Central business district intelligent parking management system and method based on Internet of Things

Abstract

The invention relates to the technical field of parking management systems, in particular to a central business district intelligent parking management system and method based on the Internet of Things. The parking management system comprises a data acquisition module, a central data management module, a tense model analysis module and a dynamic parking space distribution and optimization module. The data acquisition module acquires multi-dimensional operation data in real time through Internet of Things equipment; the central data management module is used for classifying and analyzing the data and generating a structured data set containing spatial distribution, a time sequence and transaction characteristics; the tense model analysis module constructs a comprehensive decision packet including regional scheduling priorities, time division coefficients and a service level mapping table; the dynamic parking space distribution and optimization module implements a three-level distribution mechanism and a differentiated parking space distribution strategy. Through the closed-loop intelligent control system, peak congestion is relieved, the parking space turnover rate is increased, the parking space searching time of the temporarily-parked vehicles is effectively shortened, and dynamic optimal configuration of parking resources in the central business district is achieved.

Description

Technical Field

[0001] The present invention relates to the technical field of parking management systems, and in particular to an intelligent parking management system and method for a central business district based on the Internet of Things. Background Art

[0002] As the core area of ​​urban economic activity, the Central Business District (CBD) faces severe parking management challenges year-round. For example, during the morning rush hour (7:30-9:30), its underground parking lot handles over 80% of the influx of vehicles, while over 60% of the parking spaces are occupied by fixed vehicles for long periods of time, causing temporary parked vehicles to spend over 15 minutes searching for parking spaces. This resource mismatch not only reduces parking space turnover efficiency but also exacerbates congestion at entrances and exits during peak hours. During the evening rush hour (17:30-19:30), the concentrated departure of fixed and temporarily parked vehicles further triggers payment queues and traffic congestion, causing the congestion index at conflict points to increase by 40%.

[0003] While the fixed-space allocation model has achieved a certain degree of electronic parking management, it still has significant shortcomings in dynamically optimizing parking resource allocation. Specifically, this model relies too heavily on a static allocation mechanism for fixed parking spaces. This makes it impossible to monitor changes in parking lot traffic in real time, nor can it dynamically adjust parking allocation strategies based on actual demand. Ultimately, temporarily parked vehicles are forced to spend extended periods searching for available spaces due to the rigid allocation of parking resources. This static management model severely restricts parking lot operational efficiency, particularly during peak hours, when it is impossible to flexibly release parking resources to meet temporary parking needs. Summary of the Invention

[0004] The present invention provides an IoT-based intelligent parking management system for central business districts. By constructing a comprehensive decision package containing regional scheduling priorities, time period diversion coefficients, and a service level mapping table, and implementing dynamic allocation of parking space resources based on this decision package, the present invention solves the problems raised in the above-mentioned background technology, namely:

[0005] The traditional fixed parking space allocation model relies too much on the static division mechanism of fixed parking spaces, resulting in temporary parking vehicles being forced to cruise for a long time looking for available parking spaces due to the rigid allocation of parking space resources.

[0006] To achieve the above objectives, the intelligent parking management system includes a data acquisition module, which is used to collect key parking lot operation data and transmit it to a central data management module. The central data management module receives the key parking lot operation data, classifies and analyzes it, and generates a structured feature data set including a spatial distribution set, a time series set, and a transaction feature set. It also includes a temporal model analysis module and a dynamic parking space allocation and optimization module.

[0007] The temporal model analysis module makes a strategic execution basis for the dynamic parking allocation and optimization module through a comprehensive decision package, and the comprehensive decision package includes:

[0008] Regional scheduling priority for allocating parking resources in high-load areas;

[0009] Time period diversion coefficient for differentiated diversion control of vehicles;

[0010] A service level mapping table for matching user classification labels with service policies;

[0011] The steps of constructing the comprehensive decision package by the temporal model analysis module include:

[0012] The spatial distribution set is estimated by Gaussian kernel density to generate a heat map to reflect the load intensity of each area in the parking lot;

[0013] Predict the vehicle arrival rate curve for the time series set using the LSTM neural network;

[0014] Generate user classification labels for transaction feature sets through payment behavior cluster analysis;

[0015] Normalize the heat map, vehicle arrival rate curve, and user classification labels to generate a comprehensive decision package;

[0016] The dynamic parking space allocation and optimization module dynamically adjusts the parking space allocation strategy based on the comprehensive decision package to dynamically configure parking space resources.

[0017] The above technical solution adopts a hierarchical and progressive data processing architecture design, which is mainly based on the in-depth deconstruction of the complexity of parking management in the central business district. Traditional systems often stop at basic data collection and simple statistical analysis, and are unable to cope with the ever-changing nature of parking demand. This solution solves the defect of the single data dimension of the static management system by constructing a multi-dimensional feature data set to transform discrete parking data into decision-making factors with spatiotemporal correlation. The temporal model analysis module adopts a fusion of machine learning and traditional algorithms to capture macro traffic patterns and identify micro user characteristics, avoiding the decision-making bias caused by a single algorithm. The dynamic allocation module innovatively introduces a multi-parameter collaborative control mechanism. Through the organic coordination of priority scheduling, traffic diversion and service grading, it forms a three-dimensional decision-making system, overcoming the shortcomings of the traditional system with a single adjustment dimension and slow response.

[0018] On this basis, the central data management module establishes an association map between vehicle information and transaction records, and comprehensively evaluates parking space utilization efficiency through a dynamic weight calculation model.

[0019] In another technical solution, the parking space allocation strategy includes:

[0020] Release fixed parking spaces as temporary buffer zones based on regional scheduling priorities;

[0021] Implement a three-level diversion mechanism based on the time period diversion coefficient;

[0022] Allocate parking spaces based on user rating tags.

[0023] This technical solution breaks through the limitations of data silos in traditional parking management systems by establishing a deep correlation mapping between vehicle information and transaction records, upgrading parking space utilization efficiency assessment from simple quantitative statistics to multi-dimensional value analysis. Without this data correlation mechanism, the system would be unable to identify the value characteristics of high-frequency users, resulting in a loss of precision in resource allocation strategies. The introduction of a dynamic weight calculation model transforms discrete operational parameters into unified decision-making indicators, resolving the issue of inconsistent traditional empirical allocation standards. In terms of parking space allocation strategies, the design of a three-level diversion mechanism fills the gap in the lack of adaptability of a single diversion solution, ensuring that the system maintains optimal operating status under different pressure scenarios through graded responses. The precise matching of user classification tags with parking space allocation fundamentally changes the traditional service model, achieving a qualitative shift from extensive management to refined operations.

[0024] A second object of the present invention is to provide a central business district intelligent parking management method based on the Internet of Things, comprising the following method steps:

[0025] S1. Receive a structured feature dataset including a spatial distribution set, a time series set, and a transaction feature set;

[0026] S2: Generate heatmaps for spatial distribution sets using Gaussian kernel density estimation, predict vehicle arrival rate curves for time series sets using LSTM neural networks, and generate user classification labels for transaction feature sets using payment behavior clustering analysis.

[0027] S3. Normalize the heat map, vehicle arrival rate curve, and user classification labels to generate a comprehensive decision package;

[0028] S4. Make a basis for strategy execution through comprehensive decision-making package;

[0029] S5. Release fixed parking spaces as temporary buffer zones based on regional scheduling priorities, implement a three-level diversion mechanism based on time period diversion coefficients, and allocate parking spaces based on user classification tags.

[0030] Compared with the prior art, the present invention has the following beneficial effects:

[0031] 1. By building a closed-loop control system based on "data collection - feature extraction - temporal modeling - dynamic decision-making," the system achieves intelligent management of parking resources in the central business district. This system uses a temporal model analysis module to accurately predict peak traffic periods. Combined with the three-dimensional collaborative mechanism of the dynamic parking allocation module, this system effectively alleviates entrance and exit congestion during morning and evening peak hours, significantly improving vehicle traffic efficiency.

[0032] 2. By monitoring parking space usage in real time and dynamically adjusting allocation strategies, the efficiency of temporarily parked vehicles finding parking spaces has been significantly improved. The LSTM neural network-based parking demand prediction and intelligent navigation guidance significantly shorten the average parking search time for temporarily parked vehicles, resolving the issue of long search times for temporarily parked vehicles under the traditional fixed parking space allocation model. BRIEF DESCRIPTION OF THE DRAWINGS

[0033] Figure 1 It is a schematic diagram of the overall process structure of the present invention;

[0034] Figure 2 This is a flow chart of the data acquisition module of the present invention;

[0035] Figure 3 This is a flow chart of the temporal model analysis module of the present invention;

[0036] Figure 4 This is a flow chart of the intelligent parking management method for central business districts based on the Internet of Things of the present invention.

[0037] The meaning of each number in the figure is:

[0038] 100. Data acquisition module; 200. Central data management module; 300. Temporal model analysis module; 400. Dynamic parking space allocation and optimization module. DETAILED DESCRIPTION

[0039] The following will clearly and completely describe the technical solutions in the embodiments of the present invention in conjunction with the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative efforts are within the scope of protection of the present invention.

[0040] At the same time, some technical terms are explained here:

[0041] The LSTM neural network is a special recurrent neural network (RNN) that can effectively capture long-term dependencies in time series data by introducing input gates, forget gates, and output gates. In this paper, the LSTM model dynamically predicts parking lot traffic flow over the next two hours by memorizing the temporal characteristics of historical vehicle entry and exit data (such as the peak traffic flow fluctuation cycle in the morning and evening).

[0042] NB-IoT narrowband communication is a low-power wide-area network technology designed for the large-scale connection needs of IoT devices;

[0043] Gaussian kernel density estimation is a non-parametric statistical method that estimates the spatial density distribution by placing a Gaussian kernel function at each data point (such as the parking space location). In this paper, this method is used to calculate the degree of vehicle aggregation in each area of ​​the parking lot and generate a heat map reflecting the intensity of parking space usage.

[0044] At present, the traditional fixed parking space allocation model relies too much on the static division mechanism of fixed parking spaces, resulting in the problem that temporarily parked vehicles are forced to cruise for a long time to find available parking spaces due to the rigid allocation of parking space resources. The present invention provides a central business district intelligent parking management system based on the Internet of Things, see Figure 1 As shown, it includes a data acquisition module 100, a central data management module 200, a temporal model analysis module 300 and a dynamic parking space allocation and optimization module 400. Through the collaborative work of multiple modules, dynamic optimization configuration of parking space resources is realized, achieving the goals of maximizing parking space utilization, significantly improving the efficiency of temporarily parked vehicles finding parking spaces, and increasing parking lot revenue.

[0045] like Figure 2 As shown, data acquisition module 100 is the fundamental data source for the entire intelligent parking management system. It enables comprehensive data collection through IoT devices deployed at key parking nodes. This module primarily relies on intelligent barcode scanning devices installed at parking entrances and exits. These devices are equipped with high-precision license plate recognition cameras and mobile payment terminals, enabling 24-hour, uninterrupted collection of key parking lot operational data.

[0046] During vehicle entry and exit, the data acquisition module 100 records three types of vehicle entry and exit data in real time: first, the license plate number, which serves as the basis for subsequent differentiation between fixed-parking and temporary parking types. The license plate number is a unique identifier for each vehicle. Accurately identifying the license plate number through a scanner enables precise management and information tracking of specific vehicles. Second, time data, including vehicle entry and exit times, is accurate to the second. These timestamps provide the basis for subsequent parking duration calculation and fee accounting. Finally, parking space information, namely the specific parking space number or area identifier, is key to precise parking space management. All of this data is uploaded to the system in real time via the IoT communication protocol.

[0047] In addition to basic vehicle information, the data collection module 100 also collects parking lot financial data. This includes the specific usage of various payment methods, such as mobile payments (apps, WeChat, etc.), traditional cash payments, pre-paid card payments, and emerging contactless payments. The system records the incoming amount, refund amount, and final actual amount collected for each transaction, and also notes the transaction time and update date to ensure the integrity and traceability of financial data.

[0048] To meet the needs of parking lot operations and management, the data acquisition module 100 also includes a dedicated traffic monitoring unit to collect parking lot data. Sensors located at entrances and exits such as T1, T2, and T3 provide real-time statistics on the total number of vehicles entering and exiting during each period, distinguishing between vehicles in fixed and non-fixed parking spaces. The system continuously updates key indicators such as the total number of vehicles on site and the number of remaining temporary parking spaces, providing data support for subsequent dynamic parking space allocation.

[0049] All collected critical parking lot operational data, including vehicle entry and exit data, financial data, and parking lot flow data, undergoes preliminary verification and formatting to ensure data accuracy and consistency. This pre-processed data is transmitted in real time to the central data management module 200 via a dedicated data interface, providing a reliable data source for subsequent data storage, analysis, and application. The data collection module 100 utilizes a distributed architecture, ensuring that even a single collection point failure will not affect the normal operation of the entire system, ensuring the continuity and integrity of data collection.

[0050] The Central Data Management Module 200, the system's core data processing unit, is responsible for standardized storage and intelligent analysis of various parking data collected by the front-end. Based on a relational database architecture, this module strictly adheres to system design requirements and fully records and manages all parking-related data.

[0051] The central data management module 200 first categorizes the collected data. Regarding fund data management, the system comprehensively collects various payment methods, including project name, app payment, WeChat MeiTian payment, cash payment, MeiTian unlicensed vehicle scan payment, WeChat Financial City payment, Financial City unlicensed vehicle scan payment, pre-stored value card payment, and contactless payment. It also records the amount of each transaction received, refunded, and actually received, and annotates the transaction update time and date. After cleaning and standardization, this data forms a complete fund flow record, providing an accurate basis for financial management and revenue analysis.

[0052] In terms of vehicle entry and exit data management, the module records detailed information on entry and exit traffic, including project name, total number of entries today, total number of exits today, total number of entries from T1, total number of exits from T1, total number of entries from T2, total number of exits from T2, total number of entries from T3, and total number of exits from T3. It also tracks parking status indicators such as the number of contracted fixed parking spaces, number of contracted non-fixed parking spaces, number of fixed vehicles on site, number of non-fixed vehicles on site, number of long-term rental vehicles on site, number of temporary parking spaces on site, number of whitelisted vehicles on site, number of public inspection vehicles on site, number of non-long-term rental vacant parking spaces, number of non-fixed vacant parking spaces, total number of vehicles on site, number of vehicles exceeding the limit of entrants, number of temporary parking spaces, and number of remaining non-fixed parking spaces. The system updates the total number of temporary parking entries and exits today, as well as the update time, in real time to ensure data timeliness.

[0053] During data processing, the module integrates and analyzes data using a multi-stage algorithm. First, the raw data is formatted and outliers are filtered. Then, a mapping between vehicle information and transaction records is established. Finally, a dynamic weight calculation model is used to comprehensively evaluate parking space utilization efficiency. This model uses the following formula to calculate the weight of temporary parking vehicle access:

[0054]

[0055] Where W represents the admission weight of temporary parking vehicles;

[0056] C temp Indicates the number of temporary parking spaces in the data acquisition module 100;

[0057] C non-fix Indicates the available capacity of non-fixed parking spaces. Its value is generated by dividing the number of contracted non-fixed parking spaces by the number of non-fixed parking spaces on site.

[0058] T avg Indicates the average parking search time calculated based on the difference between the vehicle entry time and the system parking space allocation time;

[0059] 15 means the operating threshold for temporarily parked vehicles to find a parking space for up to 15 minutes

[0060] In the parking management system, the average time for vehicles to find their location is T avg Efficiency reduction factor relative to the maximum allowed duration of 15 minutes;

[0061] P mobile Indicates the total transaction amount of the integrated data collection module 100APP payment + WeChat Meitian payment + Financial City payment;

[0062] P total Indicates the total value of the actual amount received field in the data collection module 100.

[0063] The model output consists of three core datasets: a spatial distribution set, a time series set, and a transaction feature set. The spatial distribution set includes a parking map composed of temporary parking spaces, the number of non-fixed remaining parking spaces, and the total number of vehicles on site. The time series set generates a traffic flow trend curve based on the time-of-day distribution of total parking spaces and the duration of temporary parking on site. The transaction feature set includes a payment method statistics table showing the proportion of mobile payments (app + WeChat) and the frequency of pre-paid card payments. The system ultimately generates a structured feature dataset consisting of the spatial distribution set, time series set, and transaction feature set, and passes it to the temporal model analysis module 300 for dynamic parking allocation strategy formulation and vehicle diversion plan generation.

[0064] like Figure 3 As shown, the temporal model analysis module 300 receives a structured feature data set (including a spatial distribution set, a time series set, and a transaction feature set) from the central data management module 200, and first processes the spatial distribution set data: the data on the number of temporary parking spaces, the number of non-fixed remaining parking spaces, and the total number of vehicles on site are input into a temporal analysis model based on Gaussian kernel density estimation. The algorithm calculates the vehicle density distribution within a radius of 50 meters around each parking space (using a Gaussian kernel function with a standard deviation of 10 meters) to generate a heat map reflecting the load intensity of each area of ​​the parking lot. At the same time, the time period distribution data of the total number of entries today and the length of stay data of the temporary parking spaces on site in the time series set are imported into a traffic prediction model constructed based on an LSTM neural network. By analyzing the historical traffic fluctuation patterns and the deviation from real-time data, the vehicle arrival rate curves of each entrance and exit within the next 2 hours are predicted.

[0065] Based on the mobile payment ratio and pre-stored value card payment frequency data in the transaction feature set, the module uses a payment behavior clustering analysis algorithm to establish payment methods and a parking duration association matrix. By performing spatiotemporal matching of APP + WeChat payment data with vehicle entry and exit timestamps, it identifies high-frequency short-term parking user groups (60% of single parking sessions <30 minutes) and low-frequency long-term parking user groups (20% of single parking sessions >4 hours), and generates user classification labels accordingly. All analysis results are normalized to form three types of decision parameters: a spatial load coefficient based on a heat map, a time pressure index based on traffic prediction, and user classification labels based on payment clustering. These parameters are fused through a dynamic weight algorithm to generate a comprehensive decision package containing regional scheduling priorities, time period diversion coefficients, and a service level mapping table. Ultimately, this package is passed to the dynamic parking space allocation and optimization module 400 as the basis for policy execution.

[0066] Among them, the regional scheduling priority is a numerical indicator calculated based on the real-time load conditions of each parking lot area, which guides the system to prioritize parking resources in high-load areas. The time-period diversion coefficient is a dynamic parameter generated based on historical data and real-time prediction results, which is used to differentiate the diversion control of vehicles at each entrance and exit during different time periods. The service level mapping table matches user classification labels with corresponding service policies. For example, high-frequency, short-term users are automatically assigned parking spaces near the exit, while low-frequency, long-term users are allocated parking spaces in marginal areas. These three parameters work together to form the core basis for the system's intelligent decision-making, ensuring efficient utilization of parking resources and orderly guidance of vehicles. This process enables the system to respond to fluctuations in parking demand in real time, optimizing the path length of temporary parking vehicles by 40% during the morning peak period and reducing the waiting time for vehicles leaving the parking lot by 35% during the evening peak period.

[0067] The dynamic parking allocation and optimization module 400, serving as the system's policy execution layer, receives the comprehensive decision package (including regional scheduling priorities, time-period diversion coefficients, and user classification tags) output by the temporal model analysis module 300. Using a dynamic rules engine, it implements real-time parking resource allocation and vehicle guidance. This module spatially overlays regional scheduling priority data with the parking lot's electronic map. During the morning rush hour (7:30-9:30 AM), when the load factor in the non-fixed area exceeds 0.8, it automatically releases 20% of fixed parking spaces as a temporary buffer zone. High-priority temporary parking vehicles (W > 0.6) are prioritized for allocation to this zone using the admission weight formula W. A three-tiered diversion mechanism is implemented based on the time-period diversion coefficients. For fixed vehicles, the system generates an optimal navigation path based on their contracted parking space number (calculating the shortest travel distance using the Dijkstra algorithm), which is displayed in real time on the on-site LED guidance screen. Temporary parking vehicles are allocated spaces differently based on their user classification tags: high-frequency, short-term users are automatically assigned to spaces within 50 meters of the exit, while low-frequency, long-term users are allocated to the edge of the parking lot. After the diversion plan is generated, the system pushes lane-level navigation instructions to the user's mobile phone APP through the IoT gateway, and synchronizes the regional scheduling heat map to the on-duty personnel's mobile terminal (such as a tablet computer).

[0068] During operation, the module achieves dynamic scheduling through the following real-time interaction mechanism: When a vehicle enters the identification area, the millimeter-wave radar at the entrance scans the license plate at a rate of 30 frames per second. After compression by the edge computing device, the data is transmitted to the central processing unit via the LoRa wireless protocol. The system completes vehicle type determination within 150 milliseconds. For fixed vehicles, the real-time status of the bound parking space is immediately verified. If the load factor of the target area exceeds 0.8, dynamic path planning based on the Dijkstra algorithm is triggered. This algorithm integrates real-time vehicle speed data collected by the geomagnetic sensor (update frequency 5Hz) to automatically generate alternative routes around high-congestion nodes. The distributed LED guidance system refreshes the navigation arrow with a 0.2-second delay. If the load factor of the target area is normal, the standard navigation path is directly pushed to the vehicle owner's app.

[0069] The processing flow for temporarily parked vehicles is deeply integrated with the access weight mechanism: the system calls the mobile payment ratio, the number of non-fixed remaining parking spaces, and the average parking time output by the temporal model analysis module 300 in real time and performs dynamic calculations to derive the current vehicle weight value. When the weight value exceeds the preset threshold of 0.6, the vehicle is classified as a Class B user and the system automatically assigns it to a dynamic buffer zone within 80 meters of the nearest exit (the fixed parking zone released during the morning rush hour). A text message containing a three-dimensional navigation map is sent to the owner via NB-IoT narrowband communication. Low-weight vehicles (W<0.6) enter the waiting queue at the edge of the area, and the system dynamically displays the estimated waiting time curve on the entrance display screen (based on the LSTM-predicted traffic attenuation model for the next 15 minutes).

[0070] When the pressure difference of traffic flows at multiple entrances reaches the diversion trigger condition set by the system (such as the instantaneous traffic volume at T1 port reaches 2.5 times that of T3 port), the dynamic allocation engine automatically activates the pressure equalization protocol. The protocol includes a three-level response mechanism: in the primary response stage (pressure ratio <1.8), diversion suggestions are only pushed through the APP; in the intermediate response stage (1.8≤pressure ratio<2.5), the backup channel is forced to open and the LED guidance plan is modified; in the advanced response stage (pressure ratio ≥2.5), the lane function is temporarily adjusted (such as changing exit E to an entrance). During specific execution, the system generates a control instruction set including diversion ratio, path weight and exception handling rules based on the 30-minute traffic flow trend predicted by the temporal model. These instructions are synchronously transmitted to the on-duty terminal through a dedicated 5G slicing network, and presented on the tablet device as a dynamic streamline map superimposed with a heat map.

[0071] To handle emergencies in special scenarios, the system establishes a dual-channel decision-making mechanism. When special vehicles such as ambulances and fire trucks are identified, green channel mode is immediately activated: all parked vehicles on the target route are forcibly cleared (with an emergency vehicle relocation notification pushed via the app), and the barrier gate remains open. For unlicensed vehicles or vehicles that fail identification (with an error rate of <0.3%), a manual verification interface automatically pops up on the on-site terminal, requiring on-site personnel to take a photo of the vehicle's features (such as the annual inspection sticker on the window) for a secondary match. Once a match is successful, the vehicle is manually added to the dispatch queue. All emergency operation data is encrypted and labeled for subsequent model reinforcement learning training.

[0072] In summary, the present invention predicts parking space demand by integrating real-time traffic flow monitoring with historical data analysis, evaluates parking space usage priority based on a dynamic weight model, and links the on-site navigation system to adjust the allocation strategy in real time. It effectively solves the problems of idle resources and difficulty in finding parking spaces caused by the traditional fixed parking space allocation model, and achieves the operational optimization goals of improving parking space turnover and reducing congestion during peak hours.

[0073] like Figure 4 As shown, the second purpose of this embodiment is to provide a central business district intelligent parking management method based on the Internet of Things, including the following method steps:

[0074] S1. Receive a structured feature dataset including a spatial distribution set, a time series set, and a transaction feature set;

[0075] S2: Generate heatmaps for spatial distribution sets using Gaussian kernel density estimation, predict vehicle arrival rate curves for time series sets using LSTM neural networks, and generate user classification labels for transaction feature sets using payment behavior clustering analysis.

[0076] S3. Normalize the heat map, vehicle arrival rate curve, and user classification labels to generate a comprehensive decision package;

[0077] S4. Make a basis for strategy execution through comprehensive decision-making package;

[0078] S5. Release fixed parking spaces as temporary buffer zones based on regional scheduling priorities, implement a three-level diversion mechanism based on time period diversion coefficients, and allocate parking spaces based on user classification tags.

[0079] The above shows and describes the basic principles, main features, and advantages of the present invention. Those skilled in the art should understand that the present invention is not limited to the above embodiments. The above embodiments and descriptions are merely preferred examples of the present invention and are not intended to limit the present invention. Various changes and improvements may be made to the present invention without departing from the spirit and scope of the present invention. Such changes and improvements fall within the scope of the present invention. The scope of protection claimed in the present invention is defined by the appended claims and their equivalents.

Claims

1. An intelligent parking management system for a central business district based on the Internet of Things, comprising a data acquisition module (100), wherein the data acquisition module (100) is used to collect key parking lot operation data and transmit the data to a central data management module (200), wherein the central data management module (200) receives the key parking lot operation data and classifies and analyzes the data to generate a structured feature data set including a spatial distribution set, a time series set, and a transaction feature set, wherein: It also includes a temporal model analysis module (300) and a dynamic parking space allocation and optimization module (400); The temporal model analysis module (300) makes a strategy execution basis for the dynamic parking space allocation and optimization module (400) through a comprehensive decision package, wherein the comprehensive decision package includes: Regional scheduling priority for allocating parking resources in high-load areas; Time period diversion coefficient for differentiated diversion control of vehicles; A service level mapping table for matching user classification labels with service policies; The temporal model analysis module (300) constructs the comprehensive decision package in the following steps: The spatial distribution set is estimated by Gaussian kernel density to generate a heat map to reflect the load intensity of each area in the parking lot; Predict the vehicle arrival rate curve for the time series set using the LSTM neural network; Generate user classification labels for transaction feature sets through payment behavior cluster analysis; Normalize the heat map, vehicle arrival rate curve, and user classification labels to generate a comprehensive decision package; The dynamic parking space allocation and optimization module (400) dynamically adjusts the parking space allocation strategy based on the comprehensive decision package to dynamically configure parking space resources.

2. The IoT-based central business district intelligent parking management system according to claim 1, characterized in that: The key parking lot operation data includes vehicle entry and exit data, capital data and parking lot data flow.

3. The central business district intelligent parking management system based on the Internet of Things according to claim 2 is characterized by: The vehicle entry and exit data includes a license plate number, time data and parking space information, and the time data includes a vehicle entry time and a vehicle exit time.

4. The central business district intelligent parking management system based on the Internet of Things according to claim 1 is characterized by: The central data management module (200) classifies parking lot key operation data into capital data and vehicle entry and exit data, and performs format verification and abnormal value filtering on the data.

5. The central business district intelligent parking management system based on the Internet of Things according to claim 4 is characterized by: The central data management module (200) establishes an association map between vehicle information and transaction records, and comprehensively evaluates parking space utilization efficiency through a dynamic weight calculation model.

6. The central business district intelligent parking management system based on the Internet of Things according to claim 1 is characterized by: The spatial distribution set includes the number of temporary parking spaces, the number of non-fixed remaining parking spaces, and the total number of vehicles on site.

7. The central business district intelligent parking management system based on the Internet of Things according to claim 1 is characterized by: The time series set is based on the traffic trend curve generated by the time period distribution of the total number of entries today and the length of time the temporary parking number is parked.

8. The central business district intelligent parking management system based on the Internet of Things according to claim 1 is characterized by: The transaction feature set includes a payment method statistics table showing the proportion of mobile payments and the frequency of pre-stored value card payments.

9. The central business district intelligent parking management system based on the Internet of Things according to claim 1 is characterized by: The parking space allocation strategy includes: Release fixed parking spaces as temporary buffer zones based on regional scheduling priorities; Implement a three-level diversion mechanism based on the time period diversion coefficient; Allocate parking spaces based on user rating tags.

10. An intelligent parking management method, characterized in that: It is used to use the central business district intelligent parking management system based on the Internet of Things as described in any one of claims 1 to 9, and the specific steps are as follows: S1. Receive a structured feature dataset including a spatial distribution set, a time series set, and a transaction feature set; S2: Generate heatmaps for spatial distribution sets using Gaussian kernel density estimation, predict vehicle arrival rate curves for time series sets using LSTM neural networks, and generate user classification labels for transaction feature sets using payment behavior clustering analysis. S3. Normalize the heat map, vehicle arrival rate curve, and user classification labels to generate a comprehensive decision package; S4. Make a basis for strategy execution through comprehensive decision-making package; S5. Release fixed parking spaces as temporary buffer zones based on regional scheduling priorities, implement a three-level diversion mechanism based on time period diversion coefficients, and allocate parking spaces based on user classification tags.

Images