A parking space management system based on a parking space platform lock

Abstract

The application relates to the technical field of intelligent traffic management and Internet of Things, in particular to a parking space management system based on a parking space flat plate lock, which comprises a positioning fusion unit, a parking management unit, a parking space flat plate lock cluster and a navigation cooperation unit; the positioning fusion unit constructs a unified virtual space and unifies the position identification of roadside and parking lot parking spaces; the parking management unit generates personalized parking space recommendations through a multi-target decision algorithm, realizes accurate charging by combining a dynamic hierarchical pricing model, and avoids resource idling through a dynamic locking-releasing mechanism; the navigation cooperation unit interacts with third-party navigation software, realizes parking space position superposition and dynamic path planning; and a license plate recognition unit adopts two-stage verification to ensure accurate identity verification. The application realizes unified management of parking space, improves resource utilization and user parking experience, considers charging fairness and operation benefits, and is suitable for various parking space management scenes in cities.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of intelligent traffic management and Internet of Things, in particular to a parking space management system based on parking space flat plate lock. BACKGROUND

[0002] With the continuous growth of urban motor vehicle ownership, the contradiction between supply and demand of parking space resources is increasingly prominent, and the efficiency and accuracy of parking space management have become a key link in urban traffic governance. At present, parking spaces are mainly divided into roadside parking spaces and parking lot parking spaces, but the management mode is generally "decentralized and extensive", which cannot meet the needs of efficient parking and resource optimization allocation of users. The specific situation and pain points are as follows:

[0003] From the perspective of space management, the positioning and management system of roadside parking spaces and parking lot parking spaces are mutually separated. Roadside parking spaces mostly rely on GPS positioning combined with manual inspection or simple lock management, and their location information is presented in the form of geographic coordinates. While parking lot parking spaces (especially indoor parking lots) mostly use local coordinates or number management in the parking lot and rely on the internal system of the parking lot to maintain the location data. The spatial representation of the two types of parking spaces is different, and the management systems often run independently, lacking a unified spatial reference, which leads to the need for users to switch between different platforms to query parking space information, and the system cannot cross-type to coordinate and dispatch parking space resources. For example, users may find that roadside parking spaces are full in navigation software, but they do not know that there are still free parking spaces in the adjacent parking lot, or vice versa, causing waste of resources that "can be seen but cannot be used".

[0004] At the resource scheduling level, the existing parking lock and recommendation mechanism has obvious inefficiency. Although some systems support parking lock function, they lack dynamic recycling mechanism after locking: if the user locks the parking space and does not arrive on time, the parking space will be "occupied" for a long time, and other users cannot use it, leading to resource idling. In the parking recommendation link, most of them only filter according to distance or idle state, without considering user preferences (such as price sensitivity, urgency) and parking turnover rate, etc., so that high turnover parking spaces are not recommended first, and low use rate parking spaces are idle for a long time, further exacerbating the parking space shortage and resource mismatch during peak hours.

[0005] The user's parking whole-process experience pain points are also more prominent. In the parking space screening stage, the recommendation results of the existing system are often inconsistent with the actual needs of the user - for example, users sensitive to price may be preferentially recommended high-priced parking spaces, and users sensitive to distance need to detour to a far parking space; in the navigation link, traditional navigation can only guide to the parking lot entrance or the approximate area of the roadside parking space, and cannot accurately locate to a specific parking space, and the user needs to repeatedly search in the parking lot or on the roadside, especially in large parking lots, the situation of "navigation but not found" occurs frequently, and even the locked parking space is released due to too long time consumption; in the identity verification link, most need the user to manually scan the code or input the verification code to unlock, the operation is cumbersome, and the parking space occupation dispute is caused by operation delay.

[0006] Therefore, a parking space management system based on a parking space flat lock is proposed for the above problems. SUMMARY

[0007] The purpose of the present application is to provide a parking space management system based on a parking space flat lock to solve the problems raised in the background art.

[0008] To achieve the above purpose, the present application provides the following technical solutions:

[0009] A parking space management system based on a parking space flat lock, comprising a positioning fusion unit, a parking management unit, a parking space flat lock cluster and a navigation coordination unit:

[0010] The positioning fusion unit is used to construct a unified virtual space covering roadside parking spaces and parking lot parking spaces, and map the physical positions and numbers of roadside parking spaces and the three-dimensional space coordinates of parking lot parking spaces to the virtual space;

[0011] Each parking space flat lock in the parking space flat lock cluster is integrated with a positioning module, a license plate recognition unit and a lock control module; the positioning module reports the position data of the parking space in the virtual space in real time, and the license plate recognition unit collects the license plate information of the vehicle driving in;

[0012] The parking management unit stores the charging strategies of each parking space, and performs the following operations:

[0013] Receive the parking space locking request of the user terminal;

[0014] Based on the real-time parking space state in the virtual space, the user navigation path and the charging strategy, a recommended parking space list is generated through a multi-objective decision algorithm model;

[0015] Send a locking instruction to the target parking space flat lock, so that the lock control module only allows the vehicle matching the license plate information to drive in within a set time length;

[0016] The navigation cooperation unit interacts with the third-party navigation software, and real-time superimposes the position information and locking state of the recommended parking space to the navigation interface, and plans a dynamic path to reach the parking space.

[0017] As a preferred solution, the method for constructing a unified virtual space comprises:

[0018] For roadside parking spaces, associate their GPS coordinates with municipal coded parking space numbers using geofencing technology;

[0019] For parking lot spaces, obtain three-dimensional space coordinates (X, Y, Z) through BIM models or point cloud scanning ), and convert the three-dimensional space coordinates (X, Y, Z) into two-dimensional mapping coordinates (X', Y') in the virtual space ), and the conversion formula is:

[0020] , wherein is the three-dimensional coordinate of the parking lot space; is the two-dimensional mapping coordinate in the virtual space; is the rotation compensation angle of the coordinate system (unit: radian); is the height compression factor, and ;

[0021] Establish a unique position identifier for all parking spaces in the virtual space

[0022] As a preferred solution, the positioning fusion unit dynamically refreshes the range of the virtual space, which comprises:

[0023] According to the real-time position coordinates of the user terminal is the longitude value of the user's location, is the latitude value of the user's location, and a preset radius dynamically loads the parking space data in the virtual space;

[0024] An incremental update algorithm using a sliding window is adopted:

[0025] , wherein is the incremental update area; is the ratio of the circumference to the diameter; is the radius of the virtual space where the parking space data is loaded; is the user displacement.​​​​​​​​​​​​ is the average moving speed of users; is the congestion coefficient of the area.

[0026] As a preferred solution, the locking mechanism of the parking space flat lock comprises:

[0027] starting the locking countdown after receiving the locking instruction ;

[0028] the license plate recognition unit verifies the consistency of the license plate of the incoming vehicle and the license plate in the locking instruction within the countdown time window ;

[0029] if no matching vehicle is identified within the time window, triggering the parking space release signal:

[0030] wherein,

[0031] is the release signal, 1 indicates triggering release, and 0 indicates not triggering; is any time within the time window; is the locking start time; is the locking countdown duration; is the license plate matching result, 0 indicates no match, and 1 indicates match.

[0032] As a preferred solution, the parking management unit adopts a dynamic hierarchical pricing model:

[0033] total fee calculation: wherein, is the locking fee; is the parking fee; is the total fee; is the dynamic weight function of locking duration; is the locking base rate; is the dynamic weight function of layer height; is the parking base rate; is the actual parking duration;

[0034] dynamic weight function adjusted in real time by the following function:

[0035] wherein, is a natural constant; is the locking duration decay factor; is the real-time demand volatility rate;

[0036] dynamic weight function of layer height wherein, is the layer height sensitivity coefficient; ​ln(x) is the natural logarithm of x; is the floor height coefficient, 1.0 for ground floor, and increases by 0.2 for each underground floor.

[0037] As a preferred solution, the multi-objective decision algorithm model performs:

[0038] For candidate parking spaces Calculate the comprehensive score: wherein, is the comprehensive score of candidate parking space ; is the objective function weight; is the value of the th objective function;

[0039] The objective function includes:

[0040] wherein, is the value of the distance factor on candidate parking space , is the value of the cost factor on candidate parking space , is the value of the timeliness factor on candidate parking space , , are the maximum and minimum distances from the candidate parking space to the destination, respectively; is the three-dimensional distance from the parking space to the destination ; , are the maximum and minimum rates of the candidate parking space, respectively; is the maximum locking duration; is the current remaining locking duration of candidate parking space ;

[0041] Dynamic weight adjustment: wherein, is the exponential function; , are the regional feature calibration parameters; is the preference component of the th dimension in the user preference vector; is the preference component of the th dimension in the user preference vector.

[0042] As a preferred solution, the multi-objective decision algorithm model adds a turnover rate optimization term:

[0043] The revised comprehensive score: wherein, is the optimization intensity coefficient; is the turnover factor; This indicates an assignment operation;

[0044] Turnover factor: ,in, Turnover factor in candidate vehicles

[0045] Bit The value of ; For parking spaces Historical hourly turnover rate; Minimum historical hourly turnover; This represents the maximum historical hourly turnover rate. The number of currently available parking spaces; This represents the total number of parking spaces in the area.

[0046] Optimized strength coefficient: ,in, To adjust the slope; This represents the real-time occupancy rate of the area, and .

[0047] As a preferred option, the path replanning trigger condition for the navigation cooperative unit is:

[0048] ,in, The Euclidean distance between the user's actual location and the nodes of the planned path; This represents the remaining path length. Basic threshold; This is a real-time traffic fluctuation index; This represents the user's current driving speed.

[0049] As a preferred solution, the license plate recognition unit performs two-stage verification:

[0050] Phase 1: m: ,in, For stage one confidence level, It is a natural constant; This represents the sensitivity coefficient for stage one. For visual feature matching degree; The feature matching threshold for stage one;

[0051] Phase Two: m: ,in, The confidence level for stage two; For character recognition accuracy; This is the absolute value of the license plate tilt angle; The angle influence coefficient;

[0052] The time-to-lock control module sends an unlock instruction to the time-to-lock control module, wherein, The unlock instruction triggers a threshold.

[0053] The technical solutions provided by the above-mentioned application can be seen that the parking space management system based on the parking space flat lock provided by the application has the beneficial effects that:

[0054] 1. Unified management of parking space is realized, and the "information island" is broken: the application constructs a unified virtual space through the positioning fusion unit: on the one hand, the positioning information and the code of the roadside parking space are associated, and on the other hand, the spatial coordinates of the parking space of the parking lot are mapped to the virtual space, and finally all the parking spaces are identified by a unified position identifier; this design realizes the standardization of the spatial data of "roadside-parking lot" parking space, so that the system can integrate parking space resources across types, provide a consistent spatial reference for subsequent intelligent recommendation and navigation guidance, and solve the problem of "resource dispersion and lack of unified basis for scheduling" in traditional management, and improve the compatibility of the system for diversified parking spaces;

[0055] 2. The utilization rate of parking space resources is improved, and the contradiction between "idleness and competition" is reduced: the application optimizes resource scheduling through multiple mechanisms:

[0056] Dynamic locking-release mechanism: the parking space flat lock continuously verifies through the license plate recognition unit within the locking duration, and if no matching vehicle is identified, a release signal is triggered to avoid long-term idling of the parking space;

[0057] Turnover rate optimization recommendation algorithm: the multi-objective decision model adds an optimization item related to turnover in the comprehensive score to improve the priority of high-turnover parking spaces and guide the flow of resources to high-frequency use parking spaces;

[0058] Dynamic range refresh: the positioning fusion unit dynamically loads parking space data according to the user's position to ensure that resource scheduling focuses on the user's demand range and reduces invalid data processing; the combination of the three effectively improves the utilization rate of regional parking spaces and alleviates the contradiction between "parking space shortage and idleness";

[0059] 3. Optimize the user's parking process experience and solve the "difficult to find a parking space and difficult to arrive" pain points: the application optimizes the whole process from "recommendation-navigation-verification":

[0060] Personalized recommendation: the multi-objective decision algorithm adapts to user preferences through dynamic weight adjustment, adjusts the recommendation priority according to different needs (such as distance sensitivity and price sensitivity), and improves the recommendation matching degree;

[0061] Accurate navigation guidance: the navigation coordination unit superimposes the parking space position on the third-party navigation interface, dynamically replans the path in real time, and prompts specific information when approaching the parking space, solving the "last mile parking" problem;

[0062] Non-inductive identity verification: The license plate recognition unit completes unlocking through two-stage verification (long-distance feature matching + short-distance character verification), without the need for manual operation by the user, reducing the number of operation steps while improving matching accuracy, avoiding "verification is complicated or mislocking"; The whole process is optimized to greatly shorten the user's parking time, and improve the probability of "successfully reaching and using after locking the parking space";

[0063] 4. Realize dynamic and accurate pricing, and balance "fairness and operating income": The dynamic hierarchical pricing model of the application realizes accurate pricing: distinguish "locking fee" and "parking fee", dynamically adjust the weight of the locking fee according to the locking time, real-time demand, etc., and suppress "long-time invalid locking"; At the same time, adjust the parking fee according to the height difference of the parking space, so that the pricing matches the actual operating cost; This model not only guarantees user fairness (the more resources occupied, the higher the cost, the higher the fee), but also helps the operator to realize income optimization;

[0064] 5. Enhance system stability and robustness, adapt to complex actual scenarios: The application improves stability through multi-unit fault-tolerant design:

[0065] The license plate recognition unit fuses pre-processing (noise reduction, tilt correction, etc.) and multi-stage verification to maintain high recognition accuracy in complex scenarios such as rainy days, night, and license plate tilt;

[0066] The navigation coordination unit uses an efficient state synchronization mechanism to ensure that the parking lock state is updated in real time, avoiding "state synchronization leading to user misjudgment";

[0067] The data flow of each unit is carried by unified parameters, forming a "space construction-recommendation-navigation-verification" technical closed loop, so that even if a single module temporarily appears, the core function can still run through local caching and other means;

[0068] The application solves the problems of low resource efficiency and fragmented experience of existing parking space management through the technical scheme of "space unification, intelligent scheduling, whole-process cooperation, and accurate adaptation", and provides a data-driven management tool for operators, which has significant value in improving user experience, optimizing resource allocation, and ensuring operating income. BRIEF DESCRIPTION OF DRAWINGS

[0069] Figure 1 The application is a parking space management system based on a parking space flat lock. DETAILED DESCRIPTION

[0070] In order to make the purpose, technical scheme and advantages of the application clearer and more understandable, the application will be further described in detail below in combination with the drawings and examples. It should be understood that the specific examples described herein are only used to explain the application and do not limit the application.

[0071] To better understand the above technical solutions, the following will provide a detailed explanation of the technical solutions in conjunction with the accompanying drawings and specific embodiments.

[0072] like Figure 1 As shown, this embodiment of the invention provides a parking space management system based on parking space tablet locks, including a positioning fusion unit, a parking management unit, a parking space tablet lock cluster, and a navigation coordination unit:

[0073] The positioning fusion unit is used to construct a unified virtual space covering roadside parking spaces and parking lot parking spaces, and to map the physical location and number of roadside parking spaces and the three-dimensional spatial coordinates of parking lot parking spaces to this virtual space;

[0074] Each parking space flat lock in the parking space flat lock cluster integrates a positioning module, a license plate recognition unit, and a lock control module; the positioning module reports the location data of the parking space in the virtual space in real time, and the license plate recognition unit collects the license plate information of the vehicles entering the parking space;

[0075] The parking management unit stores the charging policy for each parking space and performs the following operations:

[0076] Receive parking space locking requests from user terminals;

[0077] Based on the real-time parking space status, user navigation path and charging strategy in the virtual space, a recommended parking space list is generated through a multi-objective decision-making algorithm model.

[0078] Send a locking command to the target parking space's flat lock, so that the lock control module only allows vehicles with matching license plate information to enter within a set time period;

[0079] The navigation collaboration unit interacts with third-party navigation software to overlay the location information and locking status of recommended parking spaces onto the navigation interface in real time, and plans a dynamic route to reach the parking space.

[0080] In this embodiment, the method for constructing a unified virtual space includes:

[0081] For roadside parking spaces, geofencing technology is used to link their GPS coordinates with the municipal code parking space number;

[0082] For parking spaces, three-dimensional spatial coordinates are obtained through BIM models or point cloud scanning. , , ), and the three-dimensional spatial coordinates ( , , Converting ) to two-dimensional mapped coordinates in virtual space ( , The conversion formula is:

[0083] wherein, , , is the three-dimensional coordinate of the parking space; , is the two-dimensional mapping coordinate in the virtual space; is the coordinate system rotation compensation angle (unit: radian); is the height compression factor, and ;

[0084] establishes a unique position identifier for all parking spaces in the virtual space , ;

[0085] The positioning fusion unit dynamically refreshes the range of the virtual space, including:

[0086] According to the real-time position coordinates of the user terminal , is the longitude value of the user's location, is the latitude value of the user's location, and the preset radius dynamically loads the parking space data in the virtual space;

[0087] The sliding window incremental update algorithm is adopted:

[0088] wherein, is the incremental update area; is the constant pi; is the radius of the virtual space; is the user displacement; is the average moving speed of the user; is the regional congestion coefficient;

[0089] Further, the positioning fusion unit is the "core hub" of the parking space management system based on the parking space flat plate lock to realize space unification and accurate positioning. It provides a consistent spatial reference for the system's parking space management, navigation planning, and other functions through the construction of a virtual space covering multiple parking spaces and a dynamic refresh mechanism. The following will elaborate on this unit from the overall to the details:

[0090] I. Overall function overview:

[0091] The positioning fusion unit is mainly responsible for constructing a unified virtual space covering roadside parking spaces and parking lot parking spaces, mapping physical position information of different types of parking spaces to the virtual space to form a standardized position identification system. Meanwhile, the range of the virtual space is dynamically refreshed according to the real-time position of the user, ensuring that the system efficiently loads relevant parking space data and providing accurate and consistent spatial data support for path planning of the parking management unit and the navigation coordination unit, thereby guaranteeing the accuracy and coordination of the system when processing diversified parking space information.

[0092] II. Submodule composition and functions

[0093] (I) Virtual space construction unit

[0094] Parking space position information collection: For roadside parking spaces, the GPS coordinates thereof are obtained through the geographic fence technology, and the corresponding municipal coded parking space number is associated. For parking lot parking spaces, the three-dimensional space coordinates thereof are collected by using the BIM model or point cloud scanning technology, thereby providing original data for subsequent space mapping. , ,

[0095] Coordinate conversion and mapping: The three-dimensional coordinates of the parking lot parking spaces are converted to obtain the two-dimensional mapping coordinates in the virtual space , , and the conversion formula is:

[0096] wherein, , , are the three-dimensional coordinates of the parking lot parking spaces; , are the two-dimensional mapping coordinates in the virtual space; is the coordinate system rotation compensation angle (unit: radian); and is the height compression factor ).

[0097] Through the conversion, the accurate mapping of the parking lot three-dimensional parking spaces in the virtual two-dimensional space is realized, and the GPS coordinates of the roadside parking spaces are uniformly represented in the same virtual space.

[0098] Dynamic range refreshing unit

[0099] User position monitoring: The real-time position , is the longitude value of the position where the user is located, and is the latitude value of the position where the user is located, which is used as the reference point for dynamically refreshing the range of the virtual space.

[0100] ​Sliding window incremental update: the sliding window incremental update algorithm is used to calculate the incremental update area, and the formula is:

[0101] wherein, is the incremental update area; is the radius of the virtual space loaded with parking space data; is the user displacement amount; is the user average moving speed; is the area congestion coefficient (C); According to the user position change, moving speed and area congestion, the range of parking space data that needs to be updated is dynamically determined;

[0102] Data loading and updating: according to the calculated incremental update area and range, the parking space data in the virtual space is dynamically loaded, and when the user moves, the parking space information within a certain radius around the user is updated in time, which not only ensures that the user can obtain relevant parking space data, but also avoids unnecessary data loading, improving the system running efficiency;

[0103] III. Key technical principles:

[0104] (I) Virtual space construction technology principle:

[0105] The virtual space construction technology is based on coordinate conversion and fusion theory. For different position representation methods of roadside parking spaces and parking lot parking spaces, through the geographic fence technology and three-dimensional coordinate conversion algorithm, they are unified into the same virtual space. For roadside parking spaces, GPS coordinates and municipal codes are used to achieve accurate positioning. For parking lot parking spaces, a conversion relationship between three-dimensional coordinates and two-dimensional mapping coordinates is established to eliminate the difference between three-dimensional space and two-dimensional space, and then combined with the generation of unique position identifiers, the standardization and uniqueness of all parking spaces in the virtual space are realized, providing consistent spatial reference for each module of the system.

[0106] (II) Dynamic range refreshing technology principle:

[0107] The dynamic range refreshing technology is based on user position change and sliding window algorithm. According to the user's real-time position, moving speed and area congestion, the range of parking space data that needs to be loaded is dynamically adjusted. The sliding window incremental update algorithm determines the amount of data that needs to be updated each time by calculating the incremental update area, which not only ensures that the user can obtain effective parking space information, but also reduces the load of data transmission and processing, improves the response speed of the system to user position changes and data processing efficiency, and ensures that the user can obtain real-time parking space status around them during movement.

[0108] IV. Workflow of the unit:

[0109] ​(I) initialization phase:

[0110] After the positioning fusion unit is started, the initialization of its hardware and software is completed, including the loading of coordinate conversion parameters, the starting of the geographic fence technology module, the connection detection with the parking space data acquisition device, etc., to ensure the normal operation of each functional module;

[0111] The communication connection with other units of the system is established, such as the parking management unit, the navigation coordination unit, etc., to prepare to receive and send parking space position related data;

[0112] (II) virtual space construction phase:

[0113] The GPS coordinates and municipal coded parking space numbers of roadside parking spaces are collected, and the two are associated through the geographic fence technology to determine the initial position of roadside parking spaces in the virtual space;

[0114] The three-dimensional coordinates of parking lot parking spaces are collected , , , and the coordinate conversion formula is used to convert them into two-dimensional mapping coordinates in the virtual space , ;

[0115] A unique position identifier is generated for all parking spaces , completing the construction of a unified virtual space;

[0116] (III) dynamic range refresh phase:

[0117] The real-time position of the user terminal is received in real time , and the displacement of the user is calculated ;

[0118] According to the user displacement, the average moving speed , the regional congestion coefficient and the radius , the incremental update area is calculated through the sliding window incremental update algorithm ;

[0119] According to the incremental update area and the range, the parking space data in the virtual space is dynamically loaded and updated, and the updated parking space position information is sent to the parking management unit and the navigation coordination unit;

[0120] (IV) end phase:

[0121] When the system stops running or receives a stop instruction, the positioning fusion unit stops data collection, coordinate conversion and dynamic refresh operations, disconnects the communication connection with other units, saves the related parameters and data of the virtual space construction, and releases system resources;

[0122] V. Application value of the unit:

[0123] (I) Realize the unified management of parking space:

[0124] By constructing a unified virtual space, the problem of different representation methods of roadside parking spaces and parking lot parking spaces is solved, the management of various parking spaces under the same spatial framework is realized, a consistent spatial reference is provided for the subsequent operation of the system, and the processing capacity of the system for diversified parking space information is improved;

[0125] (II) Improve the data processing efficiency of the system:

[0126] The dynamic range refresh mechanism dynamically loads parking space data according to the user's position, avoids the waste of resources caused by full data loading, reduces the burden of data transmission and processing, improves the running efficiency and response speed of the system, and ensures that the user can obtain relevant parking space information in time;

[0127] (III) Ensure the accuracy of parking space positioning:

[0128] Through accurate coordinate conversion and generation of unique position identifiers, accurate positioning of parking spaces in virtual space is ensured, reliable position data support is provided for functions such as parking space recommendation and navigation coordination unit path planning of the parking management unit, and the overall service quality of the system is improved;

[0129] (Four) Enhance the compatibility and expansibility of the system:

[0130] The unified virtual space and position identification system enable the system to be compatible with different types of parking spaces and positioning methods, facilitate the subsequent access of new parking space types or update of positioning technology, enhance the compatibility and expansibility of the system, and provide protection for the long-term development of the system.

[0131] In this embodiment, the locking mechanism of the parking space flat plate lock includes:

[0132] Start the lock countdown after receiving the lock instruction ;

[0133] The license plate recognition unit verifies the consistency of the license plate of the entering vehicle and the license plate in the lock instruction within the countdown time window;

[0134] If a matching vehicle is not recognized within the time window, a parking space release signal is triggered:

[0135] wherein,

[0136] is a release signal, 1 indicates triggering release, and 0 indicates not triggering; is any time within the time window;​ is the lock start time; is the lock countdown duration; is the license plate matching result, 0 means not matching, 1 means matching;

[0137] The parking management unit adopts a dynamic hierarchical pricing model:

[0138] Total cost calculation: wherein, is the lock fee; is the parking fee; is the total fee; is the dynamic weight function of the lock duration; is the lock reference rate; is the dynamic weight function of the layer height; is the parking reference rate; is the actual parking duration;

[0139] Dynamic weight function adjusted in real time by the following function:

[0140] wherein, is a natural constant; is the lock duration attenuation factor; is the real-time demand fluctuation rate;

[0141] Dynamic weight function of layer height wherein, is the layer height sensitivity coefficient; is the natural logarithm; is the layer height coefficient, 1.0 for the ground floor, and 0.2 for each underground floor;

[0142] Multi-objective decision algorithm model execution:

[0143] For candidate parking spaces Calculate the comprehensive score: wherein, is the comprehensive score of the candidate parking space ; is the objective function weight; is the value of the th objective function;

[0144] Objective function includes:

[0145] wherein, is the value of the distance factor on the candidate parking space , is the value of the cost factor on the candidate parking space , The value of the aging factor on the candidate parking space , , The maximum and minimum distance from the candidate parking space to the destination, respectively ; The three-dimensional distance from the parking space to the destination ; , The highest and lowest rate of the candidate parking space, respectively ; The maximum locking duration ; The current remaining duration of the candidate parking space ;

[0146] Dynamic weight adjustment: , wherein , The regional feature calibration parameter ; The preference component of the first dimension in the user preference vector ; The preference component of the first dimension in the user preference vector ;

[0147] Multi-objective decision algorithm model adds turnover rate optimization term:

[0148] The revised comprehensive score: , wherein is the optimization intensity coefficient ; Indicates the assignment operation

[0149] Turnover factor: , wherein

[0150] The value of the turnover factor on the candidate parking space ; The historical hourly turnover number of the parking space ; The minimum historical hourly turnover number ; The maximum historical hourly turnover number ; The current number of available parking spaces ; The total number of parking spaces in the region

[0151] Optimization intensity coefficient: , wherein is the adjustment slope ;

[0152] Further, the parking management unit is the "smart hub" of the parking space management system based on the parking space flat lock, and undertakes the core functions of parking space resource scheduling, dynamic pricing, intelligent recommendation, etc. Through the integration of parking space status, user demand and algorithm model, efficient management and accurate service of parking space are realized. The following will be described in detail from the whole to the details:

[0153] I. Overall function overview:

[0154] The core function of the parking management unit is to receive the parking lock request of the user, based on the real-time parking space status of the virtual space provided by the positioning fusion unit, the user path data of the navigation coordination unit, combined with the preset charging strategy, to generate a recommended parking space list through a multi-objective decision algorithm, send a lock instruction to the target parking flat lock, and calculate the fee according to the dynamic layered pricing model. At the same time, this unit is also responsible for monitoring the parking lock state, and executing the parking release operation when the release condition is triggered, forming a complete closed loop of "request-recommendation-locking-pricing-release", to ensure efficient use of parking space resources and accurate matching of user demand.

[0155] II. Submodule composition and function:

[0156] (I) Parking lock management unit:

[0157] Request processing and instruction generation: Receive the parking lock request sent by the user terminal (including user location, target area, estimated lock duration, etc.), combined with the real-time state of the parking space in the virtual space (idle / occupied), to generate a preliminary screened candidate parking pool;

[0158] Lock instruction execution: Send a lock instruction to the target parking flat lock, which contains user vehicle license plate information, lock duration and other parameters, triggering the lock control module to enter the locked state, allowing only the vehicle with the matching license plate to enter within the set time;

[0159] Release signal monitoring: Real-time receive the license plate recognition result of the parking flat lock, and judge whether the release signal is triggered according to the locking mechanism. The release signal formula is:

[0160] , where

[0161] is the release signal, 1 indicates that the release is triggered (i.e. no matching vehicle is identified within the lock duration), and 0 indicates that the release is not triggered; is any time within the time window; is the lock start time; is the lock countdown duration; is the license plate matching result, 0 indicates no match, and 1 indicates match; when the release signal is 1, send an unlock instruction to the parking flat lock, and re-include the parking space into the available resource pool;

[0162] (II) Dynamic tiered pricing unit:

[0163] Cost parameter acquisition: Real-time acquisition of parameters such as lock duration , actual parking duration , layer height coefficient (Ground floor = 1.0, +0.2 for each underground floor), and real-time demand fluctuation rate , as the basis for pricing;

[0164] Total cost calculation: Calculate the total cost based on the dynamic tiered pricing model, with the formula:

[0165] , where is the lock fee; is the parking fee; is the total cost (unit: yuan); is the dynamic weight function of lock duration; is the lock reference rate (yuan / hour); is the dynamic weight function of layer height; is the parking reference rate (yuan / hour);

[0166] Dynamic weight adjustment: Adjust the weight coefficients in real time through the following function:

[0167] , where is a natural constant; is the lock duration decay factor , used to suppress resource idling caused by long-term locking; is the real-time demand fluctuation rate, with the coefficient increasing as demand increases;

[0168] , where is the layer height sensitivity coefficient ; is the natural logarithm, with the coefficient increasing as the layer height increases (e.g., deep underground), reflecting the parking cost difference of different layer heights;

[0169] (III) Multi-objective decision recommendation unit:

[0170] Candidate parking space screening: Screen idle parking spaces from the virtual space as candidates, and obtain data such as the location identifier , real-time rate , and lock remaining duration for each candidate parking space ;

[0171] Comprehensive score calculation: Calculate the comprehensive score of the candidate parking space through a multi-objective decision algorithm, with the basic formula being:

[0172] ,in, Candidate parking spaces Overall score; The weights are the objective function weights. For the first The values ​​of the objective function;

[0173] The objective function includes:

[0174] ;

[0175] in, , These represent the maximum and minimum distances from the candidate parking spaces to the destination, respectively. For parking spaces Arrive at the destination The three-dimensional distance; , These are the highest and lowest rates for the candidate parking spaces, respectively. Maximum lock duration; Candidate parking spaces Current remaining lock duration;

[0176] Weighting and scoring optimization: Adjusting the weights of each objective function using a dynamic weighting formula to adapt to user preferences.

[0177] ,in, It is an exponential function; Calibrate parameters for regional features; For the user preference vector, the first... The preference components of each dimension, such as high price sensitivity among users. Take a larger value; For the user preference vector, the first... Preference components in each dimension;

[0178] At the same time, the turnover rate optimization item has been revised and the scoring has been improved:

[0179] ,in, To optimize the strength coefficient; (This is a turnover factor; the larger the value, the higher the parking space turnover rate), further improving the overall efficiency of resource utilization.

[0180] Recommendation list generation: based on overall rating Sort the parking spaces from highest to lowest, generate a recommended parking space list, and send it back to the user's terminal.

[0181] III. Key Technology Principles:

[0182] (I) Multi-objective decision algorithm principle:

[0183] The algorithm is based on the dual objectives of user preference and system efficiency, through the weighted fusion of distance, cost, and time efficiency factors, to achieve the balance of "user satisfaction" and "resource optimization"; the dynamic weight mechanism makes the recommendation results adapt to different user needs (such as commuting users prefer distance, temporary users prefer cost), and the turnover rate optimization item reduces resource idling and improves overall utilization of regional parking spaces by increasing the priority of high-turnover parking spaces;

[0184] (II) Dynamic hierarchical pricing principle:

[0185] By introducing the lock duration decay factor and demand fluctuation rate , the lock fee decreases with the increase of duration, avoiding "occupying but not using"; at the same time, the weight adjustment of the layer height coefficient reflects the differences in construction and maintenance costs of underground parking spaces, achieving "cost-oriented" accurate pricing, taking into account fairness and economy;

[0186] (III) Locking-releasing mechanism principle:

[0187] Based on the time window license plate verification logic, ensure that the locked resources are only used by the target user, and the release signal mechanism timely recovers the parking space when the user does not arrive on time, avoiding resource waste, forming an efficient scheduling mode of "locking has a time limit, idle can be recycled";

[0188] Four, the working process of the unit:

[0189] (I) Request receiving stage:

[0190] Receive the parking lock request sent by the user terminal, parse the user location, destination, estimated use time, etc. in the request;

[0191] Call the virtual space data of the positioning fusion unit to get the idle parking space list and location identifier in the target area;

[0192] (II) Recommendation generation stage:

[0193] The multi-objective decision recommendation unit calculates the comprehensive score of the idle parking spaces , generates a recommended parking space list and returns it to the user;

[0194] Receive the target parking space information selected by the user, confirm the lock duration ;

[0195] (III) Locking execution stage:

[0196] Send a lock instruction to the target parking space flat lock, including license plate information, parameters such as Start the lock countdown;

[0197] Real-time monitoring of license plate recognition results of parking space flat lock to determine whether to trigger a release signal;

[0198] (Four) cost calculation phase:

[0199] When the vehicle enters and completes parking, record the actual parking duration And the height coefficient ;

[0200] The dynamic hierarchical pricing unit calculates the total cost , , Based on parameters such as And push to the user terminal to complete payment;

[0201] (Five) release and recovery phase:

[0202] If the vehicle enters on time, the parking space is automatically unlocked after parking is completed; if a matching vehicle is not identified within the lock duration, the release signal is triggered, the parking space is unlocked and is re-included in the available resource pool;

[0203] Record this transaction data (such as parking space number, cost, duration, etc.) for subsequent rate optimization and user behavior analysis.

[0204] In this embodiment, the path re-planning trigger condition of the navigation coordination unit is:

[0205] Where, Is the Euclidean distance between the user's actual position and the planned path node; Is the remaining path length; Is the basic threshold; Is the real-time traffic fluctuation index; Is the user's current driving speed;

[0206] Further, the navigation coordination unit is the "path guide" of the parking space management system based on the parking space flat lock. Through deep interaction with third-party navigation software, it realizes the real-time fusion of parking space information and navigation path, and provides dynamic guidance services for users from the starting point to the target parking space. The following is a detailed description from the whole to the details:

[0207] I. Overall function overview:

[0208] The core function of the navigation collaboration unit is to interface with third-party navigation software, overlaying information such as the recommended virtual parking space location provided by the positioning fusion unit and the locking status feedback from the parking management unit onto the navigation interface in real time. It also dynamically plans the optimal arrival route based on the user's real-time location, traffic conditions, and changes in parking space status. At the same time, this unit has the ability to monitor route deviation. When the user deviates from the planned route or traffic conditions change suddenly, it automatically triggers route replanning to ensure that the user arrives at the target parking space efficiently, forming a closed-loop service of "information overlay - route planning - dynamic adjustment".

[0209] II. Submodule Composition and Functions:

[0210] (a) Navigation Data Interaction Unit:

[0211] Interface adaptation and protocol conversion: Establish communication connections with mainstream third-party navigation software (such as Gaode and Baidu Maps) through standardized APIs (Application Programming Interfaces), and convert the system's internal virtual spatial location identifiers. Converted into latitude and longitude coordinates recognizable by navigation software, enabling cross-platform data interoperability;

[0212] Information synchronization and push: Push key information about recommended parking spaces to navigation software in real time, including:

[0213] The actual geographical location mapped from virtual space (based on the location fusion unit) (Conversion);

[0214] The parking management unit provides feedback on the locking status ("locked", "about to be released", etc.);

[0215] Estimated arrival time (calculated using real-time traffic data);

[0216] User behavior data feedback: Receives real-time user location feedback from navigation software. , planning path nodes Current driving speed This data provides input parameters for path replanning;

[0217] (ii) Dynamic path planning unit:

[0218] Initial route generation: based on the user's starting point and the virtual spatial location of the target parking space. It calls the route planning interface of third-party navigation software and combines it with real-time traffic data (such as congested road sections and traffic light durations) to generate an initial optimal route, including distance. Expected time Information such as;

[0219] Route replanning trigger judgment: Real-time monitoring of the deviation between the user's actual driving trajectory and the planned route, triggering replanning when the following conditions are met:

[0220] wherein, is the Euclidean distance between the user's actual position and the planned path node (unit: meters); is the remaining path length (unit: meters); is the base threshold value (default value 0.15); is the real-time traffic fluctuation index (value 0-5, the larger the value, the more unstable the traffic); is the user's current driving speed (unit: km / h);

[0221] Dynamic path adjustment: after triggering re-planning, the path from the user's current position to the target parking space is recalculated, new congestion sections are preferentially avoided, and the path display and estimated arrival time on the navigation interface are updated synchronously;

[0222] (Three) Parking space status superimposed display unit:

[0223] Real-time labeling of locked state: superimpose dynamic identification at the position of the target parking space on the navigation interface:

[0224] Green flashing icon: indicates that the parking space has been locked and is within the effective duration;

[0225] Yellow warning icon: indicates that the remaining locking duration is less than 5 minutes;

[0226] Red prohibited icon: indicates that the parking space has been released (due to timeout without arrival);

[0227] Key information pop-up prompt: when the user is within 1 km of the target parking space, the pop-up window displays:

[0228] Parking space number and specific parking area (such as "underground second floor B area 32");

[0229] Remaining locking duration (such as "still locked for 8 minutes");

[0230] Entrance of parking lot / nearby road parking lot intersection guidance;

[0231] Multi-modal guidance fusion: combined with the voice broadcast function of navigation software, customized voice prompts (such as "turn right into the parking lot 300 meters ahead, the target parking space has been reserved for you"), realizing the coordinated guidance of vision and hearing;

[0232] Three, key technical principles:

[0233] (One) Path re-planning trigger mechanism principle:

[0234] ​The mechanism judges whether to need to re-plan by "relative deviation rate" instead of "absolute distance", the core logic is: when the ratio of user deviation degree and remaining path length exceeds the dynamic threshold, the adjustment is triggered; among them, the traffic fluctuation index and the introduction of driving speed make the threshold change dynamically with traffic conditions-the more congested or the lower the speed , the higher the threshold, to avoid frequent re-planning affecting user experience; otherwise, the threshold is reduced to ensure timely correction of the path;

[0235] (2) Cross-platform data fusion principle:

[0236] Through the virtual space position identifier as an intermediate bridge, the mapping problem of system internal coordinates (virtual space) and third-party navigation coordinates (real geographical coordinates) is solved; the provided by the positioning fusion unit contains latitude and longitude information, which can be directly converted into WGS84 coordinate system coordinates supported by navigation software, while the layer depth information is supplemented through a pop-up window to realize "two-dimensional navigation + three-dimensional layer depth" accurate guidance;

[0237] (3) State synchronization real-time principle:

[0238] The "incremental update + heartbeat detection" mechanism is adopted to ensure state synchronization: the navigation coordination unit sends a heartbeat request to the parking management unit every 2 seconds to obtain the lock state change of the target parking space; when the state changes (such as from "locked" to "about to be released"), only the change data is transmitted instead of the full information, reducing the communication bandwidth occupation and ensuring that the state update delay is controlled within 1 second;

[0239] Four, the working process of the unit:

[0240] (1) Initialization and docking phase:

[0241] After the navigation coordination unit starts, it automatically scans and docks with the third-party navigation software already installed in the user's mobile phone, completes interface authentication and protocol negotiation;

[0242] Receive the target parking space information (position , lock duration ) pushed by the parking management unit, convert it into coordinates and text description that can be recognized by navigation software;

[0243] (2) Initial navigation phase:

[0244] Input the target parking space coordinates into the navigation software to generate an initial path from the user's current location to the parking space, and superimpose the parking lock state identifier on the interface;

[0245] Real-time receive navigation software feedback user position And planning path node , Calculate the remaining path length ;

[0246] (Three) Dynamic monitoring and adjustment phase:

[0247] Every 0.5 seconds to calculate the path deviation rate, to determine whether to meet the re-planning trigger condition;

[0248] If the re-planning is triggered, call the navigation software interface to generate a new path and update the interface display; if not, continue to monitor the user position change;

[0249] When the user is within 1 km of the parking space, trigger a pop-up prompt and voice guidance, and update the remaining locking time simultaneously;

[0250] (Four) Arrive end phase:

[0251] When the user arrives at the target parking space (navigation software determines arrival), send an "arrival" confirmation signal to the parking management unit, and close the state identifier;

[0252] If the user does not arrive in time and the parking space has been released, display "parking space has been released" in the navigation interface, and recommend other idle parking spaces in the surrounding area;

[0253] End navigation coordination service, save this navigation data (such as path deviation rate, arrival time, etc.) for subsequent optimization;

[0254] Five, the application value of the unit:

[0255] (I) Improve parking accessibility:

[0256] Through dynamic path planning and accurate guidance, solve the pain point of "finding a parking space but not finding a specific location", especially in large parking lots, which can reduce the user's search time by more than 30%;

[0257] (II) Ensure effective use of locking resources:

[0258] Real-time synchronization of locking state and warning of remaining time, reminding users to arrive in time, reducing the probability of "getting lost and releasing the parking space", and improving the effectiveness of the locking mechanism;

[0259] (III) Optimize navigation service experience:

[0260] Customized parking state identifier and guidance information, which makes up for the defects of traditional navigation that can only reach the parking lot entrance and cannot directly reach the parking space, achieving "door-to-door" closed-loop navigation;

[0261] (Four) Enhance system compatibility:

[0262] Through the standard interface to the mainstream navigation software, without the need for users to install a dedicated APP, reduce the threshold of use, improve the popularity of the system and user acceptance.

[0263] In this embodiment, the license plate recognition unit performs two-stage verification:

[0264] Stage one: m: wherein, is the confidence of stage one, is a natural constant; is the sensitivity coefficient of stage one; is the visual feature matching degree; is the feature matching threshold of stage one;

[0265] Stage two: m: wherein, is the confidence of stage two; is the character recognition accuracy; is the absolute value of the license plate inclination angle; is the angle influence coefficient;

[0266] Send an unlock instruction to the lock control module, wherein, is the unlock instruction trigger threshold;

[0267] Further, the license plate recognition unit is the core component of the parking space flat lock to realize "identity verification", through the double-stage visual recognition and confidence fusion mechanism, accurately collects and verifies the license plate information of the vehicle entering, provides the decision basis for the unlock / lock operation of the lock control module, and at the same time guarantees that the parking space locking resources are only opened to the target vehicle; The following is described in detail from the whole to the details:

[0268] I. Overall function overview:

[0269] The license plate recognition unit is integrated in each parking space flat lock, and the core function is to collect the license plate information of the vehicle entering the parking space in real time. Through the double-stage verification logic of "long-distance feature matching + short-distance character verification", the comprehensive confidence of license plate matching is calculated, and when the confidence meets the threshold condition, an unlock instruction is sent to the lock control module; If a matching license plate is not recognized within the locking time, cooperate with the parking management unit to trigger the parking space release signal; This unit needs to adapt to different distances, light, license plate inclination and other complex scenes to ensure the accuracy and real-time performance of license plate recognition, and is the key link to realize "license plate-parking space" accurate binding;

[0270] II. Submodule composition and function:

[0271] (I) Image acquisition and preprocessing subunit:

[0272] Multi-distance image capture: Utilizing an integrated high-definition camera (supporting 1080P resolution and 30 frames per second acquisition), the system dynamically adjusts shooting parameters based on the distance between the vehicle and the parking space lock. When m, activate wide-angle mode to capture the overall outline of the vehicle and the license plate area; when When the camera is in macro mode, switch to macro mode to focus on the details of the license plate and ensure that the characters are clearly visible.

[0273] Image preprocessing: Performing a series of optimization operations on the acquired raw images:

[0274] Noise reduction: Gaussian filtering is used to remove image noise in rainy or backlit scenes;

[0275] Tilt Correction: The license plate border is identified using an edge detection algorithm, and the tilt angle is calculated. (Unit: degrees), and perform rotation correction on the image;

[0276] Contrast Enhancement: Histogram equalization is used to improve the contrast between license plate characters and the background, providing high-quality image input for subsequent recognition;

[0277] (II) Long-distance feature matching subunit (Phase 1:) m):

[0278] License plate region localization: In the preprocessed long-distance image, the license plate region is located by color (blue background / yellow background / white background) and shape (rectangular outline) features, and visual features such as texture and character arrangement within the region are extracted;

[0279] Feature matching degree calculation: The extracted features are compared with the target license plate feature template (including key features such as character spacing and font style) issued by the parking management unit to calculate the visual feature matching degree. (Value range 0-1, 1 indicates a complete match);

[0280] Stage 1 Confidence Output: The feature matching degree is converted into a Stage 1 confidence degree using the Sigmoid function. The formula is:

[0281] ,in, The confidence level for stage one (values ​​0-1), It is a natural constant; This is the sensitivity coefficient for Phase 1 (default value is 2.5; the larger the value, the more significant the impact of changes in the matching degree on the confidence level).

[0282] For visual feature matching degree; Threshold for feature matching in phase 1 (default value 0.6, below which the confidence decreases rapidly);

[0283] (Three) Close-range character verification sub-unit (phase 2: m):

[0284] Character recognition and accuracy calculation: for the rectified license plate image taken by micro-lens, the OCR (optical character recognition) algorithm is used to extract the license plate characters (including Chinese characters, letters and numbers), and the character recognition accuracy is calculated by comparing with the target license plate character sequence (value range 0-1, 1 means complete consistency);

[0285] Inclination angle compensation: the influence of license plate inclination angle on recognition accuracy is introduced for correction, and the absolute value of the angle Adjust the confidence: the greater the inclination angle, the more serious the character distortion, and the greater the confidence correction amplitude;

[0286] Phase 2 confidence output: combine character accuracy and angle compensation to calculate phase 2 confidence , the formula is:

[0287] , wherein is the confidence of phase 2 (value 0-1); is the character recognition accuracy; is the absolute value of the license plate inclination angle (unit: degree); is the angle influence coefficient (default value 0.02, the greater the value, the stronger the weakening of the angle on the confidence);

[0288] (Four) Comprehensive verification and instruction sub-unit:

[0289] Two-stage confidence fusion: calculate the product of the confidence of the two stages as the final verification result:

[0290] ;

[0291] Unlock instruction triggering: when ( is the unlock instruction triggering threshold, default value 0.7) sends an unlock instruction to the lock control module, allowing the vehicle to enter the parking space; if the threshold is not reached, it is marked as "non-matching vehicle";

[0292] Result feedback: the verification result (match / non-match) is fed back to the parking management unit in real time as the basis for determining whether to trigger the release signal (such as parameter value);

[0293] Three, key technology principles:

[0294] ​(1) The hierarchical logic principle of two-stage verification:

[0295] The hierarchical strategy of "long-distance feature matching + short-distance character verification" essentially improves the recognition robustness in complex scenarios through "rough screening + precise verification": the long-distance stage ( m) quickly excludes obvious non-target vehicles (such as vehicles with large differences in vehicle type and license plate color) through feature matching, reducing subsequent computational load; the short-distance stage ( m) achieves precise verification through character-level verification, while introducing tilt angle compensation to correct the influence of image distortion - the combination of the two stages avoids the low precision problem of single long-distance recognition and solves the image missing problem caused by the vehicle not being completely stopped in the short-distance recognition;

[0296] (2) Mathematical principle of confidence fusion:

[0297] The two-stage confidence is fused in the form of product ( ), rather than simple addition, and the core logic is "double condition constraint": only when the long-distance feature matching is reliable ( high) and the short-distance character recognition is accurate ( high), the total confidence will meet the threshold - this "and logic" can effectively reduce the probability of false unlocking (such as vehicles with similar features but different characters, or vehicles with occasional accurate character recognition but large feature differences, which cannot pass the verification);

[0298] (3) Dynamic adaptive environmental compensation principle:

[0299] For interference such as light changes and license plate stains, environmental compensation is achieved through "parameter self-adaptive adjustment" of the preprocessing subunit: the camera has a built-in light sensor that automatically turns on infrared fill light when the light intensity is below 50 lux; for character fuzzy scenes caused by stains, morphological operations (such as erosion-dilation) are used to enhance character edges to ensure the stability of OCR recognition - these compensation mechanisms enable the unit to maintain an accuracy rate of over 95% in cloudy, night, and slightly worn license plate scenarios;

[0300] Four, the working process of the unit:

[0301] (1) Initialization stage:

[0302] The license plate recognition unit starts with the parking space flat lock, completes camera self-check (focus, exposure parameter calibration), OCR model loading, and target license plate feature template (obtained from the parking management unit) caching;

[0303] Set the initial parameters: , , , start the distance sensor to monitor the vehicle distance ;

[0304] (ii) Vehicle approach monitoring phase:

[0305] When the distance sensor detects m (warning distance), start the image acquisition and preprocessing subunit, and start periodic shooting (1 frame per second);

[0306] Continue to monitor changes, when m, enter phase one verification; when m, switch to phase two verification;

[0307] (iii) Two-stage verification phase:

[0308] Phase one ( ):

[0309] Locate the license plate area and extract visual features, calculate ;

[0310] Generate through the formula, if (low confidence threshold), directly marked as "non-matching vehicle" and continue to monitor; if ≥ 0.3, wait for the vehicle to approach further;

[0311] Phase two ( ):

[0312] Take a macro image and correct the tilt, extract characters through OCR, calculate ;

[0313] Combine generate through the formula;

[0314] Comprehensive verification:

[0315] Calculate ;

[0316] If , send an unlock instruction to the lock control module; otherwise, record "non-matching" results;

[0317] (iv) Continuous verification phase within the lock duration:

[0318] Within the lock countdown set by the parking management unit (time window ), repeat the two-stage verification every 3 seconds, and update ;

[0319] If the verification is passed at any time ( ), continuously send a "matching success" signal to the lock control module to maintain the unlocked state;

[0320] (V) Result feedback and end phase:

[0321] If the verification fails all the time (i.e. ), a "non-matching vehicle" feedback is sent to the parking management unit to trigger a release signal;

[0322] When the vehicle completely enters the parking space or after the end, stop image acquisition, save this identification log (including confidence level, verification result), and enter standby state;

[0323] Five, the application value of the unit:

[0324] (I) Ensure the accuracy of the parking lock:

[0325] Through two-stage verification and confidence fusion, effectively avoid "fake license plate mislocking" and "similar license plate misjudgment" and other problems, make the license plate matching accuracy rate improve to more than 98%, ensure that the locked parking space is only open to the target vehicle, and protect the rights and interests of users;

[0326] (II) Adapt to the stability of complex scenes:

[0327] The preprocessing and environment compensation mechanism can cope with common scenes such as rainy days, night, license plate inclination (≤15 degrees), etc., solve the "recognition failure" problem of traditional single OCR recognition in extreme environment, and improve the reliability of the system in actual application;

[0328] (III) Support efficient recovery of parking resources:

[0329] Through continuous verification within the locking time, identify "non-target vehicle occupation" or "user does not arrive on time" scene in time, provide accurate basis for the parking management unit to trigger the release signal, reduce the invalid locking of parking resources (according to the test, it can reduce 12% of the parking idle rate);

[0330] (IV) Simplify the user operation process:

[0331] No need for users to manually scan codes or input verification codes, complete identity verification through license plate automatic identification, realize "unconscious unlocking", reduce user operation steps, and improve the convenience of parking experience.

[0332] Although the embodiments of the present application have been shown and described, it can be understood by those skilled in the art that various changes, modifications, replacements and variations can be made to the embodiments without departing from the principles and spirits of the present application, and the scope of the present application is defined by the appended claims and their equivalents.

Claims

1. A parking space management system based on a parking space flat lock, characterized in that: Includes a positioning fusion unit, a parking management unit, a parking space flat lock cluster, and a navigation coordination unit: The positioning fusion unit is used to construct a unified virtual space covering roadside parking spaces and parking lot parking spaces, and to map the physical location and number of roadside parking spaces and the three-dimensional spatial coordinates of parking lot parking spaces to the virtual space. Each parking space flat lock in the parking space flat lock cluster integrates a positioning module, a license plate recognition unit, and a lock control module; the positioning module reports the location data of the parking space in the virtual space in real time, and the license plate recognition unit collects the license plate information of the vehicles entering the parking space; The parking management unit stores the charging strategy for each parking space and performs the following operations: Receive parking space locking requests from user terminals; Based on the real-time parking space status, user navigation path and charging strategy in the virtual space, a recommended parking space list is generated through a multi-objective decision-making algorithm model. Send a locking command to the target parking space's flat lock, so that the lock control module only allows vehicles with matching license plate information to enter within a set time period; The navigation collaboration unit interacts with third-party navigation software to overlay the location information and locking status of recommended parking spaces onto the navigation interface in real time, and plans a dynamic route to reach the parking space. Execution of multi-objective decision-making algorithm model: For candidate parking spaces Calculate the overall score: ,in, Candidate parking spaces Overall score; The weights are the objective function weights. For the first The values ​​of the objective function; The objective function includes: ,in, Distance factor in candidate parking spaces The value on, Cost factors in candidate parking spaces The value on, The timeliness factor in candidate parking spaces The value on, , These represent the maximum and minimum distances from the candidate parking spaces to the destination, respectively. For parking spaces Arrive at the destination The three-dimensional distance This serves as a unique location identifier for the parking space in the virtual space. , These are the highest and lowest rates for the candidate parking spaces, respectively. Maximum lock duration; Candidate parking spaces Current remaining lock duration; Candidate parking spaces Real-time rates; Dynamic weight adjustment: ,in, It is an exponential function; , Calibrate parameters for regional features; For the user preference vector, the first... Preference components in each dimension; For the user preference vector, the first... Preference components in each dimension.

2. The parking space management system based on a parking space flat lock according to claim 1, characterized in that: The method for constructing the unified virtual space includes: For roadside parking spaces, geofencing technology is used to link their GPS coordinates with the municipal code parking space number; For parking spaces, three-dimensional spatial coordinates are obtained through BIM models or point cloud scanning. , , ), and the three-dimensional spatial coordinates ( , , Converting ) to two-dimensional mapped coordinates in virtual space ( , The conversion formula is: ,in, , , The three-dimensional coordinates of the parking spaces in the parking lot; , These are two-dimensional mapped coordinates in virtual space. This is the coordinate system rotation compensation angle; It is a high compression factor, and ; Establish a unique location identifier for all parking spaces in the virtual space. , .

3. The parking space management system based on a parking space flat lock according to claim 1, characterized in that: The range of the virtual space dynamically refreshed by the positioning fusion unit includes: Based on the real-time location coordinates of the user terminal , The longitude value of the user's location. The latitude value of the user's location, with a preset radius. Dynamically load parking space data from the virtual space; Using a sliding window incremental update algorithm: ,in, Update the area incrementally; Pi; The radius for loading parking space data in the virtual space; For user displacement; This represents the average movement speed of the user. This represents the regional congestion coefficient.

4. The parking space management system based on a parking space flat lock according to claim 1, characterized in that: The locking mechanism of the parking space flat lock includes: Start the lock countdown after receiving the lock command. ; The license plate recognition unit is in the countdown time window Internally verify the consistency between the license plate of the entering vehicle and the license plate in the locking command; like If no matching vehicle is detected, a parking space release signal is triggered: ,in, For the release signal, 1 indicates that release is triggered, and 0 indicates that release is not triggered; Any moment within the time window; To lock the start time; To lock the countdown duration; This represents the license plate matching result, where 0 indicates no match and 1 indicates a match.

5. The parking space management system based on a parking space flat lock according to claim 4, characterized in that: The parking management unit adopts a dynamic tiered pricing model: Total cost calculation: ,in, To lock in the cost; Parking fees; Total cost; A dynamic weighting function for locking duration; To lock in the base rate; The dynamic weighting function for the floor height; The base parking fee rate; This refers to the actual parking time. Dynamic weight function for lock duration Adjust in real time using the following function: ,in, It is a natural constant; To lock the duration decay factor; Real-time demand volatility; Dynamic weighting function of floor height ,in, This is the floor height sensitivity coefficient; It is the natural logarithm; The floor height coefficient is 1.0 for the ground floor and 0.2 for each underground floor.

6. The parking space management system based on a parking space flat lock according to claim 1, characterized in that: The multi-objective decision-making algorithm model adds a turnover rate optimization term: Revised overall score: ,in, To optimize the strength coefficient; Turnover factor; This indicates an assignment operation; Turnover factor: ,in, Turnover factor in candidate vehicles Bit The value of ; For parking spaces Historical hourly turnover rate; Minimum historical hourly turnover; This represents the maximum historical hourly turnover rate. The number of currently available parking spaces; This represents the total number of parking spaces in the area. Optimized strength coefficient: ,in, To adjust the slope; This represents the real-time occupancy rate of the area, and .

7. The parking space management system based on a parking space flat lock according to claim 1, characterized in that: The path replanning trigger condition for the navigation coordination unit is: ,in, For the user's actual location With planned path nodes The Euclidean distance; This represents the remaining path length. Basic threshold; This refers to the real-time traffic fluctuation index. This represents the user's current driving speed.

8. The parking space management system based on a parking space flat lock according to claim 1, characterized in that: The license plate recognition unit performs two-stage verification: Phase 1: m: ,in, For vehicle distance, For stage one confidence level, It is a natural constant; This represents the sensitivity coefficient for stage one. For visual feature matching degree; The feature matching threshold for stage one; Phase Two: m: ,in, The confidence level for stage two; For character recognition accuracy; This is the absolute value of the license plate tilt angle; The angle influence coefficient; Send an unlock command to the lock control module, wherein, The threshold for triggering the unlock command.

Images