A parking violation management system based on internet of things big data

Abstract

The application discloses a kind of based on big data of Internet of Things illegal parking management system, comprising: Internet of Things sensing module, for integrating various Internet of Things equipment existing in city, traffic big data is collected;Data integration module is used to establish the mapping relationship between data according to the space-time relationship of the data collected, and the mapped data is integrated to generate traffic data heat map;Illegal parking judgment module is used to compare the number of parked vehicles, location and the total number of parking spaces, location in any area according to traffic data heat map and the continuously collected traffic big data, and detect illegal parking vehicles;Illegal parking management module is used to give guidance or warning information to illegal parking vehicles, record illegal parking data and feedback to law enforcement end. Through the application, intelligent and refined management of urban traffic can be realized, and problems such as incomplete monitoring, inaccurate judgment and low law enforcement efficiency in traditional illegal parking management can be effectively solved.

Description

Technical Field

[0001] This invention belongs to the field of intelligent traffic management technology, specifically relating to a parking violation management system based on Internet of Things big data. Background Technology

[0002] With the acceleration of urbanization, the contradiction between the surge in motor vehicle ownership and the shortage of parking resources is becoming increasingly acute. Data shows that the illegal parking rate in first-tier cities in China exceeds 20% during peak hours, directly leading to a traffic congestion rate increase of more than 35%, and 30% of traffic accidents are caused by illegal parking. Traditional management models rely on manual patrols and single-point monitoring equipment (such as fixed cameras), which have three major drawbacks: large monitoring blind spots—manual patrols only cover less than 20% of main roads, and efficiency drops sharply at night; low judgment accuracy—single sensors (such as geomagnetic sensors) are easily affected by environmental interference, with a false judgment rate as high as 40%; and long response chain—the average time from the discovery of illegal parking to on-site handling exceeds 30 minutes.

[0003] Current illegal parking management technologies have significant shortcomings, with the core issues being as follows: Fragmentation at the perception layer: Single devices have limited functionality; for example, geomagnetic sensors only detect parking space occupancy and cannot identify the location of illegally parked vehicles; cameras rely on lighting conditions, resulting in a false detection rate exceeding 50% at night; millimeter-wave radar data is incompatible with traffic management platforms, leading to a multi-source data fusion failure rate exceeding 30%. Insufficient intelligence at the analysis layer: Existing systems are mostly based on static electronic fence rules, unable to dynamically adapt to temporary traffic control (such as school drop-off / pick-up times) or sudden road conditions (such as accident zones). A pilot project in one city, due to a rigid rule base, mistakenly identified vehicles that were legally temporarily parked as illegally parked, resulting in a false alarm rate exceeding 25%. Lack of a closed-loop execution layer: Traditional early warning systems rely on SMS or broadcast notifications, with an actual vehicle departure rate of less than 15%, and lack a complete record of the evidence chain for illegal parking (such as location, duration, and license plate correlation), leading to enforcement disputes accounting for over 60%. Furthermore, the independent operation of multiple systems (such as the separation of data collection and enforcement platforms) makes it difficult to achieve a complete closed loop from early warning to handling. Summary of the Invention

[0004] In view of this, the present invention proposes an illegal parking management system based on Internet of Things big data. Through this invention, intelligent and refined management of urban traffic can be achieved, effectively solving the problems of incomplete monitoring, inaccurate judgment, and low enforcement efficiency in traditional illegal parking management. This will improve urban traffic order, alleviate traffic congestion, ensure smooth roads and traffic safety, create a better travel environment for citizens, and provide decision-making basis for urban traffic management departments, thus assisting in urban traffic planning and optimization.

[0005] To achieve the above objectives, the present invention provides the following technical solution:

[0006] This invention provides a parking violation management system based on Internet of Things big data, comprising:

[0007] The Internet of Things (IoT) sensing module is used to integrate various IoT devices in the city and collect traffic big data.

[0008] The data integration module is used to establish mapping relationships between data based on the spatiotemporal relationships of the collected data, and to integrate the mapped data to generate a traffic data heat map;

[0009] The illegal parking detection module is used to compare the number and location of parked vehicles and the total number and location of available parking spaces in any area based on traffic data heatmaps and continuously collected traffic big data to detect illegally parked vehicles.

[0010] The illegal parking management module is used to provide guidance or warning information to illegally parked vehicles, record illegal parking data, and provide feedback to law enforcement.

[0011] Preferably, the IoT sensing module includes:

[0012] The multi-source device access unit is used to integrate multiple types of device data sources through a heterogeneous protocol conversion gateway. These devices include millimeter-wave radar, road monitoring cameras, geomagnetic sensors, vehicle-mounted GPS positioning devices, electronic fences, and Bluetooth sniffing devices.

[0013] The data spatiotemporal alignment unit is used to synchronize the clock errors of each device to the millisecond level using the NTP protocol, establish a unified spatial reference under the WGS-84 coordinate system, and map the vehicle position to the lane lines of the centimeter-level high-precision map by fusing multi-source positioning data through Kalman filtering.

[0014] The data verification and alignment unit is used to perform reverse correlation between vehicle location and corresponding data for data with spatiotemporal attributes based on the analysis results of vehicle location by the data spatiotemporal alignment unit, establish the mapping relationship between vehicle location, associated data, and time, and determine the data status, location status of each device, and location status of the vehicle at each judgment time point.

[0015] Preferably, the data integration module performs the following operations:

[0016] The preset target area is divided into multiple monitoring units according to blocks, and the location boundaries of each monitoring unit are determined.

[0017] Determine the current location of the data source device or the location of the monitored area, and determine the monitoring unit to which the data source device belongs;

[0018] Based on the multiple data source devices to which the monitoring unit belongs, acquire various types of traffic data generated by them;

[0019] Based on the vehicle location status determined by various traffic data, a mapping relationship between data and vehicles is established in different times and spaces, and data that has not been associated with any determined vehicle is filtered out.

[0020] Based on the vehicles currently stored in the monitoring unit and the data that establishes a mapping relationship between the monitoring unit and each stored vehicle, the attribute tensor of the multidimensional data associated with the monitoring unit is defined. The attribute tensor includes vehicle density, vehicle dwell time, vehicle location, road load, and parking space status.

[0021] Based on the multidimensional data and attribute tensors corresponding to each monitoring unit, a traffic data heatmap for the region is generated.

[0022] Preferably, the illegal parking detection module performs the following operations:

[0023] Based on traffic data heatmaps and continuously collected traffic big data, the number of parked vehicles, their location, and duration of stay are tracked and determined in any given area.

[0024] Based on traffic big data, the total number of reserved parking spaces in the area and the spatial distribution of available parking locations are determined.

[0025] Based on the tracking and judgment results of the number of parked vehicles, vehicle parking locations, and vehicle parking duration within the area, the occupancy status and occupancy duration of available parking locations are determined, and vehicles not parked at the designated locations and their corresponding parking durations are also determined.

[0026] Based on the duration of the parked vehicle's stay and the preset parking management policy corresponding to the vehicle's location, a multi-level management policy for the parked vehicle is generated and sent to the illegal parking management module.

[0027] Preferably, it also includes a parking violation risk prediction module, which performs the following operations:

[0028] For the data that has been integrated and mapped, feature vectors related to vehicles, time, and space are extracted, and feature vectors suitable for model input are constructed.

[0029] An improved spatiotemporal graph convolutional network, STGCN, is constructed, and a spatiotemporal attention mechanism is applied to process the input feature vector. In the spatiotemporal graph convolutional network:

[0030] For each IoT device, the correlation between the current time step and the previous time step is calculated using its historical data sequence to obtain the time attention weight. The time attention weight is used to characterize the importance of data at different time points to the current prediction.

[0031] Calculate the spatial correlation between each IoT device, taking into account the topology and geographical adjacency of the road network, and derive spatial attention weights to measure the influence of data from different locations on the prediction of the current location.

[0032] The input features are weighted and fused using the calculated temporal and spatial attention weights;

[0033] The weighted and fused features are input into the resource game computation network, which then performs the following operations:

[0034] Obtain the parking space supply-demand ratio and ratings of surrounding alternative parking resources in each area;

[0035] Resource game weights are calculated based on the parking space supply-demand ratio and the scores of surrounding alternative resources to obtain resource game weights between different areas. These resource game weights reflect the parking resource competition relationship between the area and surrounding alternative parking resources.

[0036] The resource game computational network takes the weighted and fused feature vector as input and concatenates it with multiple resource game weights to obtain a comprehensive feature vector;

[0037] The comprehensive feature vector is passed through a fully connected layer and output. Using a positive activation or a constrained activation function, a risk score that changes continuously with time is output as the prediction result of the illegal parking risk in the target area.

[0038] Preferably, the scoring calculation process for nearby alternative parking resources includes:

[0039] Identify multiple parking areas within a preset range around the target area as alternative parking resources in the vicinity;

[0040] Determine the distance between any alternative parking resource and the target area, the size of the parking lot, the charging standard, and user reviews, and assign weights to each of them. ,and ;

[0041] Collect relevant data on this alternative parking resource, including distance d, number of remaining parking spaces s, charging standard c, and user rating a, and normalize the data so that its value ranges between 0 and 1.

[0042] The distance is normalized to obtain distance-normalized data. : ;

[0043] The remaining parking space quantity is normalized to obtain normalized data. : ;

[0044] The fee standard is normalized to obtain the fee standard normalized data. : ;

[0045] Normalize the user reviews to obtain normalized user review data. The scale is adjusted to keep the score within the range of 0 to 5.

[0046] Its overall score is calculated using a weighted formula: ;

[0047] The overall score will be used as the evaluation result for this alternative parking resource.

[0048] Preferably, it also includes a user interaction module, which performs the following operations:

[0049] Receive parking request information sent by users, including vehicle information, estimated arrival time, destination location, and parking duration requirement;

[0050] Based on the destination location, parking resource information from multiple surrounding monitoring units is extracted from the traffic data heat map. The parking resource information includes the number of available parking spaces, their location distribution, parking fees, and parking time restrictions.

[0051] Based on parking demand and parking resource information, the system uses a pre-set recommendation algorithm to recommend the best parking location for users and generate parking navigation information that includes driving routes, estimated parking fees, and estimated parking duration.

[0052] The system sends parking navigation information to the user's terminal and updates the navigation information in real time to guide the vehicle to the recommended parking location.

[0053] After the vehicle arrives at the recommended parking location, the system monitors the vehicle's parking status. If the vehicle is illegally parked, the system sends an illegal parking reminder to the user's terminal and provides suggestions for correction and information on nearby legal parking locations.

[0054] Preferably, the recommendation algorithm adopts a deep reinforcement learning-based algorithm, which specifically includes the following steps:

[0055] Initialize the network parameters of the deep reinforcement learning model, including the policy network and the value network;

[0056] The user's parking demand information and surrounding parking resource information are used as the input state of the model;

[0057] Based on the input state, a set of candidate actions for recommending parking locations is generated through a policy network;

[0058] Each candidate action is evaluated using a value network, and the corresponding action value is calculated.

[0059] The optimal recommended parking location is selected as the output action based on the action value.

[0060] By interacting with the environment, the system collects users' actual parking choices and parking experience feedback, which serves as a reward signal.

[0061] Based on the reward signal, a deep reinforcement learning algorithm is used to iteratively train the policy network and value network, continuously optimizing the performance of the recommendation algorithm and improving the accuracy of recommended parking locations and user satisfaction.

[0062] Preferably, it also includes a data security module, which performs the following operations:

[0063] The collected traffic big data is classified and managed hierarchically to determine the security level and protection requirements for different types of data;

[0064] Encryption algorithms are used to encrypt sensitive data to ensure the confidentiality and integrity of the data during transmission, storage and use;

[0065] Establish a data access control mechanism to authorize and restrict data access based on user identity and permissions, preventing unauthorized access and data leakage;

[0066] Monitor data security risks during system operation, including network attacks, malware intrusions, and abnormal data access;

[0067] When a data security risk is detected, the emergency response mechanism should be activated immediately, and corresponding protective measures should be taken, including blocking the source of the attack, isolating infected devices, and restoring data.

[0068] Regularly audit and assess data security, and adjust and optimize data security strategies in a timely manner based on the audit results.

[0069] The present invention has achieved at least the following beneficial effects:

[0070] 1. This invention enables intelligent and refined management of urban traffic, effectively solving problems such as incomplete monitoring, inaccurate judgment, and low enforcement efficiency in traditional illegal parking management. This improves urban traffic order, alleviates traffic congestion, ensures smooth roads and traffic safety, and creates a better travel environment for citizens. At the same time, it provides decision-making basis for urban traffic management departments and assists in urban traffic planning and optimization.

[0071] 2. By leveraging a spatiotemporal attention mechanism, the correlation between traffic data in both time and space dimensions is fully explored, resulting in more comprehensive and accurate feature representations. The introduction of a resource game computing network considers the supply and demand relationship of parking spaces and the competition for alternative parking resources in the surrounding area. This allows for a more reasonable assessment of the risk of illegal parking, enabling proactive measures such as optimizing parking resource allocation and strengthening law enforcement patrols to effectively prevent illegal parking, improve the intelligence level of urban traffic management, and ensure smooth traffic flow and traffic safety.

[0072] Other advantages, objectives, and features of the invention will be set forth in the following description and will be apparent to those skilled in the art in some respects, or may be learned by practice of the invention. The objectives and other advantages of the invention can be realized and obtained through the following description. Attached Figure Description

[0073] To make the objectives, technical solutions, and beneficial effects of this invention clearer, the following figures are provided for illustration:

[0074] Figure 1 This is a schematic diagram of the structure of an illegal parking management system based on Internet of Things big data in an embodiment of the present invention;

[0075] Figure 2 This is a flowchart illustrating the execution steps of the illegal parking risk prediction module in an embodiment of the present invention.

[0076] Figure 3 This is a flowchart illustrating the execution steps of the user interaction module in an embodiment of the present invention. Detailed Implementation

[0077] The preferred embodiments of the present invention will be described below with reference to the accompanying drawings. It should be understood that the preferred embodiments described herein are for illustration and explanation only and are not intended to limit the present invention.

[0078] This invention provides a parking violation management system based on Internet of Things big data, with reference to... Figure 1 ,include:

[0079] The Internet of Things (IoT) sensing module is used to integrate various IoT devices in the city and collect traffic big data.

[0080] The data integration module is used to establish mapping relationships between data based on the spatiotemporal relationships of the collected data, and to integrate the mapped data to generate a traffic data heat map;

[0081] The illegal parking detection module is used to compare the number and location of parked vehicles and the total number and location of available parking spaces in any area based on traffic data heatmaps and continuously collected traffic big data to detect illegally parked vehicles.

[0082] The illegal parking management module is used to provide guidance or warning information to illegally parked vehicles, record illegal parking data, and provide feedback to law enforcement.

[0083] The working principle and beneficial effects of the above technical solution are as follows: By integrating various IoT devices within the city, such as millimeter-wave radar, road surveillance cameras, and geomagnetic sensors, through an IoT sensing module, comprehensive collection of traffic big data is achieved, covering multi-dimensional information such as vehicle location, speed, and direction of travel. This provides a rich and accurate data source for subsequent data processing. Taking millimeter-wave radar as an example, it transmits millimeter-wave signals and receives reflected waves. Based on the frequency changes and time differences of the signals, it accurately calculates the vehicle's location and speed information with centimeter-level accuracy, thus achieving real-time and accurate monitoring of vehicle driving status. Then, the data integration module, based on the spatiotemporal relationship of the collected data, uses specific algorithms to establish mapping relationships between the data, integrating the scattered data to generate an intuitive traffic data heatmap. This heatmap visually presents the traffic flow and vehicle distribution within a region using different colors or densities; for example, darker colors indicate higher vehicle density, making the traffic situation immediately clear. Based on this heat map and continuously collected traffic big data, the illegal parking detection module accurately compares the number and location of parked vehicles in any area with the total number and location of available parking spaces. Using preset parking rules and algorithm models, it quickly determines whether there are illegally parked vehicles. For example, if a vehicle is parked in a non-preset parking area and stays there for more than a specified time, the system determines it to be illegally parked. Once illegal parking is detected, the illegal parking management module immediately sends guidance or warning messages to the illegally parked vehicle via SMS, in-vehicle terminals, or mobile applications to guide the driver to move the vehicle or park it properly. At the same time, the module records illegal parking data in detail, including the time, location, and license plate number of the illegal parking, and promptly reports this data to the enforcement department, providing strong data support for traffic enforcement, improving enforcement efficiency and accuracy, and reducing the workload and resource consumption of on-site patrols. This technical solution enables intelligent and refined management of urban traffic, effectively addressing issues such as incomplete monitoring, inaccurate judgment, and low enforcement efficiency in traditional illegal parking management. This improves urban traffic order, alleviates traffic congestion, ensures smooth traffic flow and safety, and creates a better travel environment for citizens. At the same time, it provides decision-making support for urban traffic management departments, assisting in urban traffic planning and optimization.

[0084] In a preferred embodiment, the IoT sensing module includes:

[0085] The multi-source device access unit is used to integrate multiple types of device data sources through a heterogeneous protocol conversion gateway. These devices include millimeter-wave radar, road monitoring cameras, geomagnetic sensors, vehicle-mounted GPS positioning devices, electronic fences, and Bluetooth sniffing devices.

[0086] The data spatiotemporal alignment unit is used to synchronize the clock errors of each device to the millisecond level using the NTP protocol, establish a unified spatial reference under the WGS-84 coordinate system, and map the vehicle position to the lane lines of the centimeter-level high-precision map by fusing multi-source positioning data through Kalman filtering.

[0087] The data verification and alignment unit is used to perform reverse correlation between vehicle location and corresponding data for data with spatiotemporal attributes based on the analysis results of vehicle location by the data spatiotemporal alignment unit, establish the mapping relationship between vehicle location, associated data, and time, and determine the data status, location status of each device, and location status of the vehicle at each judgment time point.

[0088] The working principle and beneficial effects of the above technical solution are as follows: The multi-source device access unit integrates various device data sources, including millimeter-wave radar, road monitoring cameras, geomagnetic sensors, vehicle-mounted GPS positioning devices, electronic fences, and Bluetooth sniffing devices, through a heterogeneous protocol conversion gateway. This enables comprehensive collection of multi-dimensional urban traffic data, providing a rich and accurate data foundation for subsequent data processing. Taking millimeter-wave radar as an example, it transmits millimeter-wave signals and receives reflected waves. Based on the frequency changes and time differences of the signals, it accurately calculates the vehicle's position and speed information with centimeter-level accuracy, thus achieving real-time and accurate monitoring of vehicle driving status. Then, the data spatiotemporal alignment unit uses the NTP protocol to synchronize the clock errors of each device to the millisecond level, ensuring high consistency of data in the time dimension. Simultaneously, a unified spatial reference is established under the WGS-84 coordinate system, providing accurate coordinate references for spatial data fusion. Kalman filtering is then used to fuse multi-source positioning data, effectively filtering out noise interference and further improving the accuracy of vehicle position information. Finally, the vehicle position is accurately mapped to centimeter-level high-precision map lane lines, making the vehicle position information more intuitive and accurate on the map. Based on this, the data verification and alignment unit, according to the analysis results of the data spatiotemporal alignment unit on vehicle location, performs reverse correlation between vehicle location and corresponding data for data with spatiotemporal attributes. That is, it matches vehicle location information with data collected by various associated devices, thereby establishing a mapping relationship between vehicle location, associated data, and time. This allows for accurate determination of the data status, location status, and vehicle location status of each device at each judgment time point, providing an accurate and reliable basis for subsequent illegal parking judgments. For example, by comparing vehicle location with parking space location information and combining time status, it accurately determines whether a vehicle is parked in a legal parking space. This technical solution enables efficient integration and accurate processing of urban traffic data, effectively solving problems such as incomplete data collection and inaccurate data fusion in traditional illegal parking management systems. This improves the efficiency and accuracy of illegal parking management, ensures urban traffic order, and provides strong data support and technical assurance for intelligent urban traffic management.

[0089] In a preferred embodiment, the data integration module performs the following operations:

[0090] The preset target area is divided into multiple monitoring units according to blocks, and the location boundaries of each monitoring unit are determined.

[0091] Determine the current location of the data source device or the location of the monitored area, and determine the monitoring unit to which the data source device belongs;

[0092] Based on the multiple data source devices to which the monitoring unit belongs, acquire various types of traffic data generated by them;

[0093] Based on the vehicle location status determined by various traffic data, a mapping relationship between data and vehicles is established in different times and spaces, and data that has not been associated with any determined vehicle is filtered out.

[0094] Based on the vehicles currently stored in the monitoring unit and the data that establishes a mapping relationship between the monitoring unit and each stored vehicle, the attribute tensor of the multidimensional data associated with the monitoring unit is defined. The attribute tensor includes vehicle density, vehicle dwell time, vehicle location, road load, and parking space status.

[0095] Based on the multidimensional data and attribute tensors corresponding to each monitoring unit, a traffic data heatmap for the region is generated.

[0096] The working principle and beneficial effects of the above technical solution are as follows: By dividing the preset target area into multiple monitoring units according to blocks and determining the location boundaries of each monitoring unit, refined management of urban traffic can be achieved. Large-scale traffic data is broken down into specific block units, making subsequent data processing more targeted and operable, providing a foundation for traffic condition analysis and illegal parking judgment. Determining the current location of the data source equipment or the location of the monitored area, and clarifying its associated monitoring unit, ensures accurate data classification and effective correlation. This ensures that the data within each monitoring unit originates from the equipment actually monitoring that area, avoiding data confusion and errors. Taking road surveillance cameras as an example, after clarifying their associated monitoring unit, the vehicle location, speed, and other data they collect can accurately reflect the traffic conditions of that specific block. Based on the multiple data source devices to which the monitoring unit belongs, various traffic data are acquired, including vehicle flow, driving speed, and parking duration. These data complement and corroborate each other, improving the reliability and completeness of the data. Based on vehicle location status determined by various traffic data, a mapping relationship between data and vehicles is established in both time and space to achieve precise correlation between data and vehicles. This enables real-time tracking of vehicle dynamics and understanding of location changes across different time and space dimensions. For illegal parking management, this process can accurately identify whether vehicles are parked in legal locations and whether parking time exceeds limits, among other key information. Simultaneously, data not associated with the identified vehicles is filtered out, effectively cleaning up irrelevant data, reducing data processing burden, improving system efficiency, and ensuring that subsequent analysis is based on valuable data. Based on the vehicle and associated data retained within the monitoring unit, multi-dimensional data attribute tensors are associated with the monitoring unit, including vehicle density, vehicle dwell time, vehicle location, road load, and parking space status. This further deepens the understanding of traffic conditions in each monitoring unit. Taking vehicle density and road load as examples, they can intuitively reflect the degree of road congestion, providing a basis for traffic management and illegal parking management decisions. Finally, based on the multidimensional data and attribute tensors of each monitoring unit, a traffic data heat map of the region is generated. Through intuitive visualization, the complex traffic data is transformed into easily understandable graphical information, enabling traffic managers to grasp the traffic conditions of each area at a glance, promptly identify areas with high rates of illegal parking and traffic congestion, thereby achieving efficient and accurate traffic management and illegal parking control, improving urban traffic efficiency, and ensuring smooth roads and traffic safety.

[0097] In a preferred embodiment, the illegal parking detection module performs the following operations:

[0098] Based on traffic data heatmaps and continuously collected traffic big data, the number of parked vehicles, their location, and duration of stay are tracked and determined in any given area.

[0099] Based on traffic big data, the total number of reserved parking spaces in the area and the spatial distribution of available parking locations are determined.

[0100] Based on the tracking and judgment results of the number of parked vehicles, vehicle parking locations, and vehicle parking duration within the area, the occupancy status and occupancy duration of available parking locations are determined, and vehicles not parked at the designated locations and their corresponding parking durations are also determined.

[0101] Based on the duration of the parked vehicle's stay and the preset parking management policy corresponding to the vehicle's location, a multi-level management policy for the parked vehicle is generated and sent to the illegal parking management module.

[0102] The working principle and beneficial effects of the above technical solution are as follows: The illegal parking judgment module analyzes traffic data heatmaps and continuously collected traffic big data. First, it tracks and judges the number of parked vehicles, their parking locations, and parking durations within any given area, thereby gaining real-time insight into vehicle parking conditions. Taking a commercial street as an example, the system can accurately identify the parking location and duration of each vehicle within that area, providing basic data for subsequent judgments. Next, based on traffic big data, it determines the total number of reserved parking spaces and the spatial distribution of available parking locations within the area. This allows the system to clearly understand the legal parking capacity and specific parking locations within the area, such as knowing that a certain street has 100 legal parking spaces and their specific locations. Then, based on the aforementioned tracking and judgment results, it determines the occupancy status and duration of available parking locations, while also identifying vehicles not parked at designated locations and their corresponding parking durations. In this way, the system can accurately distinguish between legally parked and illegally parked vehicles; for example, it can identify 10 vehicles parked in non-designated locations for longer than the permitted time. Finally, based on the duration of parking and the preset parking management strategy corresponding to the vehicle's location, a multi-level management strategy is generated for the parked vehicle and sent to the illegal parking management module. For example, a countdown warning message is sent to vehicles parked for less than 30 minutes, while illegal parking data is recorded and fed back to the enforcement terminal for vehicles parked for more than 30 minutes, thus achieving hierarchical and categorized management of illegal parking behavior. This technical solution enables precise monitoring and efficient management of vehicle parking, effectively solving the problems of difficulty in accurately grasping vehicle parking information in real time and failing to detect illegal parking behavior in a timely manner in traditional illegal parking management. This improves the efficiency and accuracy of illegal parking management, ensures urban traffic order, and the implementation of multi-level management strategies also helps to improve the flexibility and humanization of management, reduce unnecessary enforcement conflicts, and improve the overall level of urban traffic management.

[0103] In a preferred embodiment, refer to Figure 2 It also includes a parking violation risk prediction module, which performs the following operations:

[0104] For the data that has been integrated and mapped, feature vectors related to vehicles, time, and space are extracted, and feature vectors suitable for model input are constructed.

[0105] An improved spatiotemporal graph convolutional network, STGCN, is constructed, and a spatiotemporal attention mechanism is applied to process the input feature vector. In the spatiotemporal graph convolutional network:

[0106] For each IoT device, the correlation between the current time step and the previous time step is calculated using its historical data sequence to obtain the time attention weight. The time attention weight is used to characterize the importance of data at different time points to the current prediction.

[0107] Calculate the spatial correlation between each IoT device, taking into account the topology and geographical adjacency of the road network, and derive spatial attention weights to measure the influence of data from different locations on the prediction of the current location.

[0108] The input features are weighted and fused using the calculated temporal and spatial attention weights;

[0109] The weighted and fused features are input into the resource game computation network, which then performs the following operations:

[0110] Obtain the parking space supply-demand ratio and ratings of surrounding alternative parking resources in each area;

[0111] Resource game weights are calculated based on the parking space supply-demand ratio and the scores of surrounding alternative resources to obtain resource game weights between different areas. These resource game weights reflect the parking resource competition relationship between the area and surrounding alternative parking resources.

[0112] The resource game computational network takes the weighted and fused feature vector as input and concatenates it with multiple resource game weights to obtain a comprehensive feature vector;

[0113] The comprehensive feature vector is passed through a fully connected layer and output. Using a positive activation or a constrained activation function, a risk score that changes continuously with time is output as the prediction result of the illegal parking risk in the target area.

[0114] The working principle and beneficial effects of the above technical solution are as follows: For data that has been integrated and mapped, feature vectors related to vehicles, time, and space are extracted to construct feature vectors suitable for model input. For example, vehicle features include vehicle type, license plate number, driving history, etc.; time features include current time, time period, etc.; and spatial features include vehicle location coordinates, region, etc. A spatiotemporal attention mechanism is applied to process the input feature vectors.

[0115] For each IoT device, the correlation between the current time step and previous time steps is calculated using its historical data sequence to obtain the time attention weight. The calculation formula is as follows:

[0116] Time attention weight

[0117] in, This represents the query vector, which is generated from the features of the current time step and is used to capture the features of historical moments of interest at the current time step. This represents the key vector, which is generated from features of previous time steps and used to match the query vector to determine the relevance of the current time step to previous time steps.

[0118] The spatial correlation between each IoT device is calculated, taking into account the topology and geographical adjacency of the road network, to derive the spatial attention weight. The calculation formula is as follows:

[0119] Spatial attention weights

[0120] in, This represents the spatial feature vector of device i, which contains information such as the location of device i and road connectivity. represents the spatial feature vector of device j, used for spatial correlation calculation with device i; N represents the total number of IoT devices.

[0121] The calculated temporal and spatial attention weights are used to weight and fuse the input features, resulting in a fused feature vector. The formula is:

[0122] Fusion feature vectors

[0123] in, This represents the characteristics of device i at a previous time step. This represents the characteristics of device j at the current time step. This represents the fused feature vector of device i at the current time step. In this way, features in the time and space dimensions are fused to obtain a more comprehensive feature representation.

[0124] Obtain the parking space supply-demand ratio for each area ( ) and ratings of nearby alternative parking resources ( Based on the parking space supply-demand ratio and the scores of surrounding alternative resources, resource game weights are calculated to obtain the resource game weights between different areas. Resource game weights reflect the competition for parking resources between this area and surrounding alternative parking resources, and their calculation formula is as follows:

[0125] Resource game weight

[0126] in, The parking space supply-demand ratio represents the ratio of the supply to the demand for parking spaces in region r. This represents the sum of scores for surrounding alternative parking resources. The scores for alternative parking resources are calculated based on factors such as distance, size, and pricing.

[0127] The weighted and fused feature vector is used as input and concatenated with multiple resource game weights to obtain a comprehensive feature vector. This comprehensive feature vector is then passed through a fully connected layer and output using a positive or constrained activation function to generate a continuously varying risk score as the prediction result for the illegal parking risk in the target area. The activation function can be the Sigmoid function, whose formula is:

[0128] Output value

[0129] Where z is the result of the weighted sum of the input feature vector and weights plus the bias term and other operations. This function maps the output value to the (0,1) interval, representing the probability of illegal parking.

[0130] The illegal parking risk prediction module enables accurate prediction of illegal parking risks in target areas. On one hand, it fully leverages a spatiotemporal attention mechanism to exploit the correlations between traffic data in time and space, resulting in more comprehensive and accurate feature representations. On the other hand, it introduces a resource game theory computation network, considering the supply and demand relationship of parking spaces and the competition for alternative parking resources in the surrounding area. This allows for a more reasonable assessment of illegal parking risks, enabling proactive measures such as optimizing parking resource allocation and strengthening law enforcement patrols to effectively prevent illegal parking, improve the intelligence level of urban traffic management, and ensure smooth traffic flow and traffic safety.

[0131] In a preferred embodiment, the scoring calculation process for nearby alternative parking resources includes:

[0132] Identify multiple parking areas within a preset range around the target area as alternative parking resources in the vicinity;

[0133] Determine the distance between any alternative parking resource and the target area, the size of the parking lot, the charging standard, and user reviews, and assign weights to each of them. ,and ;

[0134] Collect relevant data on this alternative parking resource, including distance d, number of remaining parking spaces s, charging standard c, and user rating a, and normalize the data so that its value ranges between 0 and 1.

[0135] The distance is normalized to obtain distance-normalized data. : ;

[0136] The remaining parking space quantity is normalized to obtain normalized data. : ;

[0137] The fee standard is normalized to obtain the fee standard normalized data. : ;

[0138] Normalize the user reviews to obtain normalized user review data. The scale is adjusted to keep the score within the range of 0 to 5.

[0139] Its overall score is calculated using a weighted formula: ;

[0140] The overall score will be used as the evaluation result for this alternative parking resource.

[0141] The working principle and beneficial effects of the above technical solution are as follows: Multiple parking areas within a preset range surrounding the target area are identified as alternative parking resources. These parking areas are geographically related to and substitutable with the target area. For example, all parking lots within a 3-kilometer radius of the target area are selected as alternative parking resources.

[0142] Determine key factors such as the distance of any alternative parking resource from the target area, parking lot size, pricing, and user reviews, and assign weights to each. Assume the weights are: distance 0.3, remaining parking spaces 0.2, pricing 0.25, and user reviews 0.25. Collect relevant data for the alternative parking resource, including distance *d*, remaining parking spaces *s*, pricing *c*, and user reviews *a*, and normalize the data to a value between 0 and 1. Normalize the distance to obtain distance-normalized data. For example, assume the distance range of all alternative parking resources in the target area is 1 km to 5 km. If an alternative parking resource is 3 km from the target area, the distance-normalized data is calculated as: (3 km - 1 km) / (5 km - 1 km) = 0.5. Normalize the remaining parking spaces to obtain remaining parking space-normalized data. For example, assume the remaining parking spaces of all alternative parking resources in the target area range from 10 to 100. If a certain alternative parking resource has 50 remaining parking spaces, the normalized calculation for the number of remaining spaces is: (50 - 10) / (100 - 10) ≈ 0.5. The pricing standard is then normalized to obtain normalized pricing data. For example, assuming the pricing standard for all alternative parking resources in the target area ranges from 2 yuan / hour to 10 yuan / hour, and the pricing standard for a certain alternative parking resource is 6 yuan / hour, the normalized pricing data is: (6 yuan / hour - 2 yuan / hour) / (10 yuan / hour - 2 yuan / hour) ≈ 0.5. User reviews are then normalized to obtain normalized user review data, and the scale is scaled to ensure the rating is within the range of 0-5 points.

[0143] For example, suppose the original user rating score for an alternative parking resource is 4 points, which is also 4 points after normalization. Its comprehensive score is calculated using a weighted formula. For instance, if the distance normalization data for an alternative parking resource is 0.5, the remaining parking space normalization data is 0.5, the fee normalization data is 0.5, and the user rating normalization data is 4 points (which needs to be converted to a value between 0 and 1, i.e., 4 / 5 = 0.8), then the comprehensive score is calculated as: 0.3 × 0.5 + 0.2 × 0.5 + 0.25 × 0.5 + 0.25 × 0.8 = 0.15 + 0.1 + 0.125 + 0.2 = 0.575.

[0144] The comprehensive score is used as the evaluation result for each alternative parking resource. This gives each surrounding alternative parking resource a comprehensive score, which can be used for subsequent illegal parking risk prediction and analysis. This technical solution enables a comprehensive and objective evaluation of the quality of surrounding alternative parking resources. The weighting of each factor reflects the importance of different factors to the attractiveness of parking resources. Distance has a higher weight because proximity directly affects users' willingness to choose parking resources. The number of remaining parking spaces, pricing standards, and user reviews are also evaluated from the perspectives of capacity, economy, and user experience. This scoring mechanism helps the illegal parking risk prediction module more accurately assess the illegal parking risk in the target area, thereby taking measures in advance to optimize parking resource allocation and reduce the occurrence of illegal parking.

[0145] In a preferred embodiment, refer to Figure 3 It also includes a user interaction module, which performs the following operations:

[0146] Receive parking request information sent by users, including vehicle information, estimated arrival time, destination location, and parking duration requirement;

[0147] Based on the destination location, parking resource information from multiple surrounding monitoring units is extracted from the traffic data heat map. The parking resource information includes the number of available parking spaces, their location distribution, parking fees, and parking time restrictions.

[0148] Based on parking demand and parking resource information, the system uses a pre-set recommendation algorithm to recommend the best parking location for users and generate parking navigation information that includes driving routes, estimated parking fees, and estimated parking duration.

[0149] The system sends parking navigation information to the user's terminal and updates the navigation information in real time to guide the vehicle to the recommended parking location.

[0150] After the vehicle arrives at the recommended parking location, the system monitors the vehicle's parking status. If the vehicle is illegally parked, the system sends an illegal parking reminder to the user's terminal and provides suggestions for correction and information on nearby legal parking locations.

[0151] The working principle and beneficial effects of the above technical solution are as follows: The system receives parking request information sent by the user, including vehicle information, estimated arrival time, destination location, and parking duration requirement. For example, the user sends information containing license plate number, estimated arrival time, destination location, and parking duration requirement to the system via a mobile application; the system receives and stores this information.

[0152] Based on the destination location, the system extracts parking resource information from multiple surrounding monitoring units in the traffic data heatmap, including the number of available parking spaces, their location distribution, parking fees, and parking time restrictions. For example, based on the destination location, the system filters parking resource information from multiple monitoring units within a 3-kilometer radius of the traffic data heatmap. This information is stored in a database for subsequent processing.

[0153] Based on parking demand and resource information, a pre-defined recommendation algorithm is used to recommend optimal parking locations to users and generate parking navigation information that includes driving routes, estimated parking costs, and estimated parking duration. For example, the recommendation algorithm filters out suitable parking resources based on the user's parking duration requirements and the parking time limits of available parking resources. It then selects the optimal parking location based on factors such as distance and parking costs. Simultaneously, it calculates the driving route from the user's current location to the recommended parking location, estimates the parking cost, determines the estimated parking duration, and integrates all this information into parking navigation information.

[0154] The system sends parking navigation information to the user's terminal and updates the navigation information in real time to guide the vehicle to the recommended parking location. For example, the system sends parking navigation information to the user's mobile application or in-vehicle navigation system via a wireless communication module, and updates the navigation information based on real-time traffic data while the vehicle is in motion, ensuring that the user can successfully reach the recommended parking location.

[0155] After the vehicle arrives at the recommended parking location, its parking status is monitored. If illegal parking is detected, a violation alert is sent to the user's terminal, along with suggestions for correction and information on nearby legal parking locations. For example, the system uses an IoT sensing module to monitor whether the vehicle is parked in the recommended parking location. If the vehicle is detected parked in a non-recommended location, it is determined to be illegal parking, and a violation alert is immediately sent to the user's terminal, reminding the user to correct the parking violation and providing information on nearby legal parking locations to guide the user to move the vehicle to a legal location.

[0156] This technical solution enables efficient interaction between the user interaction module and users, enhancing the user experience. Users can obtain accurate parking resource information and personalized parking navigation services in a timely manner, reducing the time and effort spent searching for parking spaces. The recommendation algorithm recommends optimal parking locations based on users' parking needs and real-time parking resource information, improving the utilization efficiency of parking resources and reducing traffic congestion caused by parking difficulties. Real-time updated navigation information ensures users can smoothly reach the recommended parking location, improving the accuracy and reliability of navigation. Simultaneously, monitoring vehicle parking status and promptly reminding users to correct illegal parking behavior helps regulate parking order, reduce illegal parking, and improve the intelligence level of urban traffic management.

[0157] In a preferred embodiment, the recommendation algorithm employs a deep reinforcement learning-based algorithm, specifically including the following steps:

[0158] Initialize the network parameters of the deep reinforcement learning model, including the policy network and the value network;

[0159] The user's parking demand information and surrounding parking resource information are used as the input state of the model;

[0160] Based on the input state, a set of candidate actions for recommending parking locations is generated through a policy network;

[0161] Each candidate action is evaluated using a value network, and the corresponding action value is calculated.

[0162] The optimal recommended parking location is selected as the output action based on the action value.

[0163] By interacting with the environment, the system collects users' actual parking choices and parking experience feedback, which serves as a reward signal.

[0164] Based on the reward signal, a deep reinforcement learning algorithm is used to iteratively train the policy network and value network, continuously optimizing the performance of the recommendation algorithm and improving the accuracy of recommended parking locations and user satisfaction.

[0165] The working principle and beneficial effects of the above technical solution are as follows: Initialize the network parameters of the deep reinforcement learning model, including a policy network and a value network. The policy network is used to generate a set of candidate actions for recommending parking locations, and the value network is used to evaluate the value of each candidate action. The user's parking demand information (such as vehicle information, estimated arrival time, destination location, parking duration requirements, etc.) and surrounding parking resource information (such as the number of available parking spaces, their distribution, parking fees, parking time limits, etc.) are used as the model's input state. Based on the input state, the policy network generates a set of candidate actions for recommending parking locations. For example, the policy network may output several possible parking locations as candidate actions. The value network evaluates each candidate action and calculates the corresponding action value. The action value reflects the expected effect of the parking location recommendation, such as the probability of the user choosing that location, the convenience of parking, etc. The optimal recommended parking location is selected as the output action based on the action value. The parking location with the highest action value is selected as the recommendation result. Through interaction with the environment, the user's actual parking choices and parking experience feedback are collected as reward signals. The purpose of environmental interaction is to obtain real user feedback, enabling the model to learn and adjust based on actual results. For example, if a user accepts the recommendation and successfully parks, and has a good parking experience (such as convenient parking location and reasonable cost), the model is given a positive reward; conversely, if the user does not accept the recommendation or has a poor parking experience, a negative reward is given. Based on the reward signal, a deep reinforcement learning algorithm is used to iteratively train the policy network and value network, continuously optimizing the performance of the recommendation algorithm and improving the accuracy of recommended parking locations and user satisfaction.

[0166] This technical solution enables personalized and intelligent parking location recommendations through its recommendation algorithm. The deep reinforcement learning model learns users' parking preferences and behavioral patterns by continuously interacting with the environment, thus providing recommendations that better meet user needs. Environmental interaction provides realistic feedback, allowing the model to dynamically adjust its recommendation strategy, improving accuracy and user satisfaction. Simultaneously, with continuous training and optimization, the algorithm's performance is continuously improved, providing users with accurate and efficient parking navigation services, reducing the time and effort spent searching for parking spaces, and increasing the utilization efficiency of urban parking resources.

[0167] In a preferred embodiment, a data security module is also included, which performs the following operations:

[0168] The collected traffic big data is classified and managed hierarchically to determine the security level and protection requirements for different types of data;

[0169] Encryption algorithms are used to encrypt sensitive data to ensure the confidentiality and integrity of the data during transmission, storage and use;

[0170] Establish a data access control mechanism to authorize and restrict data access based on user identity and permissions, preventing unauthorized access and data leakage;

[0171] Monitor data security risks during system operation, including network attacks, malware intrusions, and abnormal data access;

[0172] When a data security risk is detected, the emergency response mechanism should be activated immediately, and corresponding protective measures should be taken, including blocking the source of the attack, isolating infected devices, and restoring data.

[0173] Regularly audit and assess data security, and adjust and optimize data security strategies in a timely manner based on the audit results.

[0174] The working principle and beneficial effects of the above technical solution are as follows: The collected traffic big data is classified and graded for management, determining the security level and protection requirements for different types of data. For example, sensitive data such as user personal information and vehicle driving records are classified as high-security data, while some publicly available traffic flow statistics are classified as low-security data. Encryption algorithms are used to encrypt sensitive data, ensuring the confidentiality and integrity of data during transmission, storage, and use. For example, the Advanced Encryption Standard (AES) algorithm is used to encrypt user identity information, and Secure Sockets Layer (SSL) / Transport Layer Security (TLS) protocols are used for encrypted transmission. A data access control mechanism is established to authorize and restrict data access based on user identity and permissions, preventing unauthorized access and data leakage. For example, through a Role-Based Access Control (RBAC) model, different data access permissions are assigned to different user roles; administrators can access and manage all data, while ordinary users can only access data related to themselves, such as parking records. Data security risks during system operation are monitored, including network attacks, malware intrusions, and abnormal data access. For example, deploying an Intrusion Detection System (IDS) and a Security Information and Event Management System (SIEM) allows for real-time monitoring of network traffic and system logs to identify potential security threats. When a data security risk is detected, an emergency response mechanism is immediately activated, and corresponding protective measures are taken, including blocking the attack source, isolating infected devices, and restoring data. For example, if malware intrusion is detected, infected devices are immediately isolated to prevent further virus spread, and data is restored using backups. Regular audits and assessments of data security are conducted, and data security strategies are adjusted and optimized promptly based on the audit results. For example, a data security audit is conducted quarterly to check the effectiveness of data access controls, whether encryption algorithms are outdated, etc., and security strategies are updated based on the audit results. Through this technical solution, the data security module can comprehensively ensure the security of traffic big data. Data classification and hierarchical management helps clarify the protection priorities of different data and rationally allocate security resources. Encryption ensures that sensitive data is not affected by unauthorized acquisition and tampering at any stage. Data access control mechanisms effectively prevent unauthorized access, protecting user privacy and data confidentiality. Real-time monitoring of data security risks enables timely detection and early warning in the early stages of a threat. Emergency response mechanisms enable rapid handling of security incidents, minimizing losses and ensuring the continuous and stable operation of the system. Regular security audits and assessments help to continuously improve data security strategies, adapt to the ever-changing security environment, and provide solid data security guarantees for the stable operation of the urban traffic management system.

[0175] Finally, it should be noted that the above preferred 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 through the above preferred embodiments, those skilled in the art should understand that various changes can be made to it in form and detail without departing from the scope defined by the claims of the present invention.

Claims

1. A parking violation management system based on Internet of Things big data, characterized in that, include: The Internet of Things (IoT) sensing module is used to integrate various IoT devices in the city and collect traffic big data. The data integration module is used to establish mapping relationships between data based on the spatiotemporal relationships of the collected data, and to integrate the mapped data to generate a traffic data heat map; The illegal parking detection module is used to compare the number and location of parked vehicles and the total number and location of available parking spaces in any area based on traffic data heatmaps and continuously collected traffic big data to detect illegally parked vehicles. The illegal parking management module is used to provide guidance or warning information to illegally parked vehicles, record illegal parking data, and provide feedback to law enforcement. The data integration module performs the following operations: The preset target area is divided into multiple monitoring units according to blocks, and the location boundaries of each monitoring unit are determined. Determine the current location of the data source device or the location of the monitored area, and determine the monitoring unit to which the data source device belongs; Based on the multiple data source devices to which the monitoring unit belongs, acquire various types of traffic data generated by them; Based on the vehicle location status determined by various traffic data, a mapping relationship between data and vehicles is established in different times and spaces, and data that has not been associated with any determined vehicle is filtered out. Based on the vehicles currently stored in the monitoring unit and the data that establishes a mapping relationship between the monitoring unit and each stored vehicle, the attribute tensor of the multidimensional data associated with the monitoring unit is defined. The attribute tensor includes vehicle density, vehicle dwell time, vehicle location, road load, and parking space status. Based on the multidimensional data and attribute tensors corresponding to each monitoring unit, a traffic data heat map of the region is generated. The illegal parking risk prediction module performs the following operations: For the data that has been integrated and mapped, feature vectors related to vehicles, time, and space are extracted, and feature vectors suitable for model input are constructed. An improved spatiotemporal graph convolutional network, STGCN, is constructed, and a spatiotemporal attention mechanism is applied to process the input feature vector. In the spatiotemporal graph convolutional network: For each IoT device, the correlation between the current time step and the previous time step is calculated using its historical data sequence to obtain the time attention weight. The time attention weight is used to characterize the importance of data at different time points to the current prediction. Calculate the spatial correlation between each IoT device, taking into account the topology and geographical adjacency of the road network, and derive spatial attention weights to measure the influence of data from different locations on the prediction of the current location. The input features are weighted and fused using the calculated temporal and spatial attention weights; The weighted and fused features are input into the resource game computation network, which then performs the following operations: Obtain the parking space supply-demand ratio and ratings of surrounding alternative parking resources in each area; Resource game weights are calculated based on the parking space supply-demand ratio and the scores of surrounding alternative resources to obtain resource game weights between different areas. These resource game weights reflect the parking resource competition relationship between the area and surrounding alternative parking resources. The resource game computational network takes the weighted and fused feature vector as input and concatenates it with multiple resource game weights to obtain a comprehensive feature vector; The comprehensive feature vector is passed through a fully connected layer and output. Using a positive activation or a constrained activation function, a risk score that changes continuously with time is output as the prediction result of the illegal parking risk in the target area.

2. The illegal parking management system based on Internet of Things big data according to claim 1, characterized in that, The IoT sensing module includes: The multi-source device access unit is used to integrate multiple types of device data sources through a heterogeneous protocol conversion gateway. These devices include millimeter-wave radar, road monitoring cameras, geomagnetic sensors, vehicle-mounted GPS positioning devices, electronic fences, and Bluetooth sniffing devices. The data spatiotemporal alignment unit is used to synchronize the clock errors of each device to the millisecond level using the NTP protocol, establish a unified spatial reference under the WGS-84 coordinate system, and map the vehicle position to the lane lines of the centimeter-level high-precision map by fusing multi-source positioning data through Kalman filtering. The data verification and alignment unit is used to perform reverse correlation between vehicle location and corresponding data for data with spatiotemporal attributes based on the analysis results of vehicle location by the data spatiotemporal alignment unit, establish the mapping relationship between vehicle location, associated data, and time, and determine the data status, location status of each device, and location status of the vehicle at each judgment time point.

3. The illegal parking management system based on Internet of Things big data according to claim 1, characterized in that, The illegal parking detection module performs the following operations: Based on traffic data heatmaps and continuously collected traffic big data, the number of parked vehicles, their location, and duration of stay are tracked and determined in any given area. Based on traffic big data, the total number of reserved parking spaces in the area and the spatial distribution of available parking locations are determined. Based on the tracking and judgment results of the number of parked vehicles, vehicle parking locations, and vehicle parking duration within the area, the occupancy status and occupancy duration of available parking locations are determined, and vehicles not parked at the designated locations and their corresponding parking durations are also determined. Based on the duration of the parked vehicle's stay and the preset parking management policy corresponding to the vehicle's location, a multi-level management policy for the parked vehicle is generated and sent to the illegal parking management module.

4. The illegal parking management system based on Internet of Things big data according to claim 1, characterized in that, The scoring calculation process for nearby alternative parking resources includes: Identify multiple parking areas within a preset range around the target area as alternative parking resources in the vicinity; Determine the distance between any alternative parking resource and the target area, the size of the parking lot, the charging standard, and user reviews, and assign weights to each of them. ,and ; Collect relevant data on this alternative parking resource, including distance d, number of remaining parking spaces s, charging standard c, and user rating a, and normalize the data so that its value ranges between 0 and 1. The distance is normalized to obtain distance-normalized data. : ; The remaining parking space quantity is normalized to obtain normalized data. : ; The fee standard is normalized to obtain the fee standard normalized data. : ; Normalize the user reviews to obtain normalized user review data. ; Its overall score is calculated using a weighted formula: ; The overall score will be used as the evaluation result for this alternative parking resource.

5. The illegal parking management system based on Internet of Things big data according to claim 1, characterized in that, It also includes a user interaction module, which performs the following operations: Receive parking request information sent by users, including vehicle information, estimated arrival time, destination location, and parking duration requirement; Based on the destination location, parking resource information from multiple surrounding monitoring units is extracted from the traffic data heat map. The parking resource information includes the number of available parking spaces, their location distribution, parking fees, and parking time restrictions. Based on parking demand and parking resource information, the system uses a pre-set recommendation algorithm to recommend the best parking location for users and generate parking navigation information that includes driving routes, estimated parking fees, and estimated parking duration. The system sends parking navigation information to the user's terminal and updates the navigation information in real time to guide the vehicle to the recommended parking location. After the vehicle arrives at the recommended parking location, the system monitors the vehicle's parking status. If the vehicle is illegally parked, the system sends an illegal parking reminder to the user's terminal and provides suggestions for correction and information on nearby legal parking locations.

6. The illegal parking management system based on Internet of Things big data according to claim 5, characterized in that, The recommendation algorithm uses a deep reinforcement learning-based approach, specifically including the following steps: Initialize the network parameters of the deep reinforcement learning model, including the policy network and the value network; The user's parking demand information and surrounding parking resource information are used as the input state of the model; Based on the input state, a set of candidate actions for recommending parking locations is generated through a policy network; Each candidate action is evaluated using a value network, and the corresponding action value is calculated. The optimal recommended parking location is selected as the output action based on the action value. By interacting with the environment, the system collects users' actual parking choices and parking experience feedback, which serves as a reward signal. Based on the reward signal, a deep reinforcement learning algorithm is used to iteratively train the policy network and value network, continuously optimizing the performance of the recommendation algorithm and improving the accuracy of recommended parking locations and user satisfaction.

7. The illegal parking management system based on Internet of Things big data according to claim 1, characterized in that, It also includes a data security module, which performs the following operations: The collected traffic big data is classified and managed hierarchically to determine the security level and protection requirements for different types of data; Encryption algorithms are used to encrypt sensitive data to ensure the confidentiality and integrity of the data during transmission, storage and use; Establish a data access control mechanism to authorize and restrict data access based on user identity and permissions, preventing unauthorized access and data leakage; Monitor data security risks during system operation, including network attacks, malware intrusions, and abnormal data access; When a data security risk is detected, the emergency response mechanism should be activated immediately, and corresponding protective measures should be taken, including blocking the source of the attack, isolating infected devices, and restoring data. Regularly audit and assess data security, and adjust and optimize data security strategies in a timely manner based on the audit results.

Images