters of the traffic lights on the roads are obtained. Based on the traffic light dataset containing the locations and current timing parameters of the traffic lights, the target vehicle speed of the green wave traffic zone and the distance between adjacent traffic lights are calculated. The process is as follows: extract the geographic coordinates of adjacent traffic lights from the traffic light dataset, use geographic information system tools to calculate the actual road distance between two traffic lights based on latitude and longitude, and obtain the distance between adjacent traffic lights; determine the traffic light cycle and green light time window, and calculate the target vehicle speed:

[0074] ;

[0075] in, Indicates the target vehicle speed. Indicates the actual road distance between adjacent traffic lights. This indicates that the green light has shifted.

[0076] The actual distance between traffic lights and the target vehicle speed are integrated into a green wave parameter set. Based on the green wave parameter set, the start time and duration of the green light of the traffic lights are adjusted to form a continuous green light passage zone. (3) For local events, the shared parking space resources within the fence boundary coordinates are identified, the shared parking spaces are released, and temporary parking areas are added (based on the division of available road space) to obtain a scheduling scheme dataset containing the parking lot opening and closing status, route guidance scheme and three-level response strategy of shared resource location.

[0077] Based on the scheduling scheme dataset, parking lot opening / closing status, route guidance schemes, and shared resource locations are extracted and organized into a standardized scheduling instruction set. The scheduling instruction set is stored in tabular form, with each row recording one instruction, including fields such as: parking lot opening / closing status, route guidance scheme, and shared resource location. For example, the parking lot opening / closing status is "Core Parking Lot A Closed," "Backup Parking Lot B Opened"; the route guidance scheme includes the latitude and longitude of the starting and ending points of the diversion route, the road ID of the green wave traffic zone, and the signal timing parameters; the shared resource location includes, for example, the latitude and longitude coordinates of shared parking spaces and the boundary coordinates of temporary parking areas. The scheduling instruction set is encoded using a data serialization method for easy transmission. The output is a scheduling instruction set containing parking lot opening / closing status, route guidance schemes, and shared resource locations. This scheduling instruction set allocates parking resources to reserved vehicles and is transmitted to the V2X vehicle-to-infrastructure (V2X) unit.

[0078] To further explain, by combining electronic fence datasets with parking resource demand prediction map datasets, a three-tiered response strategy is formulated, generating differentiated scheduling plans for major events, emergency response events, and localized events. This tiered response mechanism can accurately match the geographical impact range and severity of emergencies, optimize the dynamic management of parking facilities around scenic areas, and improve the targeting and effectiveness of parking resource allocation.

[0079] S4: Verify the identity of the reserved vehicle through the V2X vehicle-road cooperative mechanism, match the vehicle's physical parameters with the available parking space resources in the dispatch instruction set, and generate a three-dimensional navigation package;

[0080] Specifically, the steps include the following:

[0081] From the output scheduling instruction dataset, relevant information about the reserved vehicle, such as the vehicle ID, is obtained. Roadside communication equipment is used to communicate with the vehicle's onboard terminal. The roadside communication equipment sends an authentication request and receives identity information returned by the onboard terminal, such as the vehicle's unique identifier and reservation number. The legitimacy of the onboard terminal's identity information is verified using an encrypted verification method. The verification process is as follows: The identity information dataset containing the vehicle's unique identifier, reservation number, and digital signature is received from the onboard terminal. A public key infrastructure (PKI) query is used to retrieve the public key corresponding to the vehicle's unique identifier to obtain the digital signature. The digital signature is decrypted using the PKI's digital signature verification algorithm to generate the original message digest. A decrypted dataset containing the original message digest is obtained. The vehicle's unique identifier and reservation number are extracted from the identity information dataset, and their hash values ​​are calculated to generate a computed message digest. Combined with the decrypted dataset, a comparison dataset containing the original message digest and the computed message digest is obtained. The two digests in the comparison dataset are compared; if they match, the legitimacy of the onboard terminal's identity information is confirmed. A verification dataset containing the legitimacy result of the onboard terminal's identity information is output.

[0082] The system extracts vehicle IDs from the validation dataset and queries the pre-stored vehicle information database to obtain the physical parameters of the reserved vehicle, such as length, width, height, and turning radius. Specifically, it extracts the location of shared resources and the parking lot's opening / closing status from the dispatch instruction set dataset to filter available parking spaces. A geometric constraint-based filtering method compares the vehicle's physical parameters with the spatial parameters of available parking spaces, selecting spaces that meet the length, width, height, and turning radius requirements. The output is a parking space matching dataset containing the coordinates of the matched reserved parking spaces and the vehicle IDs. The comparison process involves: extracting the physical parameters of the reserved vehicle from the vehicle information database, including length, width, height, and turning radius; and extracting the spatial parameters of available parking spaces from the dispatch instruction set dataset, including space size and aisle width. The geometric constraint-based filtering method first compares the vehicle's length, width, and height with the parking space size, ensuring the parking space size is greater than or equal to the vehicle size. Then, it compares the vehicle's turning radius with the aisle width, confirming the aisle width is sufficient to support the vehicle's turning needs. By filtering each space, parking spaces that do not meet any of the conditions are excluded, and available parking spaces that meet all geometric constraints are selected, prioritizing those closest to the current location of the reserved vehicle. Finally, the coordinates of the selected parking space are output to ensure that the vehicle can be parked safely and smoothly.

[0083] Using a route planning algorithm, starting from the current location of the reserved vehicle (obtained in real-time via the vehicle terminal) and ending at the coordinates of the reserved parking space, the optimal travel route is calculated by combining the diversion paths and green wave traffic zones in the route guidance scheme. Considering road length, traffic conditions, and signal timing, the optimal route is generated as a three-dimensional coordinate sequence, and a navigation dataset is output. Specifically, the process is as follows: The current location coordinates of the reserved vehicle are obtained from the vehicle terminal via the V2X communication protocol; the coordinates of the reserved parking space are extracted from the parking space matching dataset; the route guidance scheme is extracted from the dispatch instruction set dataset, including the latitude and longitude of the starting and ending points of the diversion paths, as well as the road IDs and signal timing parameters of the green wave traffic zones; road network data from the geographic information system is loaded, including road length, real-time traffic conditions, and signal timing information; Algorithm A is used, with the road network as a graph structure, nodes as intersection coordinates, and edge weights comprehensively considering road length, traffic conditions (increased weight for congested sections), and signal timing (decreased weight for green wave traffic zone sections). The optimal path from the current location to the coordinates of the reserved parking space is calculated, generating a 3D coordinate sequence containing latitude, longitude, and altitude information to meet the needs of 3D navigation, such as avoiding the height restrictions of overpasses. The optimal route and reserved parking space coordinates are then compiled into a navigation dataset containing both the reserved parking space coordinates and the dedicated route, and organized into a 3D navigation package. The 3D navigation package is stored in a structured data format, containing the fields: reserved parking space coordinates and dedicated route. Data serialization methods are used to encode the 3D navigation package to ensure data integrity and transmission efficiency.

[0084] The 3D navigation package is transmitted in real time to the vehicle positioning compensation unit and the optical guidance device via the V2X vehicle-to-everything (V2X) communication interface. The vehicle positioning compensation unit receives the 3D navigation package to correct the vehicle's position and optimize navigation accuracy; the optical guidance device receives the 3D navigation package to deploy guidance signals.

[0085] To further explain, the system communicates with the vehicle-mounted terminal via roadside communication equipment, receiving identity information including the vehicle's unique identifier, reservation number, and digital signature. Using a public key infrastructure-based digital signature verification algorithm, the legitimacy of the vehicle-mounted terminal's identity information is confirmed by comparing the original message digest and the calculated message digest. This encrypted verification method effectively prevents unauthorized vehicles from impersonating the reservation holder, ensuring that only legitimately reserved vehicles can obtain parking resource allocation and navigation services. This improves the reliability and security of parking management and maintains the orderly layout of parking facilities within the scenic area.

[0086] S5: Performs light strip guidance based on the 3D navigation package along the travel path, and simultaneously integrates roadside anchor points and vehicle navigation data for position correction, outputting parking coordinates;

[0087] Specifically, the steps include the following:

[0088] From the 3D navigation package, the dedicated travel path and reserved parking space coordinates are extracted. Programmable road studs are deployed along the travel path of the reserved vehicle using optical guidance equipment. Based on the coordinate sequence of the dedicated travel path, the optical signal modes of the programmable road studs are configured: a constant blue light is set for straight sections, a flashing blue light is set for turning nodes (such as intersections or curves), and red arrows are placed near the reserved parking space coordinates to indicate the parking space location. By coordinating the illumination time and position of the programmable road studs, a continuous dynamic light strip is formed to guide the vehicle along the dedicated path to the reserved parking space. The output is a programmable road stud instruction set containing the dynamic light strip configuration.

[0089] From the 3D navigation package, a dedicated travel path is extracted. The vehicle positioning compensation unit receives roadside anchor point signals via a V2X communication interface, and the Roadside Unit (RSU) provides the positioning reference signal. Simultaneously, onboard inertial navigation data, including the vehicle's real-time position and attitude, is acquired through onboard sensors. However, due to potential cumulative errors caused by onboard sensor drift, a Kalman filter algorithm is used to fuse the roadside anchor point signals and onboard inertial navigation data. Vehicle position deviation is corrected through state estimation and prediction. Specifically, the onboard inertial navigation data is used as the prediction input to estimate the vehicle's position at the next moment; the roadside anchor point signals are used as the observation input to update the estimation results, generating high-precision real-time vehicle position coordinates. The fusion process considers signal noise and sensor errors to optimize the stability of the position estimation. Finally, the real-time vehicle position coordinates are compiled into a positioning correction dataset containing the real-time vehicle position coordinates and stored in a structured format.

[0090] Real-time vehicle position coordinates are extracted from the positioning correction dataset and combined with the reserved parking space coordinates from the 3D navigation package. Geometric calculation methods are then used to determine the final parking coordinates of the vehicle as it approaches the reserved parking space. These final parking coordinates are then transmitted to the geomagnetic piezoelectric monitoring network via a V2X communication interface, resulting in a parking coordinate dataset containing the final parking coordinates.

[0091] To further explain, by extracting dedicated travel routes and reserved parking space data from the 3D navigation package, programmable road studs are deployed along the travel path of reserved vehicles using optical guidance devices. These studs are configured with constantly lit blue lights to indicate straight sections, flashing blue lights to warn of turning points (such as intersections or curves), and red arrows to indicate parking space locations, forming a continuous dynamic light strip. The visual guidance effect of this dynamic light strip is intuitive and clear, guiding vehicles precisely along the dedicated travel route to their reserved parking spaces. This significantly reduces driving deviations caused by unclear routes or misjudgments, minimizes ineffective vehicle movement within the scenic area's parking lot, and improves parking efficiency and the visitor experience.

[0092] S6: Locate the target vehicle based on the parking coordinates and collect the tire force waveform, identify vehicle characteristics and calculate the proportion, trigger spatial reconstruction, generate a real-time spatial configuration map and feed it back to the decision unit.

[0093] Specifically, the steps include the following:

[0094] From the parking coordinate dataset, the final parking coordinates are extracted. The geomagnetic piezoelectric monitoring network activates the ground pressure sensor array of the corresponding parking space to collect the force distribution data of the vehicle tires on the ground. The ground pressure sensor array records the tire pressure waveform through piezoelectric sensors, generating feature data containing timestamps, parking coordinates and pressure waveforms, which are then integrated into a pressure waveform dataset containing the tire force distribution.

[0095] Based on the pressure waveform dataset, waveform features of tire force distribution, such as pressure peak and wheel axle spacing, are extracted. The waveform features are compared with a preset vehicle model database to identify vehicle types and form a vehicle type dataset containing vehicle type and parking coordinates. The preset vehicle model database is set based on the physical characteristics of common vehicle types and historical parking lot vehicle data, combined with the parking needs of the scenic area.

[0096] Based on the vehicle type dataset, the proportion of each vehicle type in the current parking lot is statistically analyzed. By accumulating the vehicle type corresponding to each parking coordinate, the proportion of each type of vehicle is obtained, generating real-time statistical results. Based on the real-time statistical results, the vehicle type proportion is extracted, and spatial reconstruction is triggered according to predefined rules. The predefined rules are: set thresholds for the proportion of buses and new energy vehicles based on the parking lot's space capacity. If the proportion of buses exceeds the bus proportion threshold, a wave-shaped dedicated lane is activated, and mechanical parking spaces are closed. If the proportion of new energy vehicles exceeds the new energy vehicle proportion threshold, a charging area is opened. For example, the bus proportion threshold is 50%, the new energy vehicle proportion threshold is 30%, and the scenic area parking lot currently has 100 parking spaces. The geomagnetic piezoelectric monitoring network shows the vehicle type distribution as follows: 40 buses, 35 new energy vehicles, and 25 sedans. The bus proportion threshold is not triggered, but the new energy vehicle proportion threshold is triggered. The parking lot layout is updated using geographic information system tools, generating a real-time spatial configuration map that includes lane topology changes and functional area adjustments. This real-time spatial configuration map is then transmitted to the scheduling decision unit via the MQTT protocol, allowing the scheduling decision unit to update parking lot management and resource allocation, forming a closed-loop optimization and completing the parking facility layout.

[0097] This embodiment also provides a computer device applicable to the design method for optimizing the layout of parking facilities at urban attractions, comprising: a memory and a processor; the memory is used to store computer-executable instructions, and the processor is used to execute the computer-executable instructions to implement the design method for optimizing the layout of parking facilities at urban attractions as proposed in the above embodiment.

[0098] The computer device can be a terminal, comprising a processor, memory, communication interface, display screen, and input devices connected via a system bus. The processor provides computing and control capabilities. The memory includes non-volatile storage media and internal memory. The non-volatile storage media stores the operating system and computer programs. The internal memory provides an environment for the operation of the operating system and computer programs stored in the non-volatile storage media. The communication interface is used for wired or wireless communication with external terminals; wireless communication can be achieved through Wi-Fi, carrier networks, NFC (Near Field Communication), or other technologies. The display screen can be an LCD screen or an e-ink screen. The input devices can be a touch layer covering the display screen, buttons, a trackball, or a touchpad on the computer device's casing, or an external keyboard, touchpad, or mouse.

[0099] This embodiment also provides a storage medium storing a computer program, which, when executed by a processor, implements the design method for optimizing the layout of parking facilities at urban scenic spots as proposed in the above embodiments. The storage medium can be implemented by any type of volatile or non-volatile storage device or a combination thereof, such as Static Random Access Memory (SRAM), Electrically Erasable Programmable Read-Only Memory (EEPROM), Erasable Programmable Read Only Memory (EPROM), Programmable Red-Only Memory (PROM), Read-Only Memory (ROM), magnetic storage, flash memory, magnetic disk, or optical disk.

[0100] In summary, this invention addresses the shortcomings of historical data distortion caused by sudden events by quantifying the degree of external impact through an event intensity factor, enabling dynamic adaptive adjustment of model parameters, and establishing a closed-loop feedback mechanism between the impact of sudden events and the model response. This significantly improves prediction accuracy in highly volatile scenarios and avoids the decision-making failure of static models in emergency situations. Furthermore, it generates a scheduling instruction set based on the electronic fence range and parking resource demand prediction map, overcoming the problem of traditional scheduling strategies being disconnected from the scope of event impact. Through a spatial topology matching algorithm, it achieves precise adaptation of resource supply, shortens the emergency response cycle, and reduces the risk of secondary congestion caused by resource mismatch.

[0101] It should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and are not intended to limit it. Although the present invention has been described in detail with reference to preferred embodiments, those skilled in the art should understand that modifications or equivalent substitutions can be made to the technical solutions of the present invention without departing from the spirit and scope of the technical solutions of the present invention, and all such modifications or substitutions should be covered within the scope of the claims of the present invention.

Claims

1. A design method for optimizing the layout of parking facilities at urban scenic spots, characterized in that: include, Real-time acquisition of emergency data around the scenic area, parsing of emergency types and geographical coordinates, and generation of structured datasets; the structured datasets include fence boundary coordinates and event intensity factors; The generation of the structured dataset specifically includes the following steps. The data on emergencies are organized into a raw emergency dataset. Event descriptions and category information are extracted from the raw emergency dataset. A pre-trained Chinese text classifier is used to label the event types. The classified emergency dataset is output, and the event intensity factor is determined. Use GIS tools to generate fence boundary coordinates based on event intensity factors, and integrate them with event intensity factors and classified emergency event datasets into a structured dataset; Based on the structured dataset, a gradient electronic fence is generated on the GIS platform. At the same time, the fence boundary coordinates and event intensity factors are input into the pre-built parking demand prediction model for parameter recombination, and a parking resource demand prediction map is output. Based on the electronic fence range and parking resource demand forecast map, a response strategy is generated to form a scheduling instruction set; The identity of the reserved vehicle is verified through the V2X vehicle-road cooperative mechanism, and the vehicle's physical parameters are matched with the available parking space resources in the dispatch instruction set to generate a three-dimensional navigation package. The system performs light strip guidance based on the 3D navigation package along the travel path, and simultaneously integrates roadside anchor points and vehicle navigation data for position correction, outputting parking coordinates. The output berthing coordinates specifically include the following steps. The system extracts the travel path and parking space coordinates from the 3D navigation package, deploys programmable road studs through optical guidance equipment to form a dynamic light strip, sets blue light to be constantly on for straight sections, sets blue light to flash at turning nodes, and sets red light arrows for parking spaces, and outputs a programmable road stud instruction set to guide vehicles to travel along the dedicated travel path. Based on the travel path in the 3D navigation package, the Kalman filter algorithm is used to fuse roadside anchor point signals and vehicle inertial navigation data to correct the vehicle position and output a positioning correction dataset. The final parking coordinates are calculated using the positioning correction dataset and parking space coordinates, and then transmitted to the geomagnetic piezoelectric monitoring network via the V2X communication interface to output the parking coordinates. The system locates the target vehicle based on the parking coordinates, collects the tire force waveform, identifies vehicle characteristics and calculates their proportions, triggers spatial reconstruction, and generates a real-time spatial configuration map that is fed back to the decision-making unit.

2. The design method for optimizing the layout of parking facilities at urban scenic spots as described in claim 1, characterized in that: The output parking resource demand prediction map specifically includes the following steps. Event types and geographic coordinates are extracted from structured datasets. GIS tools are used to generate gradient electronic fences with predefined ranges based on event types, and the data is then compiled into an electronic fence dataset. The fence boundary coordinates and event intensity factors are extracted from the electronic fence dataset to form standardized input data. This data is then input into a parking demand prediction model based on a multilayer perceptron neural network, triggering dynamic parameter reorganization. The feature weights are adjusted using a weighted average formula, and a parking resource demand prediction map is output.

3. The design method for optimizing the layout of parking facilities at urban scenic spots as described in claim 2, characterized in that: Based on the electronic fence range and parking resource demand forecast map, a response strategy is generated, forming a scheduling instruction set, which specifically includes the following steps. Event types and fence boundary coordinates are extracted from the electronic fence dataset. A three-level response strategy is formulated by combining the parking resource demand prediction map, and a scheduling scheme dataset is generated. The parking lot opening and closing status, diversion paths, and shared resource locations are extracted from the scheduling scheme dataset, organized into a standardized scheduling instruction set stored in tabular form, and transmitted to the V2X vehicle-road cooperative unit.

4. The design method for optimizing the layout of parking facilities at urban scenic spots as described in claim 3, characterized in that: The generation of the 3D navigation package specifically includes the following steps. Extract vehicle IDs from the dispatch instruction set, verify vehicle identity using encrypted digital signatures via roadside communication equipment, and output a verification dataset confirming vehicle legitimacy. The vehicle information database is queried using the vehicle IDs in the validation dataset to obtain the vehicle's physical parameters. The available parking space resources in the scheduling instruction set are then matched using a geometric constraint filtering method to obtain the coordinates of the matched parking space. A path planning algorithm is then used to calculate the optimal passage path and generate a 3D navigation package, taking the vehicle's current position, the coordinates of the matched parking space, and the diversion path in the scheduling instruction set as inputs.

5. The design method for optimizing the layout of parking facilities at urban scenic spots as described in claim 4, characterized in that: The system locates the target vehicle based on its parking coordinates, collects tire force waveforms, identifies vehicle model characteristics and calculates their proportions, triggers spatial reconstruction, and generates a real-time spatial configuration map which is then fed back to the decision-making unit. The specific steps include the following: Based on the parking coordinates, the geomagnetic piezoelectric sensor array is activated to collect the tire force waveform, and after extracting the waveform features, it is compared with the preset vehicle model database to identify the vehicle type and output the vehicle type dataset. The vehicle type dataset is used to calculate the proportion of different vehicle types. Spatial reconstruction is triggered based on the predefined thresholds for the proportion of buses and new energy vehicles. The parking lot layout is updated using GIS tools to generate a real-time spatial configuration map, which is then transmitted to the decision-making unit via the MQTT protocol.

6. The design method for optimizing the layout of parking facilities at urban scenic spots as described in claim 1, characterized in that: Obtaining the pre-trained Chinese text classifier specifically includes the following steps: We collected a large-scale Chinese text dataset related to emergencies, manually labeled the event types, and divided it into training and test sets. Download the pre-trained Chinese BERT model, add a classification layer, and output the event type probability using the Softmax function; Set the hyperparameters of the pre-trained Chinese BERT model, train the classifier of the pre-trained Chinese BERT model using the cross-entropy loss function, and evaluate the performance on the test set to obtain the pre-trained Chinese text classifier.

7. A computer device comprising a memory and a processor, wherein the memory stores a computer program, characterized in that: When the processor executes the computer program, it implements the steps of the design method for optimizing the layout of parking facilities in urban scenic spots as described in any one of claims 1 to 6.

8. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by the processor, it implements the steps of the design method for optimizing the layout of parking facilities in urban scenic spots as described in any one of claims 1 to 6.

Images