Non-motor vehicle integrated management system based on ultrahigh frequency RFID technology

Abstract

The invention relates to the field of non-motor vehicle management systems, in particular to a non-motor vehicle integrated management system based on an ultrahigh frequency RFID (Radio Frequency Identification Device) technology, which comprises a radio frequency identification module, an intelligent prediction module, a block chain module and an anti-theft alarm module which are in wireless communication connection, life cycle management, real-time positioning and dynamic trajectory tracking of non-motor vehicles are realized; the intelligent prediction module is used for intelligently analyzing and predicting the use frequency, the moving path and the parking position of the non-motor vehicle, and optimizing urban traffic planning and parking resource configuration; the block chain module is used for guaranteeing the safety and non-tampering of the vehicle data through a block chain layer; and the anti-theft alarm module is used for automatically triggering an alarm for the non-motor vehicle which illegally moves or is forbidden to enter an area, and sending a real-time notification to a management department and a vehicle owner. Traffic planning and parking resource configuration can be optimized, safety management is enhanced, and management transparency and user credibility are improved.

Description

Technical Field

[0001] The present invention relates to the field of non-motor vehicle management systems, and in particular to a non-motor vehicle integrated management system based on ultra-high frequency RFID technology. Background Art

[0002] With the continuous development of urban transportation and the continuous growth of non-motorized vehicles, non-motorized vehicle management faces many challenges. The traditional non-motorized vehicle management method mainly relies on manual registration and inspection, which is inefficient and difficult to achieve real-time monitoring and full life cycle management of vehicles.

[0003] On the one hand, there is a lack of effective intelligent analysis and prediction methods for the use frequency, movement paths and parking locations of non-motor vehicles, which makes it difficult to accurately optimize urban traffic planning and parking resource allocation. During peak traffic hours, the disorderly flow of non-motor vehicles may cause traffic congestion, and unreasonable parking arrangements will occupy public space and affect the beauty and order of the city. On the other hand, there are also hidden dangers in the safety management of non-motor vehicles. Vehicles are easily stolen, and it is difficult to track the location of stolen vehicles without effective monitoring methods. At the same time, there is a lack of timely alarm and notification mechanisms for non-motor vehicles that move illegally or enter prohibited areas, making it impossible to effectively manage and regulate them.

[0004] In addition, the management data of non-motor vehicles is at risk of being tampered with and lacks sufficient security and transparency, which reduces users' trust in the management system. Summary of the invention

[0005] To solve the above problems, the present invention provides a non-motor vehicle integrated management system based on ultra-high frequency RFID technology, which can optimize traffic planning and parking resource allocation, enhance safety management, and improve management transparency and user trust.

[0006] To achieve the above object, the technical solution adopted by the present invention is:

[0007] A non-motor vehicle integrated management system based on ultra-high frequency RFID technology, including a radio frequency identification module, an intelligent prediction module, a blockchain module and an anti-theft alarm module connected by wireless communication, wherein:

[0008] The radio frequency identification module is used to achieve life cycle management, real-time positioning and dynamic trajectory tracking of non-motor vehicles through ultra-high frequency RFID tags and reading and writing devices;

[0009] The intelligent prediction module is used to intelligently analyze and predict the use frequency, movement path and parking location of non-motor vehicles, and optimize urban traffic planning and parking resource allocation;

[0010] The blockchain module is used to ensure the security and immutability of vehicle data, improve management transparency and user trust;

[0011] The anti-theft alarm module is used to automatically trigger an alarm for non-motor vehicles that are illegally moved or enter prohibited areas, and send real-time notifications to management departments and vehicle owners.

[0012] Furthermore, the method of realizing life cycle management, real-time positioning and dynamic trajectory tracking of non-motor vehicles through ultra-high frequency RFID tags and reading and writing devices includes the following steps:

[0013] A1: The RFID reading and writing equipment deployed at the main traffic nodes automatically reads the tag information bound to the vehicle information management system that stores the detailed information of non-motor vehicles within the identification range. The reading and writing equipment transmits the read tag information to the vehicle information management system through the wireless network;

[0014] A2: The vehicle information management system receives the tag information data transmitted by the reading and writing device for comparison, determines the identity and location information of the vehicle, updates the vehicle's location information in real time according to the tag information read by the reading and writing device, and displays the vehicle's current location on the map;

[0015] A3: The vehicle owner queries the real-time location of the vehicle through the location information display terminal to obtain the vehicle's whereabouts, and records and stores the vehicle's location information at different time points to form a dynamic trajectory of the vehicle.

[0016] Furthermore, the RFID reading and writing device automatically reads the tag information, including: the tag recognition algorithm determines the type and unique identification of the tag through feature extraction and pattern recognition of the tag signal;

[0017] The determining of the position information of the vehicle includes: measuring the strength of the tag signal to estimate the distance between the vehicle and the reading and writing device through a positioning algorithm based on signal strength, and determining the position of the vehicle through a triangulation positioning or a multilateral positioning algorithm based on the position information of multiple reading and writing devices;

[0018] The formation of the dynamic trajectory of the vehicle includes: recording and storing the position information of the vehicle at different time points through a trajectory tracking algorithm to form the dynamic trajectory of the vehicle; performing time series analysis on the vehicle position data to calculate the displacement and speed of the vehicle in different time periods to determine the driving direction and speed of the vehicle; and performing rationality analysis on the driving trajectory of the vehicle in combination with map data and traffic rules to determine abnormal behavior.

[0019] Furthermore, the intelligent analysis and prediction of the use frequency, movement path and parking location of non-motor vehicles to optimize urban traffic planning and parking resource allocation includes the following steps:

[0020] B1: Continuously collect non-motor vehicle tag information data through RFID reading and writing devices at major traffic nodes and parking areas, store the data according to the distributed framework, remove noise data, fill in missing values ​​and standardize the data;

[0021] B2: Through the counting algorithm, the data is distributed to multiple computing nodes in a distributed environment. Each node counts the number of times the tag is read in a specific time period, and the results of each node are summarized to obtain the frequency of use of non-motor vehicles, and a frequency heat map is drawn;

[0022] B3: Through the graph construction algorithm, the locations of the reading and writing devices in the city are regarded as the nodes of the graph, and the movement of tags between different nodes is regarded as the edges of the graph. The weight of the edge is determined according to the time sequence and time interval of the tags at different reading and writing device locations, and then the main movement paths of non-motor vehicles are analyzed through the graph analysis algorithm;

[0023] B4: Determine the main parking locations of non-motor vehicles through clustering algorithm;

[0024] B5: Use time series analysis algorithms to predict historical data and obtain trend values ​​of the usage frequency, movement paths, and parking locations of non-motor vehicles in future time periods. Combine these trend values ​​with environmental characteristics as input and input them into the machine learning algorithm for training and prediction.

[0025] Furthermore, the counting algorithm includes: extracting label data within a specific time period from the database and distributing it to each node of the distributed computing framework, running a mapper task on each computer point to count the assigned label data, and after the shuffle and sort stages, the redducer task aggregates the mapper results to obtain the final frequency;

[0026] The graph analysis algorithm includes: analyzing the importance of nodes in the graph through the PageRank algorithm to obtain PageRank values, nodes with high PageRank values ​​represent locations where non-motor vehicles pass more frequently, and the paths between these locations are the main movement paths.

[0027] Furthermore, the method of determining the main parking location of non-motor vehicles by using a clustering algorithm in combination with a distributed computing framework includes:

[0028] Read RFID data, extract the geographic location coordinates and residence time of the tag, and use the K-Means algorithm in the distributed computing framework to determine the number of clusters k according to the elbow rule. Then assign the data points to the cluster where the nearest cluster center is located to obtain preliminary clustering results. Based on the residence time, the clustering results are screened and optimized to obtain the main parking locations.

[0029] Furthermore, the historical data is predicted by a time series analysis algorithm to obtain trend values ​​of the use frequency, movement path and parking location of non-motor vehicles in the future time period, including:

[0030] Time series data on the usage frequency, movement paths and parking locations of non-motor vehicles are collected, and the data are tested for stationarity using the unit root test method. The autoregressive order, difference order and moving average order of the ARIMA model are determined based on the observed autocorrelation function and partial autocorrelation function. The prediction model is obtained by training the historical data using the ARIMA model with determined parameters.

[0031] Furthermore, the trend values ​​are combined with environmental features as inputs to a machine learning algorithm for training and prediction, including:

[0032] The trend value predicted by the ARIMA model is merged with the environmental characteristic data through the numpy library to generate a multidimensional array, wherein the environmental characteristic data includes: weather data, season data, holiday data and time data;

[0033] The merged data set is divided into training set, validation set and test set. The initial structure of the recurrent neural network algorithm is two layers, with 128 neurons in each layer. Softmax is used as the activation function. The dimension of the input layer matches the dimension of the merged trend value and environmental feature data. When predicting the frequency of use of non-motor vehicles, the dimension of the output layer is a scalar. When predicting the moving path or parking position, the dimension of the output layer is a vector or matrix.

[0034] The weights and biases of the recurrent neural network are initialized by the He initialization method. The training cycle is started through the training set according to the cross entropy loss function and the Adam optimization algorithm. The value of the loss function is calculated on the validation set to evaluate the model performance. The learning rate, number of layers or number of neurons are adjusted according to the performance. The best model parameters saved during the training process are loaded. The new trend value and environmental feature data are input into the recurrent neural network loaded with the best model parameters, and the prediction results are obtained through forward propagation.

[0035] Furthermore, the blockchain layer is used to ensure the security and immutability of vehicle data, improve management transparency and user trust, including:

[0036] Interact and verify the key data of non-motor vehicles with the IOTA blockchain platform through the blockchain layer. The data on the chain includes: the unique identification, registration information, usage frequency, movement path and parking location of non-motor vehicles, and pre-process the data;

[0037] The nodes in the IOTA network will verify the newly uploaded transactions to ensure the integrity and authenticity of the data. If the data is tampered with or invalid, the transaction will be rejected. The management software sends a query request through the interface with the IOTA node, and the IOTA node returns the query result;

[0038] The data is encrypted before being uploaded to the chain, and only authorized users can decrypt and access the data. Access to the data on the blockchain is restricted by setting access permissions. Security audits are performed regularly on the data on the blockchain to detect any abnormal transactions or data tampering.

[0039] Furthermore, the automatic triggering of an alarm for a non-motor vehicle that moves illegally or enters a prohibited area and sending a real-time notification to the management department and the vehicle owner include:

[0040] According to the rules for determining areas where non-motor vehicles are prohibited from entering and illegal movement, additional RFID reading and writing devices are deployed around the prohibited areas and at key traffic nodes. When the RFID reading and writing devices read the tag information of the non-motor vehicle, they immediately transmit the data to the management system. The management system continuously monitors the location and movement trajectory of the vehicle, and determines whether the vehicle has entered the prohibited area by comparing it with the pre-set boundary coordinates of the prohibited area. It also determines whether there is illegal movement based on the vehicle's movement speed and direction information. If the system detects that the non-motor vehicle is moving illegally or entering a prohibited area, an alarm is triggered immediately.

[0041] The beneficial effects of the present invention are:

[0042] 1. Through ultra-high frequency RFID technology and distributed computing framework, the unique identity of non-motor vehicles can be quickly and accurately identified to achieve full life cycle management of vehicles. Compared with the traditional manual registration method, it greatly improves management efficiency and reduces labor and time costs. Real-time positioning and dynamic trajectory tracking functions enable management departments to grasp the location information of non-motor vehicles at any time, facilitate timely handling of violations and emergencies, and improve management response speed.

[0043] 2. Provide a scientific basis for urban traffic planning through intelligent analysis and prediction of the frequency of use, movement paths and parking locations of non-motor vehicles. Increase the width of non-motor vehicle lanes or improve road conditions in areas with high usage frequency, and set up more traffic signs and facilities on the main movement paths to improve the traffic efficiency of non-motor vehicles. Accurately determine the main parking locations of non-motor vehicles to help optimize the allocation of parking resources. Increase the number of parking spaces and improve parking facilities in areas with high parking demand to improve the utilization efficiency of urban space.

[0044] 3. For non-motor vehicles that move illegally or are prohibited from entering areas, the system can automatically trigger an alarm and send real-time notifications to management departments and vehicle owners, so that violations can be discovered and handled in a timely manner, enhancing the safety management of non-motor vehicles. The combination of blockchain technology ensures the security and non-tamperability of vehicle data, reducing the risk of vehicle theft. Once a vehicle is stolen, the management system can assist in quickly recovering the stolen vehicle through the real-time positioning function, improving the safety of the vehicle.

[0045] 4. The introduction of blockchain technology makes non-motor vehicle management data more secure and reliable, and its tamper-proof nature increases the transparency of management. Management departments and vehicle owners can query the vehicle's historical records and current status at any time, which enhances trust in the management system and facilitates data exchange with vehicle management departments. Users can receive real-time notifications through mobile phone applications and other means to understand the vehicle's status, which improves user participation and satisfaction. BRIEF DESCRIPTION OF THE DRAWINGS

[0046] Figure 1 This is a module diagram of a non-motor vehicle integrated management system based on ultra-high frequency RFID technology.

[0047] Figure 2 It is a flow chart of the steps of the RFID module.

[0048] Figure 3 It is a step flow chart of the intelligent prediction module. DETAILED DESCRIPTION

[0049] See also Figure 1 , Figure 2 and Figure 3 As shown, the present invention relates to a non-motor vehicle integrated management system based on ultra-high frequency RFID technology, including a radio frequency identification module, an intelligent prediction module, a blockchain module and an anti-theft alarm module connected by wireless communication, wherein:

[0050] The radio frequency identification module is used to achieve life cycle management, real-time positioning and dynamic trajectory tracking of non-motor vehicles through ultra-high frequency RFID tags and reading and writing devices;

[0051] The intelligent prediction module is used to intelligently analyze and predict the use frequency, movement path and parking location of non-motor vehicles, and optimize urban traffic planning and parking resource allocation;

[0052] The blockchain module is used to ensure the security and immutability of vehicle data, improve management transparency and user trust;

[0053] The anti-theft alarm module is used to automatically trigger an alarm for non-motor vehicles that are illegally moved or enter prohibited areas, and send real-time notifications to management departments and vehicle owners.

[0054] Furthermore, the method of realizing life cycle management, real-time positioning and dynamic trajectory tracking of non-motor vehicles through ultra-high frequency RFID tags and reading and writing devices includes the following steps:

[0055] A1: The RFID reading and writing equipment deployed at the main traffic nodes automatically reads the tag information bound to the vehicle information management system that stores the detailed information of non-motor vehicles within the identification range. The reading and writing equipment transmits the read tag information to the vehicle information management system through the wireless network;

[0056] A2: The vehicle information management system receives the tag information data transmitted by the reading and writing device for comparison, determines the identity and location information of the vehicle, updates the vehicle's location information in real time according to the tag information read by the reading and writing device, and displays the vehicle's current location on the map;

[0057] A3: The vehicle owner queries the real-time location of the vehicle through the location information display terminal to obtain the vehicle's whereabouts, and records and stores the vehicle's location information at different time points to form a dynamic trajectory of the vehicle.

[0058] Specifically, the tag information includes the unique identification information of the vehicle, such as vehicle model, license plate number and owner information. The reading distance of the tag should be moderate, which should ensure that it can be accurately read by the reading and writing equipment at traffic nodes and parking areas, and avoid misreading the tags of other vehicles. The main traffic nodes include intersections, bridges, tunnels and other locations, as well as parking areas, such as parking lots, community parking spaces and other places. The installation location of the reading and writing equipment should take into account the signal coverage range and reading angle to ensure that the tag information of the non-motor vehicles passing by can be accurately read. The vehicle information management system is used to store detailed information of non-motor vehicles, including vehicle model, license plate number, owner's name, contact information, purchase date, etc. The database has data backup and recovery functions to prevent data loss.

[0059] After purchasing a non-motor vehicle, the owner brings the vehicle-related documents and his / her identity certificate to the designated registration point for vehicle registration. The registration staff uses a dedicated RFID reader / writer to read the tag information on the vehicle and enter the vehicle information into the database. The registration staff reviews the entered vehicle information to ensure the accuracy and completeness of the information. After the review is passed, the vehicle information is bound to the tag to achieve the unique identity of the vehicle.

[0060] RFID reading and writing devices arranged at traffic nodes and parking areas continuously scan the surrounding environment. When a non-motor vehicle passes by, the reading and writing device automatically reads the tag information on the vehicle. The location information display terminal includes a mobile phone APP or a web page. The owner can query the vehicle's driving trajectory within a certain period of time and understand the vehicle's driving route and activity range.

[0061] Furthermore, the RFID reading and writing device automatically reads the tag information, including: the tag recognition algorithm determines the type and unique identification of the tag through feature extraction and pattern recognition of the tag signal;

[0062] The determining of the position information of the vehicle includes: measuring the strength of the tag signal to estimate the distance between the vehicle and the reading and writing device through a positioning algorithm based on signal strength, and determining the position of the vehicle through a triangulation positioning or a multilateral positioning algorithm based on the position information of multiple reading and writing devices;

[0063] The formation of the dynamic trajectory of the vehicle includes: recording and storing the position information of the vehicle at different time points through a trajectory tracking algorithm to form the dynamic trajectory of the vehicle; performing time series analysis on the vehicle position data to calculate the displacement and speed of the vehicle in different time periods to determine the driving direction and speed of the vehicle; and performing rationality analysis on the driving trajectory of the vehicle in combination with map data and traffic rules to determine abnormal behavior.

[0064] Specifically, the tag recognition algorithm is mainly used to accurately identify the tag information read by the RFID reader and writer and extract the unique identification of the vehicle. The algorithm needs to consider factors such as the tag's signal strength, reading angle, interference, etc. to improve the accuracy and reliability of tag recognition. For example, the frequency, phase, amplitude and other characteristics of the tag signal are combined with a machine learning algorithm to classify and identify the tags. The tag recognition algorithm is mainly used in the data collection step. When the RFID reader and writer scans the surrounding environment, it reads signals from multiple tags. The tag recognition algorithm processes these signals, identifies the unique identification of each tag, and transmits it to the vehicle information management system. The tag recognition algorithm is the basis of the entire system, which provides the vehicle's unique identification for subsequent positioning algorithms and trajectory tracking algorithms. Only by accurately identifying the tag information can the vehicle be located and tracked.

[0065] The positioning algorithm based on signal strength, when the vehicle information management system receives the tag information transmitted by the read-write device, the positioning algorithm will calculate the location of the vehicle based on the unique identifier of the tag and the location information of the read-write device, and display the current location of the vehicle on the map. The positioning algorithm based on signal strength needs to rely on the unique identifier of the vehicle provided by the tag recognition algorithm, as well as the location information of the read-write device. At the same time, the results of the positioning algorithm will also provide the vehicle's location data to the trajectory tracking algorithm, which is used to record the vehicle's driving trajectory.

[0066] The trajectory tracking algorithm is used to record and analyze the vehicle's driving trajectory. It uses a time series-based analysis method to process the vehicle's location information at different time points and extract the vehicle's driving direction, speed and other information. The trajectory tracking algorithm will record and store the vehicle's location information at different time points to form a dynamic trajectory of the vehicle. The owner can query the vehicle's driving trajectory within a certain period of time to understand the vehicle's driving route and activity range. The trajectory tracking algorithm needs to rely on the vehicle location data provided by the positioning algorithm and the vehicle's unique identification provided by the tag recognition algorithm. At the same time, the results of the trajectory tracking algorithm can also provide a decision-making basis for the traffic management department to optimize traffic planning and management.

[0067] Furthermore, the intelligent analysis and prediction of the use frequency, movement path and parking location of non-motor vehicles to optimize urban traffic planning and parking resource allocation includes the following steps:

[0068] B1: Continuously collect non-motor vehicle tag information data through RFID reading and writing devices at major traffic nodes and parking areas, store the data according to the distributed framework, remove noise data, fill in missing values ​​and standardize the data;

[0069] B2: Through the counting algorithm, the data is distributed to multiple computing nodes in a distributed environment. Each node counts the number of times the tag is read in a specific time period, and the results of each node are summarized to obtain the frequency of use of non-motor vehicles, and a frequency heat map is drawn;

[0070] B3: Through the graph construction algorithm, the locations of the reading and writing devices in the city are regarded as the nodes of the graph, and the movement of tags between different nodes is regarded as the edges of the graph. The weight of the edge is determined according to the time sequence and time interval of the tags at different reading and writing device locations, and then the main movement paths of non-motor vehicles are analyzed through the graph analysis algorithm;

[0071] B4: Determine the main parking locations of non-motor vehicles through clustering algorithm;

[0072] B5: Use time series analysis algorithms to predict historical data and obtain trend values ​​of the usage frequency, movement paths, and parking locations of non-motor vehicles in future time periods. Combine these trend values ​​with environmental characteristics as input and input them into the machine learning algorithm for training and prediction.

[0073] Specifically, the distributed architecture is based on Apache Dor is for data analysis and query, and is a distributed SQL data warehouse based on MPP (massive parallel processing) architecture.

[0074] The steps of the graph construction algorithm are as follows: reading RFID data, extracting the movement records of tags, adding corresponding edges to the graph for each movement record of tags, calculating the weight of the edges according to factors such as time intervals, and storing the constructed graph in a distributed Neo4j or Titan graph database.

[0075] Furthermore, the counting algorithm includes: extracting label data within a specific time period from the database and distributing it to each node of the distributed computing framework, running a mapper task on each computer point to count the assigned label data, and after the shuffle and sort stages, the redducer task aggregates the mapper results to obtain the final frequency;

[0076] The graph analysis algorithm includes: analyzing the importance of nodes in the graph through the PageRank algorithm to obtain PageRank values, nodes with high PageRank values ​​represent locations where non-motor vehicles pass more frequently, and the paths between these locations are the main movement paths.

[0077] Specifically, the counting algorithm distributes data to multiple computing nodes in a distributed environment. Each node is responsible for counting the number of times a part of the tags are read within a specific time period. Finally, the results of each node are summarized to obtain the usage frequency of each non-motor vehicle. According to the usage frequency of non-motor vehicles in different areas, a usage frequency heat map is drawn. For areas with high usage frequency, the width of non-motor vehicle lanes is increased, more traffic signs are set up, or the flatness of the road is improved to improve riding safety and comfort.

[0078] The PageRank algorithm is used to analyze the importance of nodes in a graph. In the mobile path analysis, nodes with high PageRank values ​​represent locations where non-motor vehicles pass more frequently, and the paths between these locations are more likely to be the main mobile paths.

[0079] Furthermore, the method of determining the main parking location of non-motor vehicles by using a clustering algorithm in combination with a distributed computing framework includes:

[0080] Read RFID data, extract the geographic location coordinates and residence time of the tag, and use the K-Means algorithm in the distributed computing framework to determine the number of clusters k according to the elbow rule. Then assign the data points to the cluster where the nearest cluster center is located to obtain preliminary clustering results. Based on the residence time, the clustering results are screened and optimized to obtain the main parking locations.

[0081] Specifically, the K-Means algorithm is used to divide data points into K clusters. In parking location analysis, the label data is clustered according to the geographic location coordinates, and each cluster represents a possible parking area. The basic idea of ​​the K-Means algorithm is to randomly select K initial cluster centers, then assign each data point to the cluster where the nearest cluster center is located, and then recalculate the center of each cluster, and repeat this process until the cluster center no longer changes or reaches a certain number of iterations.

[0082] Furthermore, the historical data is predicted by a time series analysis algorithm to obtain trend values ​​of the use frequency, movement path and parking location of non-motor vehicles in the future time period, including:

[0083] Time series data on the frequency of use, movement paths and parking locations of non-motor vehicles are collected, and the data are tested for stationarity using the unit root test method. The autoregressive order, number of differences and moving average order of the ARIMA model are determined based on the observed autocorrelation function and partial autocorrelation function. The prediction model is obtained by training the historical data using the ARIMA model with determined parameters.

[0084] Specifically, in the stationarity test, if the data is unstable, differential processing is performed to make it stable. The autocorrelation function and partial autocorrelation function can also be replaced by automatic parameter selection methods, such as AIC (Akaike Information Criterion) or BIC (Bayesian Information Criterion) to determine the parameters such as autoregressive order, difference number and moving average order.

[0085] Furthermore, the trend values ​​are combined with environmental features as inputs to a machine learning algorithm for training and prediction, including:

[0086] The trend value predicted by the ARIMA model is merged with the environmental characteristic data through the numpy library to generate a multidimensional array, wherein the environmental characteristic data includes: weather data, season data, holiday data and time data;

[0087] The merged data set is divided into training set, validation set and test set. The initial structure of the recurrent neural network algorithm is two layers, with 128 neurons in each layer. Softmax is used as the activation function. The dimension of the input layer matches the dimension of the merged trend value and environmental feature data. When predicting the frequency of use of non-motor vehicles, the dimension of the output layer is a scalar. When predicting the moving path or parking position, the dimension of the output layer is a vector or matrix.

[0088] The weights and biases of the recurrent neural network are initialized by the He initialization method. The training cycle is started through the training set according to the cross entropy loss function and the Adam optimization algorithm. The value of the loss function is calculated on the validation set to evaluate the model performance. The learning rate, number of layers or number of neurons are adjusted according to the performance. The best model parameters saved during the training process are loaded. The new trend value and environmental feature data are input into the recurrent neural network loaded with the best model parameters, and the prediction results are obtained through forward propagation.

[0089] The weather data includes temperature, humidity, precipitation, etc.; the time data includes: weekdays / weekends, different time periods, etc. Preprocess the environmental feature data, such as encoding the season data (such as spring = 1, summer = 2, autumn = 3, winter = 4), converting the holiday data into binary variables (holiday = 1, non-holiday = 0), etc. For the multidimensional array, if the trend value is a one-dimensional array and the environmental feature is a multidimensional matrix, expand the trend value to the same number of rows as the environmental feature matrix, and then splice it in the column dimension.

[0090] The training cycle is specifically as follows: randomly selecting a batch of data from the training set, inputting it into the recurrent neural network, calculating the output of the recurrent neural network through forward propagation, calculating the value of the loss function based on the output and the true label, calculating the gradient of the loss function with respect to the weight and bias of the recurrent neural network, and finally using an optimization algorithm to update the weight and bias of the recurrent neural network according to the gradient.

[0091] The learning rate, number of layers or number of neurons are all hyperparameters adjusted according to the performance on the validation set. When the performance on the validation set reaches the best, the model parameters at this time are saved. The trend value of the ARIMA model is combined with the environmental characteristics and input into the recurrent neural network algorithm for training and prediction, so as to more accurately predict the use of non-motor vehicles and provide decision support for urban traffic planning and parking resource allocation.

[0092] Furthermore, the blockchain layer is used to ensure the security and immutability of vehicle data, improve management transparency and user trust, including:

[0093] Interact and verify the key data of non-motor vehicles with the IOTA blockchain platform through the blockchain layer. The data on the chain includes: the unique identification, registration information, usage frequency, movement path and parking location of non-motor vehicles, and pre-process the data;

[0094] The nodes in the IOTA network will verify the newly uploaded transactions to ensure the integrity and authenticity of the data. If the data is tampered with or invalid, the transaction will be rejected. The management software sends a query request through the interface with the IOTA node, and the IOTA node returns the query result;

[0095] The data is encrypted before being uploaded to the chain, and only authorized users can decrypt and access the data. Access to the data on the blockchain is restricted by setting access permissions. Security audits are performed regularly on the data on the blockchain to detect any abnormal transactions or data tampering.

[0096] Specifically, IOTA is a new type of distributed ledger technology with the characteristics of no transaction fees, high scalability, and security and reliability. By adjusting the interface of RFID reading and writing equipment and management software, it can send data to the blockchain layer for processing. The data preprocessing is to format the data into JSON format on the blockchain and encrypt the data to improve the security of the data.

[0097] The specific integration of the IOTA blockchain platform is as follows: deploy one or more IOTA nodes in the system to interact with the IOTA network, choose to deploy nodes locally, or use the IOTA node service provided by a cloud service provider, and develop an interface with the IOTA node through the software development kit (SDK) provided by IOTA to implement data chain and query operations. For example, use Python's IOTA SDK to interact with the IOTA node. When the RFID reader reads the data of the non-motor vehicle, the management software sends the data to the blockchain layer. The blockchain layer packages the data into a transaction through the interface with the IOTA node and sends it to the IOTA network for verification.

[0098] In user data query, users can view vehicle history, usage and other information, improving management transparency and user trust. Only authorized users or devices can read and write data.

[0099] Furthermore, the automatic triggering of an alarm for a non-motor vehicle that moves illegally or enters a prohibited area and sending a real-time notification to the management department and the vehicle owner include:

[0100] According to the rules for determining areas where non-motor vehicles are prohibited from entering and illegal movement, additional RFID reading and writing devices are deployed around the prohibited areas and at key traffic nodes. When the RFID reading and writing devices read the tag information of the non-motor vehicle, they immediately transmit the data to the management system. The management system continuously monitors the location and movement trajectory of the vehicle, and determines whether the vehicle has entered the prohibited area by comparing it with the pre-set boundary coordinates of the prohibited area. It also determines whether there is illegal movement based on the vehicle's movement speed and direction information. If the system detects that the non-motor vehicle is moving illegally or entering a prohibited area, an alarm is triggered immediately.

[0101] Specifically, based on actual conditions, determine areas in the city where non-motor vehicles are prohibited from entering, such as specific pedestrian streets, squares, scenic spots, etc. The boundary coordinates of these areas can be accurately marked on the map through Geographic Information System (GIS) technology. Set up judgment rules for illegal movement, for example, rapid movement beyond the normal range in a short period of time may be regarded as illegal movement, or entering a prohibited area during a specific time period, etc. The additional RFID reading and writing equipment can monitor the movement of non-motor vehicles more intensively. The management system continuously monitors the location and movement trajectory of the vehicle. For each monitored non-motor vehicle, the system updates the location coordinates of the vehicle in real time based on its tag information and the location information of the reading device.

[0102] The alarm is immediately triggered, and an alarm signal is sent out through an audible and visual alarm device in the monitoring center of the management department to remind the staff to pay attention. At the same time, the system automatically generates alarm information, including the unique identification of the vehicle, the current location, the time, etc. For the management department, a notification is pushed through text messages, emails or special management software to inform relevant personnel that non-motor vehicles have violated regulations so that timely measures can be taken to deal with them. For car owners, notifications are sent through mobile phone text messages, mobile application pushes, etc. to remind the car owner that his vehicle is in an illegal state and require the car owner to deal with it as soon as possible. After receiving the notification, the management department dispatches law enforcement personnel to the scene for processing according to the alarm information. The law enforcement personnel can read the vehicle's tag information through a handheld device to confirm the vehicle's identity and violation. After receiving the notification, the car owner will promptly understand the situation of the vehicle and take corresponding measures, such as moving the vehicle out of the prohibited area or cooperating with the management department's processing.

[0103] The above implementation modes are merely descriptions of the preferred implementation modes of the present invention, and are not intended to limit the scope of the present invention. Without departing from the design spirit of the present invention, various modifications and improvements made to the technical solutions of the present invention by ordinary engineering and technical personnel in the field shall fall within the protection scope determined by the claims of the present invention.

Claims

1. A non-motor vehicle integrated management system based on ultra-high frequency RFID technology, characterized in that: It includes a radio frequency identification module, an intelligent prediction module, a blockchain module and an anti-theft alarm module connected by wireless communication, wherein: The radio frequency identification module is used to achieve life cycle management, real-time positioning and dynamic trajectory tracking of non-motor vehicles through ultra-high frequency RFID tags and reading and writing devices; The intelligent prediction module is used to intelligently analyze and predict the use frequency, movement path and parking location of non-motor vehicles, and optimize urban traffic planning and parking resource allocation; The blockchain module is used to ensure the security and immutability of vehicle data through the blockchain layer, improving management transparency and user trust; The anti-theft alarm module is used to automatically trigger an alarm for non-motor vehicles that are illegally moved or enter prohibited areas, and send real-time notifications to management departments and vehicle owners.

2. According to claim 1, a non-motor vehicle integrated management system based on ultra-high frequency RFID technology is characterized in that: The UHF RFID tags and read / write devices are used to achieve life cycle management, real-time positioning and dynamic trajectory tracking of non-motor vehicles; The following steps are involved: A1: The RFID reader / writer deployed at the main traffic nodes automatically reads the tag information bound to the vehicle information management system that stores the detailed information of non-motor vehicles within the identification range. The reader / writer transmits the read tag information to the vehicle information management system through the wireless network; A2: The vehicle information management system receives the tag information data transmitted by the reading and writing device for comparison, determines the identity and location information of the vehicle, updates the vehicle's location information in real time according to the tag information read by the reading and writing device, and displays the vehicle's current location on the map; A3: The vehicle owner queries the real-time location of the vehicle through the location information display terminal to obtain the vehicle's whereabouts, and records and stores the vehicle's location information at different time points to form a dynamic trajectory of the vehicle.

3. According to claim 2, a non-motor vehicle integrated management system based on ultra-high frequency RFID technology is characterized in that: The RFID reading and writing device automatically reads the tag information, including: the tag recognition algorithm determines the type and unique identification of the tag through feature extraction and pattern recognition of the tag signal; The determining of the position information of the vehicle includes: measuring the strength of the tag signal to estimate the distance between the vehicle and the reading and writing device through a positioning algorithm based on signal strength, and determining the position of the vehicle through a triangulation positioning or a multilateral positioning algorithm based on the position information of multiple reading and writing devices; The formation of the dynamic trajectory of the vehicle includes: recording and storing the position information of the vehicle at different time points through a trajectory tracking algorithm to form the dynamic trajectory of the vehicle; performing time series analysis on the vehicle position data to calculate the displacement and speed of the vehicle in different time periods to determine the driving direction and speed of the vehicle; and performing rationality analysis on the driving trajectory of the vehicle in combination with map data and traffic rules to determine abnormal behavior.

4. According to claim 1, a non-motor vehicle integrated management system based on ultra-high frequency RFID technology is characterized in that: The intelligent analysis and prediction of the use frequency, movement path and parking location of non-motor vehicles to optimize urban traffic planning and parking resource allocation includes the following steps: B1: Continuously collect non-motor vehicle tag information data through RFID reading and writing devices at major traffic nodes and parking areas, store the data according to the distributed framework, remove noise data, fill in missing values ​​and standardize the data; B2: Through the counting algorithm, the data is distributed to multiple computing nodes in a distributed environment. Each node counts the number of times the tag is read in a specific time period, and the results of each node are summarized to obtain the frequency of use of non-motor vehicles, and a frequency heat map is drawn; B3: Through the graph construction algorithm, the locations of the reading and writing devices in the city are regarded as the nodes of the graph, and the movement of tags between different nodes is regarded as the edges of the graph. The weight of the edge is determined according to the time sequence and time interval of the tags at different reading and writing device locations, and then the main movement paths of non-motor vehicles are analyzed through the graph analysis algorithm; B4: Determine the main parking locations of non-motor vehicles through clustering algorithm; B5: Use time series analysis algorithms to predict historical data and obtain trend values ​​of the usage frequency, movement paths, and parking locations of non-motor vehicles in future time periods. Combine these trend values ​​with environmental characteristics as input and input them into the machine learning algorithm for training and prediction.

5. The non-motor vehicle integrated management system based on UHF RFID technology according to claim 4 is characterized in that: The counting algorithm includes: extracting label data within a specific time period from the database and distributing it to each node of the distributed computing framework, running a mapper task on each computer point to count the assigned label data, and after the shuffle and sort stages, the redducer task aggregates the mapper results to obtain the final frequency; The graph analysis algorithm includes: analyzing the importance of nodes in the graph through the PageRank algorithm to obtain PageRank values, nodes with high PageRank values ​​represent locations where non-motor vehicles pass more frequently, and the paths between these locations are the main movement paths.

6. The non-motor vehicle integrated management system based on UHF RFID technology according to claim 4 is characterized in that: The method of determining the main parking location of non-motor vehicles by using a clustering algorithm in combination with a distributed computing framework includes: Read RFID data, extract the geographic location coordinates and residence time of the tag, and use the K-Means algorithm in the distributed computing framework to determine the number of clusters k according to the elbow rule. Then assign the data points to the cluster where the nearest cluster center is located to obtain preliminary clustering results. Based on the residence time, the clustering results are screened and optimized to obtain the main parking locations.

7. The non-motor vehicle integrated management system based on ultra-high frequency RFID technology according to claim 4 is characterized in that: The method of predicting historical data by using a time series analysis algorithm to obtain trend values ​​of the use frequency, movement path and parking location of non-motor vehicles in a future time period includes: Time series data on the frequency of use, movement paths and parking locations of non-motor vehicles are collected, and the data are tested for stationarity using the unit root test method. The autoregressive order, number of differences and moving average order of the ARIMA model are determined based on the observed autocorrelation function and partial autocorrelation function. The prediction model is obtained by training the historical data using the ARIMA model with determined parameters.

8. The non-motor vehicle integrated management system based on UHF RFID technology according to claim 4 is characterized in that: The trend values ​​are combined with environmental features as inputs to the machine learning algorithm for training and prediction, including: The trend value predicted by the ARIMA model is merged with the environmental characteristic data through the numpy library to generate a multidimensional array, wherein the environmental characteristic data includes: weather data, season data, holiday data and time data; The merged data set is divided into training set, validation set and test set. The initial structure of the recurrent neural network algorithm is two layers, with 128 neurons in each layer. Softmax is used as the activation function. The dimension of the input layer matches the dimension of the merged trend value and environmental feature data. When predicting the frequency of use of non-motor vehicles, the dimension of the output layer is a scalar. When predicting the moving path or parking position, the dimension of the output layer is a vector or matrix. The weights and biases of the recurrent neural network are initialized by the He initialization method. The training cycle is started through the training set according to the cross entropy loss function and the Adam optimization algorithm. The value of the loss function is calculated on the validation set to evaluate the model performance. The learning rate, number of layers or number of neurons are adjusted according to the performance. The best model parameters saved during the training process are loaded. The new trend value and environmental feature data are input into the recurrent neural network loaded with the best model parameters, and the prediction results are obtained through forward propagation.

9. The non-motor vehicle integrated management system based on ultra-high frequency RFID technology according to claim 1 is characterized in that: The blockchain layer is used to ensure the security and immutability of vehicle data, improve management transparency and user trust, including: Interact and verify the key data of non-motor vehicles with the IOTA blockchain platform through the blockchain layer. The data on the chain includes: the unique identification, registration information, usage frequency, movement path and parking location of non-motor vehicles, and pre-process the data; The nodes in the IOTA network will verify the newly uploaded transactions to ensure the integrity and authenticity of the data. If the data is tampered with or invalid, the transaction will be rejected. The management software sends a query request through the interface with the IOTA node, and the IOTA node returns the query result; The data is encrypted before being uploaded to the chain, and only authorized users can decrypt and access the data. Access to the data on the blockchain is restricted by setting access permissions. Security audits are performed regularly on the data on the blockchain to detect any abnormal transactions or data tampering.

10. The non-motor vehicle integrated management system based on ultra-high frequency RFID technology according to claim 1 is characterized in that: The automatic triggering of alarms for non-motor vehicles that are illegally moving or entering prohibited areas, and sending real-time notifications to management departments and vehicle owners, include: According to the rules for determining areas where non-motor vehicles are prohibited from entering and illegal movement, additional RFID reading and writing devices are deployed around prohibited areas and at key traffic nodes. When the RFID reading and writing devices read the tag information of non-motor vehicles, they immediately transmit the data to the management system. The management system continuously monitors the location and movement trajectory of the vehicle, and determines whether the vehicle has entered the prohibited area by comparing it with the pre-set boundary coordinates of the prohibited area. It also determines whether there is illegal movement based on the vehicle's movement speed and direction information. If the system detects that the non-motor vehicle is illegally moving or entering a prohibited area, an alarm is immediately triggered and a real-time notification is sent to the management department and the vehicle owner.

Images