Intelligent parking lot traffic management system based on traffic internet of things

Abstract

The invention specifically relates to a smart parking lot traffic management system based on traffic Internet of Things, and relates to the technical field of smart traffic and Internet of Things. A data processing and fusion module; a congestion prediction module; a prediction optimization module; and an induction information generation and release module. According to the invention, through multi-dimensional data acquisition, fusion and model optimization, accurate pre-judgment and dynamic adaptation of parking lot congestion are realized; the congestion prediction module constructs a multi-dimensional feature system of stagnation degree, unit average delay degree and occupancy degree, covers core inducements of vehicle retention, parking efficiency, resource mismatch and the like, combines historical data optimization factors and k value dynamic adjustment, and enables the prediction model to continuously adapt to changes of parking lot vehicle types, user habits and the like. Passive response of congestion treatment is upgraded to active prevention, invalid wandering of vehicles and lane congestion are reduced, the parking lot passing efficiency and the parking space turnover rate are remarkably improved, and the operation cost is reduced.

Description

Technical Field

[0001] This invention relates to the fields of intelligent transportation and Internet of Things (IoT) technology, and in particular to an intelligent parking traffic management system based on the Internet of Transportation (IoT). Background Technology

[0002] With the rapid development of the automotive industry and the acceleration of urban motorization, the number of cars has exploded, and the demand for urban parking is increasing daily. As a key node in the urban transportation system, the operational efficiency of parking lots directly affects the smoothness of urban traffic and the travel experience of citizens. However, traditional parking lot management models rely on manual labor and simple equipment, which suffers from problems such as limited data collection dimensions and insufficient accuracy. They cannot fully perceive vehicle dynamics (such as accurate positioning, driving trajectory, and parking behavior) and environmental conditions (such as temperature, humidity, and facility operation), making it difficult to identify potential congestion risks in a timely manner.

[0003] Existing congestion prediction and mitigation solutions are mostly based on single-dimensional data (such as traffic flow) and do not delve into the core causes such as vehicle congestion, entry efficiency, and resource mismatch, resulting in limited prediction accuracy. Furthermore, the guidance information lacks specificity and is difficult to dynamically adjust according to the real-time status of vehicles (such as entry delays and parking space occupancy types), which not only fails to effectively alleviate congestion but also reduces the parking experience for car owners.

[0004] Therefore, building a smart parking traffic management system that integrates IoT sensing, intelligent data processing, accurate congestion prediction, and personalized guidance has become an inevitable choice to break through traditional management bottlenecks and adapt to the needs of modern urban traffic. Summary of the Invention

[0005] The purpose of this invention is to provide a smart parking traffic management system based on the Internet of Things for transportation in order to solve the above-mentioned problems.

[0006] To achieve the above objectives, the present invention adopts the following technical solution:

[0007] A smart parking traffic management system based on the Internet of Things for transportation includes:

[0008] The traffic data sensing and acquisition module is configured to collect dynamic vehicle data and static environmental data in the parking lot through IoT devices;

[0009] The data processing and fusion module is configured to perform in-depth processing on the collected raw data, eliminate noise and errors, and achieve effective fusion of multi-source data;

[0010] The congestion prediction module is configured to obtain corresponding features based on the processed data, and input each obtained feature into the prediction model to output the corresponding congestion prediction evaluation coefficient.

[0011] The prediction optimization module is configured to obtain optimization factors based on historical data and optimize the prediction model to output the optimized congestion prediction evaluation coefficient.

[0012] The guidance information generation and publishing module is configured to generate guidance information based on congestion prediction results to guide vehicles to divert.

[0013] Preferably, the traffic data sensing and acquisition module specifically includes:

[0014] Deploying IoT devices in parking lots to obtain vehicle location and environmental information;

[0015] The positioning tag sends signals to the base station at preset time intervals, and the base station calculates the vehicle coordinates and uploads them; the camera captures an image frame at preset time intervals, extracts the vehicle position and driving direction through image recognition technology, and calculates the driving speed by combining the continuous frame images; when the ground induction coil detects a vehicle, it outputs a pulse signal and records the time it takes for the vehicle to pass.

[0016] The guidance screen status monitoring chip provides real-time feedback on screen brightness, displayed content, and network connection status; the millimeter-wave radar outputs vehicle information within its detection range at preset intervals.

[0017] Location data is transmitted to a local data aggregation node and then uploaded via Ethernet; camera data is transmitted to the data center.

[0018] Preferably, the data processing and fusion module specifically includes:

[0019] Set a vehicle speed threshold, and remove a vehicle when its speed exceeds the threshold.

[0020] For location data, if the distance between two adjacent location coordinates exceeds a preset distance, it will be discarded;

[0021] When the camera cannot recognize the vehicle, the trajectory is completed using linear interpolation based on the vehicle's previous direction and speed. If the ground loop data is lost, the lost data is estimated by combining the detection time of the front and rear loops and the vehicle's speed.

[0022] The location coordinates and camera pixel coordinates are uniformly converted into the physical coordinates of the parking lot; the vehicle IDs collected by different devices are associated.

[0023] By integrating location data and camera trajectory data, and matching vehicle trajectory data with the electronic map of the parking lot, the vehicle's lane and distance from the intersection ahead can be determined.

[0024] Preferably, the congestion prediction module specifically includes:

[0025] The non-parking area of ​​the parking lot is divided into zones with a preset area to obtain various driving zones;

[0026] Determine the license plate numbers of each vehicle in the driving area and count the number of vehicles; after obtaining the license plate numbers of each vehicle in the driving area after a preset time interval, count the number of vehicles; compare the vehicles obtained in the two times to obtain the number of vehicles with the same license plate number, and divide the number of vehicles with the same license plate number by the maximum number of vehicles that the driving area can accommodate to obtain the stagnation rate.

[0027] A preset stagnation threshold is set, and the number of unoccupied parking spaces in the driving area corresponding to stagnation levels greater than the stagnation threshold is obtained. The overlapping area between vehicles and unoccupied parking spaces within a preset range of unoccupied parking spaces is obtained at preset time intervals. If the overlapping area increases in the time series, the corresponding vehicles are marked. The overlapping area is divided by the area of ​​unoccupied parking spaces to obtain the integrity. The difference between the integrity of adjacent time intervals is calculated, and the absolute value is taken to obtain the integrity difference. Each integrity difference is divided by the duration of the adjacent time interval to obtain the progress per unit time.

[0028] After obtaining the progress of k units of time, take their average value to obtain the average progress per unit. Then, take the difference between the average progress per unit and 1, and take the absolute value to obtain the average delay per unit.

[0029] Preferably, the method further includes:

[0030] A circular area is obtained by drawing a circle with the driving area as the center and a preset size as the radius. The number of fast charging piles and the number of parking spaces occupied by the fast charging piles are obtained within the circular area. The number of non-electric vehicles is extracted from the circle. The number of non-electric vehicles occupying the parking spaces corresponding to the fast charging piles is divided by the number of parking spaces occupied by the fast charging piles to obtain the occupancy rate.

[0031] Preferably, the stagnation degree, average delay per unit, and occupancy degree are respectively labeled as... , , The input prediction model is then calculated as follows:

[0032]

[0033] Obtain congestion prediction evaluation coefficient ;

[0034] in,

[0035] , , These are the reference stagnation rate, the maximum permissible unit average delay, and the reference standard occupancy rate, respectively.

[0036] , , These are the weighting factors for stagnation, average delay per unit, and occupancy, respectively.

[0037] Multiple threshold ranges are preset, and each threshold range is matched with an estimated congestion duration. The congestion prediction evaluation coefficient is matched with the ranges of multiple thresholds to obtain the corresponding estimated congestion duration.

[0038] Preferably, the prediction optimization module specifically includes:

[0039] Obtain historical parking data for a preset time period in the parking lot, as well as the corresponding measured congestion duration;

[0040] The optimization factor is obtained by dividing the estimated congestion duration by the actual congestion duration. The reciprocal of the optimization factor is then multiplied by the congestion prediction evaluation coefficient to obtain the optimized congestion prediction evaluation coefficient.

[0041] Preferably, the method further includes:

[0042] The average progress per unit is calculated based on the k-value; the progress error rate corresponding to the k-value is obtained, and then the k-value correction weight is defined;

[0043] Iterate through k=1,2,...,n, and calculate the corrected weight for each k to find the optimal k value.

[0044] Preferably, the inducement information generation and dissemination module specifically includes:

[0045] The guidance information includes the target area, estimated travel time, and available parking spaces; a dynamic route map is displayed on the guidance screen and congestion points along the route are updated in real time; and information is disseminated through parking guidance screens, in-vehicle apps / mini-programs, and broadcast systems.

[0046] In summary, due to the adoption of the above technical solution, the beneficial effects of the present invention are:

[0047] 1. This invention achieves accurate prediction and dynamic adaptation of parking lot congestion through multi-dimensional data collection, fusion, and model optimization. The congestion prediction module constructs a multi-dimensional feature system of stagnation degree, average delay per unit, and occupancy degree, covering core causes such as vehicle lingering, parking efficiency, and resource mismatch. Combined with historical data optimization factors and dynamic adjustment of the k-value, the prediction model continuously adapts to changes in parking lot vehicle types and user habits, upgrading congestion management from passive response to proactive prevention, reducing ineffective vehicle lingering and lane congestion, significantly improving parking lot traffic efficiency and parking space turnover rate, and reducing operating costs.

[0048] 2. This invention, through its guidance information generation and dissemination module, provides personalized guidance for users with different needs and in different scenarios based on accurate congestion prediction results. For ordinary vehicles, it accurately pushes information on target areas, travel time, and available parking spaces through dynamic route maps on guidance screens, real-time navigation via in-vehicle apps, and broadcast voice prompts, helping users quickly find available parking spaces and avoid blindly detouring. For vehicles that are slow to enter parking spaces, it pushes auxiliary information such as parking angle adjustment and remaining progress prompts, reducing the difficulty of parking operations. For electric vehicle users, it accurately identifies the vacancy status of fast charging stations and guides them to charge efficiently. For non-electric vehicle users who occupy charging stations, it pushes exclusive parking space exchange guidance to optimize resource allocation. Attached Figure Description

[0049] Further details, features, and advantages of this application are disclosed in the following description of exemplary embodiments in conjunction with the accompanying drawings, in which:

[0050] Figure 1 This is a system structure diagram of the present invention. Detailed Implementation

[0051] Several embodiments of this application will now be described in more detail with reference to the accompanying drawings to enable those skilled in the art to implement this application. This application may be embodied in many different forms and for various purposes and should not be limited to the embodiments set forth herein. These embodiments are provided to make this application thorough and complete, and to fully convey the scope of this application to those skilled in the art. The embodiments described do not limit this application.

[0052] Unless otherwise defined, all terms used herein (including technical and scientific terms) shall have the same meaning as commonly understood by one of ordinary skill in the art to which this application pertains. It will be further understood that terms such as those defined in commonly used dictionaries shall be interpreted as having a meaning consistent with their meaning in the relevant field and / or the context of this specification, and shall not be interpreted in an idealized or overly formal sense unless expressly defined herein.

[0053] Example 1

[0054] Its specific implementation method is combined with the appendix Figure 1 Please provide a detailed explanation.

[0055] Appendix Figure 1 This invention provides a block diagram of a smart parking traffic management system based on the Internet of Things for transportation, which shows the connection relationship between the traffic data perception and collection module and the guidance information generation and release module, and marks the main functional interaction flow of each module.

[0056] In this embodiment, it includes:

[0057] The traffic data sensing and acquisition module is configured to use a variety of IoT devices to achieve comprehensive and high-precision acquisition of dynamic vehicle data and static environmental data in the parking lot.

[0058] Specifically, it includes:

[0059] Deploying IoT devices in parking lots to obtain vehicle location and environmental information;

[0060] Vehicle positioning equipment: UWB positioning base stations are deployed every 5-8 meters in locations such as ceilings and pillars in the parking lot. When a vehicle enters the parking lot, it receives or senses and activates a UWB positioning tag (accuracy up to 30 centimeters); smart cameras with a resolution of 2 megapixels or higher (supporting frame rates of 25fps or higher) are installed at key intersections and curves, in conjunction with license plate recognition algorithms and trajectory tracking algorithms; ground induction coils are embedded in the ground of some lanes, one every 50 meters, to detect vehicles passing by.

[0061] Environmental sensing equipment: Temperature and humidity sensors and brightness sensors are installed in various areas of the parking lot; status monitoring chips are embedded in the guidance screens and variable information signs; millimeter-wave radar (detection distance of 50 meters and angle of 90 degrees) is set on both sides of the lanes to detect the driving status of vehicles.

[0062] UWB positioning tags send signals to the base station at preset time intervals. The base station calculates the vehicle coordinates using triangulation and uploads them. Cameras capture images at preset time intervals and extract the vehicle's position and direction of travel using image recognition technology. The driving speed is calculated by combining the continuous frame images (within ±1 km / h error). When the inductive loop detects a vehicle, it outputs a pulse signal and records the time it takes for the vehicle to pass.

[0063] The temperature and humidity sensor collects data every 5 minutes; the guidance screen status monitoring chip provides real-time feedback on screen brightness, display content, and network connection status; the millimeter-wave radar outputs information on the number of vehicles, distance, and speed within its detection range at preset intervals.

[0064] UWB positioning data is transmitted to a local data aggregation node via a dedicated wireless channel and then uploaded via Ethernet; camera data is transmitted to the data center via fiber optic cable (ensuring zero image latency); inductive loop and sensor data are transmitted via the NB-IoT network (low power consumption, configurable transmission interval); all data carries a timestamp (accurate to milliseconds) to ensure time synchronization.

[0065] The data processing and fusion module is configured to perform in-depth processing on the collected raw data, eliminate noise and errors, and achieve effective fusion of multi-source data to provide high-quality data for subsequent analysis.

[0066] Specifically, it includes:

[0067] Set a vehicle speed threshold (maximum speed limit of 10km / h in the parking lot). When a vehicle speed exceeds the vehicle speed threshold, it is identified as abnormal data and removed.

[0068] For UWB positioning data, if the distance between two adjacent positioning coordinates exceeds a preset distance (and the time interval is within 1 second), it is considered signal drift and will be discarded.

[0069] When the camera is temporarily unable to recognize a vehicle due to changes in light (not exceeding 3 seconds), the trajectory is completed using linear interpolation based on the vehicle's previous direction of travel and speed; if the ground loop data is lost, the lost data is estimated by combining the detection time of the front and rear loops and the vehicle's speed.

[0070] The (X,Y) coordinates of UWB positioning and the pixel coordinates of the camera are uniformly converted into the physical coordinates of the parking lot (with the parking lot entrance as the origin, the X-axis as the horizontal direction, and the Y-axis as the vertical direction); the vehicle IDs collected by different devices are associated (e.g., associated with UWB tag IDs through license plate recognition results).

[0071] The Kalman filter algorithm is used to fuse UWB positioning data and camera trajectory data. When the UWB signal is strong, UWB data is used as the main data source; when the camera recognition is clear, camera data is used as the main data source. The positioning error of the fused data can be reduced to within 20 centimeters. The vehicle trajectory data is matched with the parking lot electronic map (including lane width, turning radius and other attributes) to determine the lane where the vehicle is located (such as lane A, lane B) and the distance to the intersection ahead.

[0072] Choose the appropriate schema to store the data and create indexes;

[0073] A distributed database architecture is adopted, with a time-series database (such as InfluxDB) storing real-time data such as vehicle location and speed (retaining data from the last 24 hours); a relational database (such as MySQL) storing static data such as basic vehicle information and parking facility parameters; and a distributed file system (such as HDFS) storing historical vehicle trajectory data (retaining data for 30 days) and creating indexes (indexed by date and region).

[0074] The congestion prediction module is configured to obtain corresponding features based on the processed data, and input each obtained feature into the prediction model to output the corresponding congestion prediction evaluation coefficient.

[0075] Specifically, it includes:

[0076] The non-parking area of ​​the parking lot is divided into zones with a preset area to obtain various driving zones;

[0077] Determine the license plate numbers of all vehicles within the driving area and count the number of vehicles; after obtaining the license plate numbers of all vehicles within the driving area after a preset time interval, count the number of vehicles; compare the two sets of vehicles to obtain the number of vehicles with the same license plate number, and divide the number of vehicles with the same license plate number by the maximum number of vehicles that the driving area can accommodate to obtain the stationarity; since the area of ​​the driving area is known, the maximum number of vehicles that the driving area can accommodate can be roughly estimated based on the size of the vehicles.

[0078] By pre-defining driving zones by area, parking lots are transformed from chaotic spaces into standardized diagnostic units. Compared to boundless and fuzzy monitoring, this makes the calculation of stagnation level more reliable, as stagnation level directly reflects the proportion of vehicles remaining in the area.

[0079] Based on standardized stagnation levels, the system dynamically tracks vehicle congestion trends and combines historical data to create regional congestion models, providing quantitative data for resource allocation and traffic management decisions. When regional stagnation levels rise, the system can predict congestion risks, accurately calculate the number of parking spaces that need to be released, and coordinate with unoccupied parking space monitoring to proactively intervene by releasing parking spaces and guiding traffic flow. Simultaneously, the standardized data architecture supports iterative system optimization, is compatible with diverse parking scenarios, and upgrades congestion management from passive response to proactive prevention, expanding from single-area management to intelligent adaptation across all scenarios. This endows smart parking lots with scalable, self-learning, and precise decision-making capabilities for long-term efficient operation.

[0080] A preset stagnation threshold is set, and the number of unoccupied parking spaces in the driving area corresponding to stagnation levels greater than the threshold is obtained (parking spaces in parking lots are generally equipped with indicator lights above them to show the vacancy status, such as red light indicating that the parking space is occupied and green light indicating that the parking space is vacant). The overlapping area of ​​vehicles and unoccupied parking spaces within a preset range is obtained at preset time intervals. If the overlapping area increases over time, the corresponding vehicles are marked. The overlapping area is divided by the area of ​​the unoccupied parking space to obtain the integrity. The difference between the integrity of adjacent time intervals is calculated, and the absolute value is taken to obtain the integrity difference. Each integrity difference is divided by the duration of the adjacent time interval to obtain the progress per unit time. After obtaining k progress per unit time, their average value is taken to obtain the average progress per unit time. The difference between the average progress per unit time and 1 is obtained, and the absolute value is taken to obtain the average delay per unit time.

[0081] By collecting the overlapping area between vehicles and unoccupied parking spaces at preset time intervals, the system dynamically tracks vehicle entry behavior, breaking down the process layer by layer from parking space occupancy status to changes in overlapping area to completeness / progress / delay, accurately depicting the details of the entire parking process.

[0082] For example, by marking vehicles with increased overlapping areas, we can distinguish between vehicles that have actually entered the parking lot and those that are idly lingering. By calculating completeness and progress per unit time, we can quantify the efficiency of parking coverage and operational speed, allowing us to refine the causes of congestion from vehicle delays to delays in parking progress. Compared to the traditional, broad-based approach that only considers the number of occupied parking spaces, this method can diagnose hidden congestion caused by inefficient vehicle operations, providing a more granular basis for traffic management strategies. If an area has no occupied parking spaces but a high average delay per unit, then priority is given to sending parking assistance guidance. If there are few parking spaces and the overlapping area is increasing rapidly, then the focus is on guiding vehicles to parking spaces in other areas.

[0083] It integrates key metrics such as stagnation rate, completeness, progress per unit time, and average delay per unit time, covering the core aspects of "vehicle stagnation rate (reflecting the proportion of vehicles stuck in the area), parking progress (completeness reflecting the area covered by the parking system), operational efficiency (progress per unit time reflecting the speed of parking), and delay trend (average delay per unit time reflecting progress deviation)." Compared to the traditional, extensive approach that only considers traffic flow, it can accurately pinpoint the causes of congestion.

[0084] For vehicles with increased overlap in the time series, targeted entry assistance information is pushed (e.g., if the current entry angle is too far to the left, a slight adjustment to the right is recommended). Compared to indiscriminate traffic management, this reduces interference with normally driving vehicles and minimizes lane congestion caused by repeated adjustments of marked vehicles, thereby improving overall traffic efficiency.

[0085] Also includes:

[0086] A circular area is obtained by drawing a circle with the driving area as the center and a preset size as the radius. The number of fast charging piles and the number of parking spaces occupied by the fast charging piles are obtained within the circular area. The number of non-electric vehicles is extracted from the circle. The number of non-electric vehicles occupying the parking spaces corresponding to the fast charging piles is divided by the number of parking spaces occupied by the fast charging piles to obtain the occupancy rate.

[0087] This system specifically targets the hidden cause of congestion: fast charging stations being occupied by non-electric vehicles. Traditional monitoring often focuses on whether charging stations are fully occupied, but non-electric vehicle occupation can lead to a chain reaction where electric vehicles cannot charge, are stuck searching for charging stations, and exacerbate lane congestion. By calculating occupancy rates, this system accurately identifies resource misallocation issues. For example, if there are 10 fast charging stations in an area, and 6 are occupied by non-electric vehicles, it means that 60% of the fast charging resources are being ineffectively occupied, representing a potential source of congestion.

[0088] Stagnation degree, average delay per unit, and occupancy degree are respectively labeled as , , The input prediction model is then calculated as follows:

[0089]

[0090] Obtain congestion prediction evaluation coefficient ;

[0091] in,

[0092] , , These are the reference stagnation rate, the maximum permissible unit average delay, and the reference standard occupancy rate, respectively.

[0093] , , These are the weighting factors for stagnation, average delay per unit, and occupancy, respectively.

[0094] Integrating three dimensions of characteristics—stagnation, average delay per unit, and occupancy—to cover vehicle congestion ( Reflecting the proportion of vehicles stationary in the area, parking efficiency ( Reflecting delays in warehousing progress and resource mismatch ( It reflects the use of charging stations by non-electric vehicles, which is a core cause of congestion. Compared to a single dimension (such as only looking at traffic flow), it can provide a more comprehensive diagnosis of the root causes of congestion.

[0095] For example, a certain area (40% of vehicles stopped) but (Small delay in receiving goods) (Few charging pile mismatches) indicate that the main cause of congestion is short-term vehicle concentration, and the traffic management strategy focuses on temporary diversion; if If there is a serious mismatch in charging stations, priority should be given to clearing the spaces occupied by non-electric vehicles.

[0096] Multiple threshold ranges are preset, and each threshold range is matched with an estimated congestion duration. The congestion prediction evaluation coefficient is matched with the ranges of multiple thresholds to obtain the corresponding estimated congestion duration.

[0097] Multiple thresholds are preset to match the "estimated congestion duration", transforming the abstract congestion prediction evaluation coefficient into an intuitive expectation of congestion duration.

[0098] The prediction optimization module is configured to obtain optimization factors based on historical data and optimize the prediction model to output the optimized congestion prediction evaluation coefficient.

[0099] Specifically, it includes:

[0100] Obtain historical parking data for a preset time period in the parking lot, as well as the corresponding measured congestion duration;

[0101] Divide the estimated congestion duration by the measured congestion duration to obtain the optimization factor h, and take the reciprocal of the optimization factor. Then multiply it by the congestion prediction evaluation coefficient to obtain the optimized congestion prediction evaluation coefficient;

[0102] And the optimized formula for calculating the congestion prediction evaluation coefficient is obtained:

[0103]

[0104] This refers to the optimized congestion prediction evaluation coefficient.

[0105] Also includes:

[0106] The unit average schedule calculated based on the k value is:

[0107]

[0108] in This represents the time progress for the i-th unit.

[0109] Let the actual unit progress of vehicles entering the warehouse be... (Through camera trajectory extraction, such as using the time between the start of reversing and the start of the vehicle's engine being turned off as the parking time, and using this to calculate the actual unit progress.)

[0110] The progress error rate corresponding to the k value is:

[0111]

[0112] The smaller the value, the better the k-value;

[0113] A formula for correcting the k value is established by combining the optimization factor h. The optimization factor h reflects the overall prediction bias (h > 1).

[0114] (If the estimate is too long, h < 1 indicates a short estimate), the value of k needs to be adjusted to accommodate this bias.

[0115] Redefine the k value to adjust the weights:

[0116]

[0117] By schedule error rate The formula quantifies the impact of the k value on the unit average progress as an error rate, directly linking the deviation between the predicted and measured values.

[0118] The optimization factor h reflects the overall prediction bias of the system (e.g., h > 1 indicates an overestimation). The weights are adjusted using the k value. The formula allows k to adapt to this bias. If the system's overall prediction is too long due to historical data, the correction of the k value will prioritize selecting samples that can smooth fluctuations and reduce the number of samples for aggressive predictions, thereby reducing the propagation of prediction bias from the underlying feature calculation stage.

[0119] Iterate through k = 1, 2, ..., n (where n is the maximum number of samples, e.g., 10), and calculate the adjusted weight for each k. To find the optimal value of k:

[0120]

[0121] The larger the k value, the smaller the progress error corresponding to the k value, and the better the adaptability to the overall prediction deviation, enabling the system to dynamically adapt to the real-time changes of the parking lot.

[0122] Each time a batch of historical data (such as parking records for one day) is accumulated, the optimal solution for the value of k can be recalculated. As the types of vehicles in the parking lot, user parking habits, and charging pile usage patterns change (such as a surge in electric vehicle charging on weekends), the system can iteratively update the formula. It continuously adapts to new scenarios to ensure long-term prediction accuracy.

[0123] The guidance information generation and release module is configured to generate scientific and reasonable guidance information based on congestion prediction results, and accurately release it through multiple terminals to guide vehicles to divert.

[0124] Specifically, it includes:

[0125] The guidance information includes the target area, estimated travel time, and available parking spaces, such as "Currently congested in area A (8 minutes), we recommend 5 available parking spaces in area B, approximately 2 minutes' drive, then turn right after going straight along the main road";

[0126] The guidance screen displays a dynamic route map (green route indicating the vehicle's current location and target parking space, and red route indicating congested areas), and updates the congestion points on the route in real time (e.g., "Entrance to Zone B is currently clear and there is no congestion").

[0127] When a vehicle approaches an intersection, a voice message will be broadcast over the parking lot or pushed through the in-vehicle app: "Turn left into Zone B in 30 meters ahead. There are 2 empty parking spaces remaining. Please slow down."

[0128] If the average progress per unit is less than 0.3 (inbound is too slow), inbound assistance information will be pushed: "The current inbound angle is too far to the left. It is recommended to slightly adjust the steering wheel to the right. The remaining inbound progress is 60%" (calculated based on the trend of overlapping area changes).

[0129] If temporary congestion occurs around the target parking space (e.g., there are 3 or more vehicles queuing behind), a "Quick Exit Notice" will be sent: "You have completed parking. There are 3 vehicles waiting behind you. Please turn off your engine and lock your car as soon as possible to avoid exacerbating the congestion."

[0130] For electric vehicle users: "There are currently 2 available fast charging stations in Zone C (0 are occupied by non-electric vehicles), which are 40 meters away from your current location and there is no congestion on the route."

[0131] For non-electric vehicle users occupying charging station parking spaces: "The parking space D12 you are currently occupying is a dedicated fast charging station space. Occupying it by non-electric vehicles will cause electric vehicle queuing and congestion. We recommend that you move to a regular parking space (E05, vacant) 30 meters away. The alternative route has been sent."

[0132] Information is disseminated through parking guidance screens, in-vehicle apps / mini-programs, parking space indicator lights, and broadcast systems.

[0133] Parking guidance screens: Deployed at key nodes such as entrances, intersections, and elevator entrances (with a resolution of 2 megapixels or higher), prioritizing the display of congestion warnings and parking information within a 50-meter radius (e.g., the entrance guidance screen displays "Today 18:00-19:00, area A is expected to be congested, parking spaces in areas B / C are recommended").

[0134] In-vehicle APP / Mini Program: Connects vehicle users through UWB location tags, pushes personalized routes (such as "Your reserved parking space has been temporarily occupied, we will change you to parking space F08 in the same area, the navigation has been updated"), and displays the real-time status of nearby congestion points.

[0135] Parking space indicator light linkage: Above the recommended vacant parking space, in addition to the green indicator light, a flashing prompt (once per second) is added, and a small display screen next to the indicator light displays "Recommended parking space, 20 meters away from the current vehicle";

[0136] The public address system will broadcast a message in a loop over the parking lot's public address system in response to widespread congestion: "Currently, areas A and B are congested. We recommend parking in areas C or D. There are 15 parking spaces remaining in area C."

[0137] The above formulas are all dimensionless calculations. The formulas are derived from software simulations based on a large amount of collected data to obtain the most recent real-world results. The preset parameters in the formulas are set by those skilled in the art according to the actual situation.

[0138] The foregoing has only described certain exemplary embodiments of the present invention by way of illustration. Undoubtedly, those skilled in the art can modify the described embodiments in various ways without departing from the spirit and scope of the present invention. Therefore, the foregoing drawings and descriptions are illustrative in nature and should not be construed as limiting the scope of protection of the claims of the present invention.

[0139] It should be noted that, in this document, the use of relational terms such as "first" and "second" is merely for distinguishing one entity or operation from another, and does not necessarily require or imply any such actual relationship or order between these entities or operations. Furthermore, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or apparatus. Without further limitations, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or apparatus that includes the element.

[0140] It should be understood that in the various embodiments of this application, the order of the above-mentioned processes does not imply the order of execution. The execution order of each process should be determined by its function and internal logic, and should not constitute any limitation on the implementation process of the embodiments of this application.

[0141] Those skilled in the art will recognize that the units and algorithm steps of the various examples described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are implemented in hardware or software depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementation should not be considered beyond the scope of this application.

[0142] Those skilled in the art will understand that, for the sake of convenience and brevity, the specific working processes of the systems, devices, and units described above can be referred to the corresponding processes in the foregoing method embodiments, and will not be repeated here.

[0143] The units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the units can be selected to achieve the purpose of this embodiment according to actual needs.

[0144] In addition, the functional units in the various embodiments of this application can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit.

[0145] The above description is merely a specific embodiment of this application, but the scope of protection of this application is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the scope of the technology disclosed in this application should be included within the scope of protection of this application. Therefore, the scope of protection of this application should be determined by the scope of the claims.

[0146] The foregoing has only described certain exemplary embodiments of the present invention by way of illustration. Undoubtedly, those skilled in the art can modify the described embodiments in various ways without departing from the spirit and scope of the present invention. Therefore, the foregoing drawings and descriptions are illustrative in nature and should not be construed as limiting the scope of protection of the claims of the present invention.

Claims

1. A smart parking traffic management system based on the Internet of Things for transportation, characterized in that, include: The traffic data sensing and acquisition module is configured to collect dynamic vehicle data and static environmental data in the parking lot through IoT devices; The data processing and fusion module is configured to perform in-depth processing on the collected raw data, eliminate noise and errors, and achieve effective fusion of multi-source data; The congestion prediction module is configured to obtain corresponding features based on the processed data, and input each obtained feature into the prediction model to output the corresponding congestion prediction evaluation coefficient. The prediction optimization module is configured to obtain optimization factors based on historical data and optimize the prediction model to output the optimized congestion prediction evaluation coefficient. The guidance information generation and publishing module is configured to generate guidance information based on congestion prediction results to guide vehicles to divert.

2. The intelligent parking traffic management system based on the Internet of Things for transportation according to claim 1, characterized in that, The traffic data sensing and acquisition module specifically includes: Deploying IoT devices in parking lots to obtain vehicle location and environmental information; The positioning tag sends signals to the base station at preset time intervals, and the base station calculates the vehicle coordinates and uploads them; the camera captures an image frame at preset time intervals, extracts the vehicle position and driving direction through image recognition technology, and calculates the driving speed by combining the continuous frame images; when the ground induction coil detects a vehicle, it outputs a pulse signal and records the time it takes for the vehicle to pass. The guidance screen status monitoring chip provides real-time feedback on screen brightness, displayed content, and network connection status; the millimeter-wave radar outputs vehicle information within its detection range at preset intervals. Location data is transmitted to a local data aggregation node and then uploaded via Ethernet; camera data is transmitted to the data center.

3. The intelligent parking traffic management system based on the Internet of Things for transportation according to claim 1, characterized in that, The data processing and fusion module specifically includes: Set a vehicle speed threshold, and remove a vehicle when its speed exceeds the threshold. For location data, if the distance between two adjacent location coordinates exceeds a preset distance, it will be discarded; When the camera cannot recognize the vehicle, the trajectory is completed using linear interpolation based on the vehicle's previous direction and speed. If the ground loop data is lost, the lost data is estimated by combining the detection time of the front and rear loops and the vehicle's speed. The location coordinates and camera pixel coordinates are uniformly converted into the physical coordinates of the parking lot; the vehicle IDs collected by different devices are associated. By integrating location data and camera trajectory data, and matching vehicle trajectory data with the electronic map of the parking lot, the vehicle's lane and distance from the intersection ahead can be determined.

4. The intelligent parking traffic management system based on the Internet of Things for transportation according to claim 1, characterized in that, The congestion prediction module specifically includes: The non-parking area of ​​the parking lot is divided into zones with a preset area to obtain various driving zones; Determine the license plate numbers of each vehicle in the driving area and count the number of vehicles; after obtaining the license plate numbers of each vehicle in the driving area after a preset time interval, count the number of vehicles; compare the vehicles obtained in the two times to obtain the number of vehicles with the same license plate number, and divide the number of vehicles with the same license plate number by the maximum number of vehicles that the driving area can accommodate to obtain the stagnation rate. A preset stagnation threshold is set, and the number of unoccupied parking spaces in the driving area corresponding to stagnation levels greater than the stagnation threshold is obtained. The overlapping area between vehicles and unoccupied parking spaces within a preset range of unoccupied parking spaces is obtained at preset time intervals. If the overlapping area increases in the time series, the corresponding vehicles are marked. The overlapping area is divided by the area of ​​unoccupied parking spaces to obtain the integrity. The difference between the integrity of adjacent time intervals is calculated, and the absolute value is taken to obtain the integrity difference. Each integrity difference is divided by the duration of the adjacent time interval to obtain the progress per unit time. After obtaining the progress of k units of time, take their average value to obtain the average progress per unit. Then, take the difference between the average progress per unit and 1, and take the absolute value to obtain the average delay per unit.

5. A smart parking traffic management system based on the Internet of Things for transportation according to claim 4, characterized in that, Also includes: A circular area is obtained by drawing a circle with the driving area as the center and a preset size as the radius. The number of fast charging piles and the number of parking spaces occupied by the fast charging piles are obtained within the circular area. The number of non-electric vehicles is extracted from the circle. The number of non-electric vehicles occupying the parking spaces corresponding to the fast charging piles is divided by the number of parking spaces occupied by the fast charging piles to obtain the occupancy rate.

6. A smart parking traffic management system based on the Internet of Things for transportation according to claim 5, characterized in that, Stagnation degree, average delay per unit, and occupancy degree are respectively labeled as , , The input prediction model is then calculated as follows: ； Obtain congestion prediction evaluation coefficient ; in, , , These are the reference stagnation rate, the maximum permissible unit average delay, and the reference standard occupancy rate, respectively. , , These are the weighting factors for stagnation, average delay per unit, and occupancy, respectively. Multiple threshold ranges are preset, and each threshold range is matched with an estimated congestion duration. The congestion prediction evaluation coefficient is matched with the ranges of multiple thresholds to obtain the corresponding estimated congestion duration.

7. A smart parking traffic management system based on the Internet of Things for transportation according to claim 6, characterized in that, The prediction and optimization module specifically includes: Obtain historical parking data for a preset time period in the parking lot, as well as the corresponding measured congestion duration; The optimization factor is obtained by dividing the estimated congestion duration by the actual congestion duration. The reciprocal of the optimization factor is then multiplied by the congestion prediction evaluation coefficient to obtain the optimized congestion prediction evaluation coefficient.

8. A smart parking traffic management system based on the Internet of Things for transportation according to claim 7, characterized in that, Also includes: The average progress per unit is calculated based on the k-value; the progress error rate corresponding to the k-value is obtained, and then the k-value correction weight is defined; Iterate through k=1,2,...,n, and calculate the corrected weight for each k to find the optimal k value.

9. A smart parking traffic management system based on the Internet of Things for transportation according to claim 1, characterized in that, The module for generating and disseminating misleading information includes: The guidance information includes the target area, estimated travel time, and available parking spaces; a dynamic route map is displayed on the guidance screen and congestion points along the route are updated in real time; and information is disseminated through parking guidance screens, in-vehicle apps / mini-programs, and broadcast systems.

Images