Intelligent parking resource dynamic allocation method

Abstract

The invention relates to the technical field of smart parking, in particular to a smart parking resource dynamic allocation method, which comprises the steps of constructing a dynamic demand prediction model based on real-time data, generating an allocation scheme by using a distributed computing framework, and providing personalized recommendation according to user demands. The system is composed of a data acquisition module, a data processing module, a prediction analysis module, a resource allocation module and a state updating module. The prediction precision is improved through fusion of time sequence analysis and a deep learning algorithm, low-delay response is ensured by adopting a distributed architecture, and the resource utilization rate and the user experience are remarkably improved. According to the application, the defects of the existing system in the aspects of dynamic response, real-time deployment and personalized service can be effectively solved, and the problem of traffic jam is relieved.

Description

Technical Field

[0001] This invention belongs to the field of intelligent transportation and urban management technology, specifically a method for dynamic allocation of intelligent parking resources. Background Technology

[0002] With the rapid development of smart parking technology, how to improve parking lot utilization, reduce resource waste, and alleviate traffic congestion through dynamic allocation of parking resources has become a research hotspot. However, existing smart parking systems still have some shortcomings in terms of resource allocation strategies, real-time performance, and intelligence level, which affect the overall efficiency of the system and user experience.

[0003] A search revealed a patent for a smart parking system to reduce resource waste, published on November 5, 2024, with publication number CN118609414B. This patent achieves efficient use of parking resources by adding CPUs, entry displays, and communication modules to each parking lot, establishing communication connections with a resource coordination platform and a payment platform, and utilizing low-cost rental of parking spaces during off-peak hours. However, in this technical solution, the resource coordination platform only performs simple rental processing based on the current off-peak data of the parking lot, which has limitations in its real-time response capability to dynamic changes in demand. Furthermore, the system does not fully consider the differences in parking demand across different areas and time periods, which may lead to uneven or excessive concentration of resources in some areas, thus affecting the overall resource utilization rate.

[0004] A search revealed a city-level smart parking management system based on big data technology and the internet, published on October 23, 2018, with publication number CN106205204B. This patent achieves real-time release and scheduling of parking resources through a static traffic big data processing center and a three-in-one virtual city parking resource database, combined with online and offline guidance systems. However, in this technical solution, the allocation of parking resources mainly relies on historical data and static analysis models, exhibiting limited capabilities in predicting and rapidly adjusting real-time dynamic demand. Furthermore, the system fails to provide flexible resource allocation strategies for the personalized needs of different user groups (such as long-term parking, short-term parking, or reserved parking), which may have a certain impact on resource allocation efficiency and user experience.

[0005] The aforementioned problems indicate that existing smart parking systems still have room for improvement in areas such as dynamic demand response, real-time resource allocation, personalized service support, and multi-dimensional data analysis. Therefore, this invention provides a method for dynamic allocation of smart parking resources, aiming to optimize the dynamic allocation strategy of parking resources by introducing advanced real-time data analysis and intelligent algorithms, thereby improving the system's response speed and resource utilization, meeting diverse user needs, and enhancing the overall parking experience. Summary of the Invention

[0006] One of the objectives of this invention is to overcome the shortcomings of existing technologies and provide a smart parking resource dynamic allocation method that optimizes parking resource allocation strategies through real-time demand prediction and multi-dimensional data analysis.

[0007] The second objective of this invention is to provide a parking resource scheduling system based on a distributed computing architecture to implement the above-mentioned dynamic allocation method.

[0008] The third objective of this invention is to provide a parking resource allocation algorithm that supports personalized services, thereby improving user experience and meeting diverse needs.

[0009] To achieve the above objectives, the technical mechanism employed in this invention is as follows: by introducing a dynamic weight adjustment mechanism, and combining real-time traffic flow data, user behavior patterns, and historical parking lot operation data, a multi-dimensional prediction model is constructed. This model is based on time series analysis and integrates deep learning algorithms to accurately predict parking demand. Simultaneously, a distributed computing framework is used to process large-scale data streams in real time, ensuring low-latency response between data updates and decision execution.

[0010] A method for dynamic allocation of intelligent parking resources, characterized by the following steps:

[0011] 1. Based on real-time collected traffic flow data and user parking request information, construct a dynamic demand prediction model;

[0012] 2. The prediction results are processed in real time using a distributed computing framework, and a parking resource allocation scheme is generated;

[0013] 3. Based on user needs (such as short-term parking, long-term parking, or reserved parking), the allocation plan is adjusted to be personalized.

[0014] 4. Push optimal parking resource recommendations to users through smart terminal devices and simultaneously update the status of available parking spaces in the parking management system.

[0015] A parking resource scheduling system based on a distributed computing architecture, characterized in that the system consists of the following modules:

[0016] 1. Data acquisition module, used to acquire real-time data from various data sources (such as traffic monitoring cameras, user mobile devices, and parking lot sensors);

[0017] 2. The data processing module uses a distributed computing framework to clean, store, and analyze the collected data;

[0018] 3. Predictive analysis module, which predicts parking demand based on time series models and deep learning algorithms;

[0019] 4. Resource allocation module: Generates dynamic allocation schemes based on prediction results and pushes recommended information to users through smart terminal devices;

[0020] 5. Status update module: Real-time synchronization of parking space status information in the parking management system to ensure data consistency.

[0021] A parking resource allocation algorithm supporting personalized services, characterized by the following specific steps:

[0022] a. The construction of the dynamic demand forecasting model involves the following steps: a-1. Collect real-time traffic flow data and user parking request information from traffic monitoring cameras and user mobile devices; a-2. Input the collected data into a time series analysis model to extract key feature variables; a-3. Use deep learning algorithms to train the feature variables and generate the dynamic demand forecasting model.

[0023] b. The implementation of the distributed computing framework involves the following steps: b-1. The collected raw data is sharded and stored in a distributed file system; b-2. The distributed computing framework is used to process the data shards in parallel to complete data cleaning and feature extraction; b-3. The processed data is input into the prediction and analysis module to generate real-time prediction results.

[0024] c. The generation of personalized resource allocation schemes involves the following steps: c-1. Based on user parking request information, classify user needs into short-term parking, long-term parking, or reserved parking; c-2. Combine prediction results to generate a priority-ranked list of recommended parking resources for each user type; c-3. Push recommended information to users through smart terminal devices and record user feedback data.

[0025] The specific implementation of step a-1 above is as follows: a-1-1. Deploy data acquisition interfaces on traffic monitoring cameras and user mobile devices to obtain traffic flow data and user parking request information in real time; a-1-2. Transmit the collected data to the data processing module through a message queue to ensure the real-time performance and reliability of data transmission; a-1-3. Perform preliminary cleaning on the data to remove invalid data and noise interference, and generate a standardized dataset.

[0026] The specific implementation of step b-2 above is as follows: b-2-1. Divide the standardized dataset into multiple data shards, each shard containing data records within a certain time range; b-2-2. Use a distributed computing framework to process the data shards in parallel and extract key feature variables; b-2-3. Store the extracted feature variables in a distributed database for subsequent prediction and analysis modules to call.

[0027] The specific implementation of step c-2 above is as follows: c-2-1. Based on user needs, prioritize recommending the nearest available parking space to short-term parking users; c-2-2. Prioritize recommending the lowest-priced available parking space to long-term parking users; c-2-3. Reserve available parking spaces for a specified time period for users who make reservations, and send reminder notifications when the reservation time approaches.

[0028] This invention's distributed computing framework combines message queues and a distributed file system to ensure efficient data transmission and processing. By integrating time series analysis models and deep learning algorithms, prediction accuracy is significantly improved. Multiple experiments have verified that this method achieves a prediction accuracy of over 90% during peak hours, with resource allocation scheme generation time controlled within seconds. Furthermore, the personalized service module can dynamically adjust recommendation strategies based on user needs, significantly enhancing the user experience.

[0029] This invention addresses the shortcomings of existing smart parking systems in dynamic demand response, real-time resource allocation, and user experience through real-time demand forecasting and personalized service support. The high accuracy of the forecasting model and the efficiency of the distributed computing framework significantly improve the utilization rate of parking resources while alleviating traffic congestion. (See attached figures.)

[0030] Figure 1 This is a flowchart illustrating the intelligent parking resource dynamic allocation method of the present invention, showing the overall process from data collection to resource allocation. Figure 2 This is a module structure diagram of the parking resource scheduling system based on a distributed computing architecture according to the present invention, which shows in detail the composition of each functional module and their interrelationships. Figure 3 This diagram illustrates the steps of the parking resource allocation algorithm supporting personalized services in this invention, focusing on the specific processes of dynamic demand prediction, distributed computing implementation, and personalized resource allocation.

[0031] The attached diagram is labeled as follows: 1. Data acquisition module; 2. Data processing module; 3. Predictive analysis module; 4. Resource allocation module; 5. Status update module; 6. Dynamic demand prediction model; 7. Distributed computing framework; 8. Personalized resource allocation scheme. Detailed implementation methods are described below.

[0032] This invention provides a method for dynamic allocation of intelligent parking resources, along with related systems and algorithms. Specific implementation methods are detailed in the appendix. Figure 1 To be continued Figure 3 Please provide a detailed explanation. Figure 1 This demonstrates the entire process from data collection to resource allocation. Figure 2 It presents the composition of each functional module and their interrelationships. Figure 3 The specific steps of the parking resource allocation algorithm that supports personalized services are broken down.

[0033] In practical applications, data acquisition module 1 acquires traffic flow data and user parking request information in real time through data acquisition interfaces deployed on traffic monitoring cameras and user mobile devices. This data is transmitted to data processing module 2 via a message queue to ensure real-time and reliable data transmission. Data acquisition module 1 and data processing module 2 are connected via a standardized protocol. Data processing module 2 performs preliminary cleaning of the acquired data, removing invalid data and noise interference to generate a standardized dataset. The standardized dataset is then divided into multiple data shards, each containing data records within a specific time range. These data shards are stored in a distributed file system for subsequent processing.

[0034] Data processing module 2 employs a distributed computing framework 7 to perform parallel processing on data shards, extract key feature variables, and store the extraction results in a distributed database. The distributed computing framework 7 consists of multiple computing nodes connected via a high-speed network to achieve efficient data stream processing. Data processing module 2 completes data cleaning and feature extraction tasks by calling parallel processing algorithms within the distributed computing framework 7, ensuring data processing efficiency reaches second-level response times. The extracted key feature variables are then input into the predictive analysis module 3 for constructing the dynamic demand prediction model 6.

[0035] The predictive analysis module 3 predicts parking demand based on time series models and deep learning algorithms. The time series model extracts periodic patterns by analyzing historical data, while the deep learning algorithm uses neural network structures to model complex nonlinear relationships. The predictive analysis module 3 fuses the results of the two models to generate a high-precision dynamic demand prediction model 6. The predictive analysis module 3 shares data with the data processing module 2 through a distributed database, ensuring that the training data for the prediction model is always up-to-date. The prediction results are transmitted in real-time to the resource allocation module 4 via a distributed computing framework 7 to generate parking resource allocation schemes.

[0036] Resource allocation module 4 generates a dynamic allocation plan based on the prediction results and pushes recommended information to users through smart terminal devices. First, resource allocation module 4 categorizes user needs into short-term parking, long-term parking, or reserved parking based on user parking request information. For short-term parking users, resource allocation module 4 prioritizes recommending the nearest available parking space; for long-term parking users, it prioritizes recommending the available parking space with the lowest price; for reserved parking users, it reserves available parking spaces for a specified time period and sends reminder notifications as the reservation time approaches. Resource allocation module 4 and status update module 5 maintain data synchronization through a real-time communication mechanism to ensure that the parking space status information in the parking management system is always accurate.

[0037] The status update module 5 is responsible for synchronizing parking space status information in the parking management system in real time. By establishing a connection with parking sensors, the status update module 5 acquires the parking space occupancy status in real time and feeds back the updated status information to the resource allocation module 4. The status update module 5 also interacts with the data processing module 2 and the predictive analysis module 3 through the distributed computing framework 7 to ensure data consistency throughout the system. Communication between the status update module 5 and the resource allocation module 4 uses a low-latency protocol to ensure that the recommended information received by the user is completely consistent with the actual parking space status.

[0038] The generation process of personalized resource allocation scheme 8 combines the outputs of dynamic demand forecasting model 6 and distributed computing framework 7. Personalized resource allocation scheme 8 generates a priority-ranked list of parking resource recommendations based on user needs and pushes the recommendations to users via smart terminal devices. The smart terminal devices and resource allocation module 4 are connected via a wireless communication protocol to ensure timely delivery of recommendations to users. User feedback data is transmitted back to data acquisition module 1 via the smart terminal devices, forming a closed-loop feedback mechanism to further optimize the accuracy of dynamic demand forecasting model 6.

[0039] The distributed computing framework 7 employs a combination of message queues and a distributed file system to achieve efficient data stream processing. The message queue is responsible for transmitting raw data fragments acquired by the data acquisition module 1 to the data processing module 2, while the distributed file system stores the processed data fragments and extracted feature variables. The computing nodes in the distributed computing framework 7 allocate tasks through a load balancing mechanism, ensuring that the computing power of each node is fully utilized. The distributed computing framework 7 is connected to the predictive analysis module 3 and the resource allocation module 4 via a high-speed network, ensuring that the generation time of prediction results and allocation schemes is controlled within the second range.

[0040] This invention achieves dynamic allocation of smart parking resources through the collaborative work of the aforementioned modules and algorithms. Real-time data acquired from the data acquisition module 1 is cleaned and its features extracted by the data processing module 2, then input into the predictive analysis module 3 to generate a dynamic demand prediction model 6. The prediction results are used by the resource allocation module 4 to generate personalized resource allocation schemes 8, and recommended information is pushed to users through smart terminal devices. The status update module 5 synchronizes parking space status information in real time to ensure data consistency throughout the system. These modules and algorithms, through a distributed computing framework 7, achieve efficient data stream processing and low-latency response, significantly improving parking resource utilization and user experience.

[0041] In this embodiment, all modules and algorithms are implemented through hardware devices and software programs. The hardware devices include server clusters, network switches, and storage devices to support the operation of the distributed computing framework 7. The software programs include data acquisition interfaces, the core algorithms of the distributed computing framework 7, and applications for smart terminal devices. Seamless connection between the hardware devices and software programs is achieved through standardized protocols, ensuring the stability and reliability of the entire system.

[0042] The specific embodiments of the present invention are as described above. Through the collaboration of the aforementioned modules and algorithms, dynamic allocation of smart parking resources is achieved, overcoming the shortcomings of existing technologies in dynamic demand response, real-time resource allocation, and user experience. To better enable those skilled in the art to fully understand and implement the present invention, the specific implementation principles of the present invention are further supplemented below with a specific application scenario.

[0043] In the core business district of a major city, a smart parking system has been deployed to alleviate parking pressure during peak hours. The area includes multiple parking lots, each connected to the smart parking resource scheduling system via sensors and monitoring equipment. Users submit parking requests through smart devices (such as mobile apps), and the system generates dynamic allocation plans based on real-time traffic flow data, parking lot status information, and user needs, then recommends the best available parking space.

[0044] Step 1: Data Acquisition Module 1 first collects real-time traffic flow data within the area using traffic monitoring cameras, and simultaneously obtains parking request information from user mobile devices, including user location, destination, and parking duration requirements. The collected data is transmitted to Data Processing Module 2 via a message queue to ensure real-time and reliable data transmission. During this process, Data Acquisition Module 1 uses a standardized protocol to connect with Data Processing Module 2 to ensure data format consistency. Subsequently, Data Processing Module 2 performs preliminary data cleaning to remove invalid data and noise interference, generating a standardized dataset. The standardized dataset is divided into multiple data shards, each containing data records within a specific time range, and stored in a distributed file system for subsequent processing.

[0045] Step 2: Data processing module 2 invokes the distributed computing framework 7 to perform parallel processing on the data shards, extract key feature variables, and store the results in a distributed database. The distributed computing framework 7 consists of multiple computing nodes connected via a high-speed network, utilizing parallel processing algorithms to complete data cleaning and feature extraction tasks. During this process, data shards are allocated to different computing nodes, with each node independently processing a portion of the data, thus significantly shortening data processing time. The extracted key feature variables include traffic flow trends, user behavior patterns, and historical parking lot operation data, which provide the foundation for constructing the dynamic demand prediction model 6 in the subsequent predictive analysis module 3.

[0046] Step 3: Predictive Analysis Module 3 predicts parking demand based on time series models and deep learning algorithms. The time series model extracts periodic patterns from historical data analysis, such as fluctuations in parking demand during morning and evening peak hours. The deep learning algorithm uses neural network structures to model complex nonlinear relationships, such as the impact of weather conditions and holidays on parking demand. Predictive Analysis Module 3 fuses the results of the two models to generate a high-precision dynamic demand prediction model 6. The prediction results are transmitted in real-time to the resource allocation module 4 via a distributed computing framework 7 to generate parking resource allocation schemes. During this process, Predictive Analysis Module 3 and Data Processing Module 2 share data through a distributed database to ensure that the training data for the prediction model remains up-to-date.

[0047] Step 4: Resource allocation module 4 generates a dynamic allocation plan based on the prediction results and pushes recommended information to users through smart terminal devices. Specifically, resource allocation module 4 first classifies user needs into short-term parking, long-term parking, or reserved parking based on user parking request information. For short-term parking users, resource allocation module 4 prioritizes recommending the nearest available parking space to reduce the user's walking distance; for long-term parking users, it prioritizes recommending the lowest-priced available parking space to reduce the user's parking costs; for reserved parking users, it reserves available parking spaces for a specified time period and sends a reminder notification when the reservation time approaches. Resource allocation module 4 and status update module 5 maintain data synchronization through a real-time communication mechanism to ensure that the parking space status information in the parking management system is always accurate.

[0048] Step 5: The status update module 5 establishes a connection with the parking lot sensors to obtain the real-time parking space occupancy status and feeds back the updated status information to the resource allocation module 4. The status update module 5 also interacts with the data processing module 2 and the predictive analysis module 3 through the distributed computing framework 7 to ensure data consistency throughout the system. Communication between the status update module 5 and the resource allocation module 4 uses a low-latency protocol to ensure that the recommended information received by the user is completely consistent with the actual parking space status. For example, when a parking space is occupied, the status update module 5 immediately updates the parking space status information and feeds back the updated information to the resource allocation module 4, avoiding recommending occupied parking spaces to the user.

[0049] Step Six: The generation process of Personalized Resource Allocation Scheme 8 combines the outputs of Dynamic Demand Forecasting Model 6 and Distributed Computing Framework 7. Personalized Resource Allocation Scheme 8 generates a priority-ranked list of parking resource recommendations based on user needs and pushes the recommendations to users via smart terminal devices. For example, for a user who wants short-term parking in a core business district, the system will prioritize recommending the vacant parking space closest to their destination; for a user who needs long-term parking, the system will prioritize recommending the vacant parking space with the lowest price. The smart terminal device and resource allocation module 4 are connected via a wireless communication protocol to ensure that the recommendation information is delivered to the user in a timely manner. Furthermore, user feedback data is transmitted back to data acquisition module 1 via the smart terminal device, forming a closed-loop feedback mechanism to further optimize the accuracy of Dynamic Demand Forecasting Model 6.

[0050] Step 7: The distributed computing framework 7 employs a combination of message queues and a distributed file system to achieve efficient data stream processing. The message queue is responsible for transmitting raw data fragments acquired by the data acquisition module 1 to the data processing module 2, while the distributed file system stores the processed data fragments and extracted feature variables. The computing nodes in the distributed computing framework 7 allocate tasks through a load balancing mechanism to ensure that the computing power of each node is fully utilized. The distributed computing framework 7 is connected to the predictive analysis module 3 and the resource allocation module 4 via a high-speed network, ensuring that the generation time of prediction results and allocation schemes is controlled within the second range.

[0051] In the above scenario, this invention achieves dynamic allocation of smart parking resources through the collaborative work of data acquisition module 1, data processing module 2, predictive analysis module 3, resource allocation module 4, status update module 5, and distributed computing framework 7. Throughout the entire process from data acquisition to resource allocation, the modules work closely together to ensure the system's real-time performance and accuracy. For example, during peak hours, the system can quickly generate dynamic allocation schemes based on real-time traffic flow data and user parking request information, significantly improving parking resource utilization and user experience. Furthermore, the personalized service module can dynamically adjust recommendation strategies according to user needs, further satisfying diverse user requirements.

[0052] All content not described in detail in this specification is prior art known to those skilled in the art, and the model parameters of each electrical appliance are not specifically limited; conventional equipment can be used. Electrical control components not mentioned in this technical solution are not shown in the figures because they are prior art, and will not be described further here.

[0053] The above description is merely a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the protection scope of the present invention.

Claims

1. A method for dynamic allocation of intelligent parking resources, characterized in that... The method includes the following steps.

2. Data acquisition module (1) collects real-time traffic flow data and user parking request information from traffic monitoring cameras and user mobile devices, and transmits them to data processing module (2) through message queue.

3. The data processing module (2) uses a distributed computing framework (7) to clean and store the collected data and generate a standardized dataset.

4. Predictive Analysis Module (3) Processes the standardized dataset based on time series model and deep learning algorithm to generate dynamic demand prediction model (6).

5. The resource allocation module (4) generates a parking resource allocation scheme based on the output of the dynamic demand prediction model (6) and pushes recommended information to users through smart terminal devices.

6. Status update module (5) synchronizes parking space status information in the parking management system in real time to ensure data consistency.

7. The intelligent parking resource dynamic allocation method according to claim 1, characterized in that... The specific implementation steps of the data processing module (2) are as follows: a-1. The data acquisition module (1) divides the acquired raw data into multiple data segments, each of which contains data records within a certain time range; a-2. The distributed computing framework (7) performs parallel processing on the data segments and extracts key feature variables; a-3. The extracted key feature variables are stored in the distributed database for the predictive analysis module (3) to call.

8. A parking resource allocation algorithm supporting personalized services, characterized in that... The specific steps of the algorithm are as follows: b-1. The dynamic demand prediction model (6) classifies user demand into short-term parking, long-term parking or reserved parking based on user parking request information; b-2. The resource allocation module (4) generates a priority-ranked parking resource recommendation list for each type of user based on the output of the dynamic demand prediction model (6); b-3. The smart terminal device pushes recommendation information to the user and records user feedback data to optimize the dynamic demand prediction model (6).

Images