Real-time vehicle-building cooperative communication data transmission and storage method and system

Abstract

The invention relates to the field of intelligent transportation, in particular to a real-time vehicle-building cooperative communication data transmission and storage method and system. A real-time vehicle-building cooperative communication data transmission and storage system comprises a vehicle owner classification module, a vehicle flow prediction module, an occupied parking space prediction module, a resource demand prediction module and a resource distribution module. According to the method, the initial resource demand index of each time period is calculated according to the predicted traffic flow and the occupied parking space, and the initial resource demand index is normalized to obtain the resource demand index; according to the resource demand indexes, resources used for transmission and storage are dynamically adjusted, so that when communication data is transmitted and stored, stable supply and demand balance can be kept, the situation of data transmission delay or insufficient storage space is effectively avoided, meanwhile, more resources can be saved, and optimal configuration of the resources is achieved; according to the method, the operation efficiency of the parking lot can be improved, the resource waste is reduced, and the user experience is improved.

Description

Technical Field

[0001] The present invention relates to the field of intelligent transportation, and in particular to a real-time vehicle-building collaborative communication data transmission and storage method and system. Background Art

[0002] With the acceleration of urbanization, urban traffic pressure is increasing, especially in commercial centers and residential areas, where parking difficulties are particularly prominent; traditional parking lot management methods often rely on manual management, which is inefficient and difficult to cope with peak traffic flow; in addition, parking lot resource allocation is usually based on experience rather than real-time data, resulting in uneven resource utilization and failure to meet the actual needs of car owners.

[0003] In the existing car-building collaborative communication systems, the predictions of vehicle flow and parking space occupancy are often not accurate enough, which is mainly due to the lack of effective data analysis and prediction models; most systems only rely on static rules or simple statistical methods and cannot adapt to complex and changeable traffic conditions; in addition, these systems also lack flexibility in resource allocation. Existing systems often adopt a fixed mode for resource allocation and cannot dynamically adjust resource allocation according to real-time data; while the situation in a parking lot is usually changeable, fixed resource allocation will cause the system to have excess or shortage of resources when transmitting and storing data generated by vehicles in the parking lot; in response to the above problems, the present invention proposes a real-time car-building collaborative communication data transmission and storage method and system, which aims to achieve accurate prediction of vehicle flow and parking space occupancy through advanced data analysis and prediction technology, and optimize resource allocation accordingly, so that when the communication system transmits and stores data, the allocated resources and the actual demand are in a state of equilibrium, thereby improving the operation efficiency and user experience of the parking lot. Summary of the invention

[0004] The present invention calculates the initial resource demand index for each time period according to the predicted vehicle flow and occupied parking spaces, and normalizes the initial resource demand index to obtain a resource demand index; dynamically adjusts the resources used for transmission and storage according to these resource demand indexes, so that when communication data is transmitted and stored, a relatively stable supply and demand balance can be maintained, effectively avoiding data transmission delays or insufficient storage space, while also saving more resources and achieving optimal resource allocation; this method can improve the operational efficiency of the parking lot, reduce resource waste, and enhance user experience.

[0005] A real-time vehicle-building collaborative communication data transmission and storage method, comprising: At the beginning of any update cycle, the historical parking data of all car owners in the previous update cycle are screened at multiple levels, and the car owners with regular behavior from Monday to Sunday are screened out by comparing with the first preset threshold, the second preset threshold and the third preset threshold. The car owners other than the car owners with regular behavior are ordinary car owners; the historical parking data includes the entry time data and the parking duration data; At the beginning of any day, for all car owners with behavioral patterns on that day, based on the historical traffic flow data of the car owners with behavioral patterns, the traffic flow of several time periods within the day is predicted, and the first traffic flow prediction data of each time period is obtained respectively; at the same time, according to the historical occupied parking space data of all car owners with behavioral patterns, the first occupied parking space prediction data of each time period of the day is calculated and obtained respectively; the historical traffic flow data and the historical occupied parking space data of ordinary car owners are used as the input of the traffic flow prediction model and the parking space prediction model respectively, and the second traffic flow prediction data and the second occupied parking space prediction data of each time period of the day are output; For any time period, the first vehicle flow prediction data obtained for the time period is combined with the second vehicle flow prediction data to obtain vehicle flow prediction data; the first occupied parking space prediction data obtained for the time period is combined with the second occupied parking space prediction data to obtain occupied parking space prediction data; Calculate the resource demand index for each time period based on the acquired traffic flow forecast data and occupied parking space data; The resource demand index of each time period is applied, and within the dispatchable range of the building resources, the linear interpolation method is used to adjust the resources used for transmission and storage in each time period.

[0006] Preferably, the historical parking data of the car owner is analyzed, and the specific operations are as follows: At the beginning of any update cycle, the historical parking data of all car owners in the previous update cycle are obtained, and the historical parking data include the entry time data and the parking duration data; an update cycle includes several weeks, and the historical parking data of all weeks marked with "Monday" are extracted. For any car owner, it is determined whether the ratio of the number of times the car owner parked on Monday to the total number of weeks in the update cycle is greater than the first preset threshold. If not, no operation is performed; if so, the formula is used Calculate the standard deviation A of the parking time of the car owner on Monday, where: represents the i-th parking duration data, i=1, 2, ..., n; is the average parking time of the car owner on Monday; determine whether the parking time standard deviation A of the car owner is less than the second preset threshold value, if not, no operation; if yes, use the formula Calculate the standard deviation B of the car owner's entry time on Monday, where: Represents the i-th entry time data; is the average entry time of the car owner on Monday; determine whether the standard deviation B of the entry time of the car owner is less than the third preset threshold, if not, do nothing; if so, classify the car owner as a regular behavior owner on Monday of the current update cycle; for the historical parking data with week mark information of "Tuesday", "Wednesday", ..., "Sunday", and so on, filter out the regular behavior owners from Tuesday to Sunday of the current update cycle respectively.

[0007] Preferably, based on the historical parking data of the car owner with behavioral patterns, the traffic flow in several time periods of the day is predicted, and the specific operations are as follows: For all the car owners with behavioral patterns on the day, obtain the historical traffic flow data with the same week mark information as the day in the previous N weeks, and form all the obtained historical traffic flow data into a historical traffic flow data set; For any time period of the day, historical traffic flow data in the current time period is selected from the historical traffic flow data set, and the average value of the selected historical traffic flow data is calculated to obtain the first traffic flow prediction data of the current time period; According to the historical parking space occupancy data of all car owners with behavioral patterns, the first occupied parking space prediction data of each time period of the day is calculated and obtained respectively. The specific operation is as follows: For all the car owners with regular behavior on that day, obtain the historical parking space occupied data of the car owners with regular behavior in the previous N weeks with the same week mark information as that of that day, and form all the obtained historical parking space occupied data into a historical parking space occupied data set; For any time period of the day, historical occupied parking space data in the current time period is selected from the historical occupied parking space data set, and the average value of the selected historical occupied parking space data is calculated to obtain the first occupied parking space prediction data of the current time period.

[0008] Preferably, based on the acquired traffic flow forecast data and occupied parking space data, the resource demand index for each time period is calculated, and the specific operations are as follows: For any time period of the day, use the formula Calculate and obtain the initial resource demand index F for this time period, where T represents the traffic flow forecast data for this time period. Indicates the prediction data of occupied parking spaces in this time period, Indicates the total number of parking spaces in the parking lot. and are the weights of traffic flow and proportion of occupied parking spaces, respectively. > ; Get the highest initial resource demand index in the last update cycle and minimum initial resource requirement index , using the formula Calculate and obtain the resource demand index H for the time period.

[0009] Preferably, the historical traffic flow data and historical occupied parking space data of ordinary car owners are used as inputs of the traffic flow prediction model and the parking space prediction model respectively. The specific operations are as follows: The historical traffic flow data and historical occupied parking space data of ordinary car owners in the previous 7 days are obtained respectively. For any time period of the day, the historical traffic flow data and historical occupied parking space data in the current time period are extracted from all the historical traffic flow data and historical occupied parking space data obtained respectively, and they are arranged in chronological order to form a traffic flow prediction set and an occupied parking space prediction set respectively. The traffic flow prediction set is used as the input of the traffic flow prediction model, and the second traffic flow prediction data in the time period of the day is output; the occupied parking space prediction set is used as the input of the parking space prediction model, and the second occupied parking space prediction data in the time period of the day is output.

[0010] Preferably, the vehicle flow prediction model is established based on the LSTM model, including 7 vehicle flow prediction LSTM units, 1 first fully connected layer and 1 first output layer, wherein the vehicle flow prediction LSTM unit is used to extract the temporal characteristics in the vehicle flow prediction set; the first fully connected layer is used to extract features from the outputs of all vehicle flow prediction LSTM units; the first output layer is used to output the second vehicle flow prediction data; For the training of the traffic flow prediction model, the specific operations are as follows: A number of traffic flow training samples are obtained, each of which stores 8 traffic flow data sorted by time, and the first 7 traffic flow data in the traffic flow training sample are used as the input of the traffic flow prediction model, and the last traffic flow data in the traffic flow training sample is used as the output of the traffic flow prediction model; all traffic flow training samples are divided into a first training set and a first test set, and the first training set is sent to the parameter-initialized traffic flow prediction model for training, and then the first test set is input to the traffic flow prediction model to obtain a first test result; a first training condition is set to determine whether the first test result meets the first training condition, and if so, the trained traffic flow prediction model is output; otherwise, the traffic flow prediction model is continuously trained through the first training set.

[0011] Preferably, the parking space prediction model is established based on the LSTM model, including 7 parking space prediction LSTM units, 1 second fully connected layer and 1 second output layer, wherein the parking space prediction LSTM unit is used to extract the temporal characteristics in the occupied parking space prediction set; the second fully connected layer is used to extract features from the outputs of all parking space prediction LSTM units; the second output layer is used to output the second occupied parking space prediction data; For the training of parking space prediction model, the specific operations are as follows: A plurality of occupied parking space training samples are obtained, each of which stores 8 occupied parking space data sorted by time, the first 7 occupied parking space data in the occupied parking space training samples are used as the input of the parking space prediction model, and the last occupied parking space data in the occupied parking space training samples are used as the output of the parking space prediction model; all occupied parking space training samples are divided into a second training set and a second test set, the second training set is sent to the parking space prediction model with initialized parameters for training, and then the second test set is input to the parking space prediction model to obtain a second test result; a second training condition is set to determine whether the second test result meets the second training condition, and if so, the trained parking space prediction model is output; otherwise, the parking space prediction model is continuously trained through the second training set.

[0012] Preferably, a real-time vehicle-building collaborative communication data transmission and storage system comprises: The car owner classification module is used to analyze the historical parking data of all car owners in the previous update cycle at the beginning of any update cycle. The historical parking data includes the entry time data and parking duration data, and select the car owners with regular behavior from Monday to Sunday; The vehicle flow prediction module is used to predict the vehicle flow in several time periods of the day according to the historical vehicle flow data of the vehicle owner with behavioral patterns, and obtain the first vehicle flow prediction data of each time period respectively; the historical vehicle flow data of ordinary vehicle owners is used as the input of the vehicle flow prediction model, and the second vehicle flow prediction data of each time period of the day is output; the first vehicle flow prediction data obtained is combined with the second vehicle flow prediction data to obtain the vehicle flow prediction data; The parking space occupancy prediction module is used to calculate and obtain the first parking space occupancy prediction data for each time period of the day according to the historical parking space occupancy data of all car owners with behavioral patterns; use the historical parking space occupancy data of ordinary car owners as the input of the parking space prediction model, and output the second parking space occupancy prediction data for each time period of the day; combine the obtained first parking space occupancy prediction data and the second parking space occupancy prediction data to obtain the parking space occupancy prediction data; The resource demand prediction module is used to calculate the resource demand index for each time period based on the acquired traffic flow prediction data and occupied parking space data; The resource allocation module is used to adjust the resources used for transmission and storage according to the resource demand index of each time period by using the linear interpolation method within the dispatchable range of the car building resources.

[0013] The present invention has the following advantages: 1. The present invention calculates the initial resource demand index for each time period according to the predicted vehicle flow and occupied parking spaces, and normalizes the initial resource demand index to obtain a resource demand index; dynamically adjusts the resources used for transmission and storage according to these resource demand indexes, so that when communication data is transmitted and stored, a relatively stable supply and demand balance can be maintained, effectively avoiding data transmission delays or insufficient storage space, while also saving more resources and achieving optimal resource allocation; this method can improve the operational efficiency of the parking lot, reduce resource waste, and enhance user experience.

[0014] 2. The present invention can accurately identify regular behavior owners and ordinary car owners by analyzing the historical parking data of car owners, including entry time and parking duration; based on the inconsistent behavior patterns of regular behavior owners and ordinary car owners, classification can be performed to ensure the accuracy of data prediction for regular behavior owners; and for ordinary car owners, the LSTM model is used for prediction, which can adapt to their nonlinear change trend; by executing targeted prediction methods for the traffic flow and parking space occupancy of the two types of car owners in a specific time period, more accurate prediction can be achieved, so that the parking lot can manage traffic flow and parking space resources more effectively. BRIEF DESCRIPTION OF THE DRAWINGS

[0015] Figure 1 This is a schematic diagram of the structure of the real-time vehicle-building collaborative communication data transmission and storage system adopted in an embodiment of the present invention. DETAILED DESCRIPTION

[0016] In order to enable persons skilled in the art to better understand the technical solutions in the present invention, the technical solutions in the embodiments of the present invention will be clearly and completely described below in conjunction with the accompanying drawings in the embodiments of the present invention.

[0017] Embodiment 1, a method for transmitting and storing real-time vehicle-building collaborative communication data, comprising: At the beginning of any update cycle, the historical parking data of all car owners in the previous update cycle are analyzed. By analyzing these data, the system can understand the parking habits and needs of car owners, and provide basic information for subsequent predictions and resource allocation. The historical parking data includes entry time data and parking duration data, and the car owners with regular behavior from Monday to Sunday are screened out respectively. Other car owners except for the car owners with regular behavior are ordinary car owners, whose parking behaviors have no obvious regularity, or the data is insufficient to judge their behavior patterns; the update cycle refers to a set time interval, which can be set to about 8 weeks here, for regular analysis and update of data; because the situation in the parking lot will continue to change, at the beginning of each update cycle, the system will start a new round of data analysis process to ensure that the car owners with regular behavior in the current update cycle are reasonably screened; identifying car owners with regular behavior helps the system more accurately predict the parking needs of these car owners on specific days in the future, so as to prepare and dispatch resources in advance; At the beginning of any day, for all car owners with behavioral patterns on that day, based on the historical traffic flow data of the car owners with behavioral patterns, the traffic flow of several time periods within the day is predicted, and the first traffic flow prediction data of each time period is obtained respectively; since the parking pattern of car owners with behavioral patterns is relatively stable, a simple method can be used to make accurate predictions. By predicting the traffic flow of car owners with behavioral patterns, the system can prepare corresponding parking resources in advance, thereby improving the operation efficiency of the parking lot and the satisfaction of car owners; at the same time, according to the historical occupied parking space data of all car owners with behavioral patterns, the first occupied parking space prediction data of each time period of the day is calculated and obtained respectively, and the historical traffic flow data and historical occupied parking space data of ordinary car owners are used as the input of the traffic flow prediction model and the parking space prediction model respectively, and the second traffic flow prediction data and the second occupied parking space prediction data of each time period of the day are output. Since the parking behavior of ordinary cars has no fixed pattern, it is difficult to predict directly by mathematical methods. In order to achieve more accurate prediction of the nonlinear data of ordinary car owners, a deep learning model is used here; accurate parking space occupancy prediction helps the system to better manage parking space resources, optimize parking space allocation, reduce parking space vacancy rate, and improve parking lot utilization; For any time period, the first vehicle flow prediction data obtained for the time period is combined with the second vehicle flow prediction data to obtain vehicle flow prediction data; the first occupied parking space prediction data obtained for the time period is combined with the second occupied parking space prediction data to obtain occupied parking space prediction data; Based on the acquired traffic flow prediction data and occupied parking space data, the resource demand index of each time period is calculated; the present invention takes into account the two effects of traffic flow and the number of occupied parking spaces, rather than simply judging by one data; for example, when the traffic flow is small but too many parking spaces are occupied, although the traffic flow is small, after the car enters the parking lot, it will be difficult to find a parking space due to the small number of available parking spaces. At this time, long-term real-time positioning and parking guidance of the vehicle will increase the resources required for data transmission and storage; if only the traffic flow is considered here, it cannot show the actual situation; The resource demand index of each time period is applied, and within the dispatchable range of the building resources, the linear interpolation method is used to adjust the resources used for transmission and storage in each time period; by applying the resource demand index, the system can realize dynamic adjustment of resources to ensure that there are enough resources available during the peak demand period, and the resources can be released or reallocated during the trough demand period to improve the efficiency of resource use and economic benefits; linear interpolation is a mathematical method used to estimate the value of an unknown point between two known points; in the present invention, the linear interpolation method is used to make a smooth transition between the resource demands of different time periods to realize the reasonable allocation of resources.

[0018] Analyze the historical parking data of the car owner. The specific operations are as follows: Set an update cycle. At the beginning of any update cycle, obtain the historical parking data of all car owners in the previous update cycle. The historical parking data includes the entry time data and the parking duration data. An update cycle includes several weeks. Extract the historical parking data of all weeks marked with "Monday". For any car owner, determine whether the ratio of the number of times the car owner parked on Monday to the total number of weeks in the update cycle is greater than the first preset threshold. If not, no operation is performed. If so, use the formula Calculate the standard deviation A of the parking time of the car owner on Monday, where: represents the i-th parking duration data, i=1, 2, ..., n; is the average parking time of the car owner on Monday; determine whether the parking time standard deviation A of the car owner is less than the second preset threshold value, if not, no operation; if yes, use the formula Calculate the standard deviation B of the car owner's entry time on Monday, where: Represents the i-th entry time data; is the average entry time of the car owner on Monday; determine whether the standard deviation B of the entry time of the car owner is less than the third preset threshold, if not, do nothing; if so, classify the car owner as a regular behavior owner on Monday of the current update cycle; for the historical parking data with week mark information of "Tuesday", "Wednesday", ..., "Sunday", and so on, filter out the regular behavior owners from Tuesday to Sunday of the current update cycle respectively.

[0019] Based on the historical parking data of the car owner with behavioral patterns, the traffic flow is predicted for several time periods within the day. The specific operations are as follows: For all the car owners with behavioral patterns on the day, obtain the historical traffic flow data with the same week mark information as the day in the previous N weeks, and form all the obtained historical traffic flow data into a historical traffic flow data set; For any time period of the day, historical traffic flow data in the current time period is selected from the historical traffic flow data set, and the average value of the selected historical traffic flow data is calculated to obtain the first traffic flow prediction data of the current time period; According to the historical parking space occupancy data of all car owners with behavioral patterns, the first occupied parking space prediction data of each time period of the day is calculated and obtained respectively. The specific operation is as follows: For all the car owners with regular behavior on that day, obtain the historical parking space occupied data of the car owners with regular behavior in the previous N weeks with the same week mark information as that of that day, and form all the obtained historical parking space occupied data into a historical parking space occupied data set; For any time period of the day, historical occupied parking space data in the current time period is selected from the historical occupied parking space data set, and the average value of the selected historical occupied parking space data is calculated to obtain the first occupied parking space prediction data of the current time period.

[0020] Based on the obtained traffic flow forecast data and occupied parking space data, the resource demand index for each time period is calculated. The specific operations are as follows: For any time period of the day, use the formula Calculate and obtain the initial resource demand index F for this time period, where T represents the traffic flow forecast data for this time period. Indicates the prediction data of occupied parking spaces in this time period, Indicates the total number of parking spaces in the parking lot. and are the weights of traffic flow and proportion of occupied parking spaces, respectively. > , here is because the impact of traffic flow on required resources is still greater than that of parking spaces; get the highest initial resource demand index in the last update cycle and minimum initial resource requirement index , using the formula Calculate and obtain the resource demand index H for the time period.

[0021] The historical traffic flow data and historical occupied parking space data of ordinary car owners are used as the input of the traffic flow prediction model and the parking space prediction model respectively. The specific operations are as follows: The historical traffic flow data and historical occupied parking space data of ordinary car owners in the previous 7 days are obtained respectively. For any time period of the day, the historical traffic flow data and historical occupied parking space data in the current time period are extracted from all the historical traffic flow data and historical occupied parking space data obtained respectively, and they are arranged in chronological order to form a traffic flow prediction set and an occupied parking space prediction set respectively. The traffic flow prediction set is used as the input of the traffic flow prediction model, and the second traffic flow prediction data in the time period of the day is output; the occupied parking space prediction set is used as the input of the parking space prediction model, and the second occupied parking space prediction data in the time period of the day is output.

[0022] The traffic flow prediction model is established based on the LSTM model, including 7 traffic flow prediction LSTM units, 1 first fully connected layer and 1 first output layer. The traffic flow prediction LSTM unit is used to extract the temporal features in the traffic flow prediction set; the first fully connected layer is used to extract features from the outputs of all traffic flow prediction LSTM units; the first output layer is used to output the second traffic flow prediction data; For the training of the traffic flow prediction model, the specific operations are as follows: A number of traffic flow training samples are obtained, each of which stores 8 traffic flow data sorted by time, and the first 7 traffic flow data in the traffic flow training sample are used as the input of the traffic flow prediction model, and the last traffic flow data in the traffic flow training sample is used as the output of the traffic flow prediction model; all traffic flow training samples are divided into a first training set and a first test set, and the first training set is sent to the parameter-initialized traffic flow prediction model for training, and then the first test set is input to the traffic flow prediction model to obtain a first test result; a first training condition is set to determine whether the first test result meets the first training condition, and if so, the trained traffic flow prediction model is output; otherwise, the traffic flow prediction model is continuously trained through the first training set.

[0023] The parking space prediction model is established based on the LSTM model, including 7 parking space prediction LSTM units, 1 second fully connected layer and 1 second output layer. The parking space prediction LSTM unit is used to extract the temporal features in the occupied parking space prediction set; the second fully connected layer is used to extract features from the outputs of all parking space prediction LSTM units; the second output layer is used to output the second occupied parking space prediction data; For the training of parking space prediction model, the specific operations are as follows: A plurality of occupied parking space training samples are obtained, each of which stores 8 occupied parking space data sorted by time, the first 7 occupied parking space data in the occupied parking space training samples are used as the input of the parking space prediction model, and the last occupied parking space data in the occupied parking space training samples are used as the output of the parking space prediction model; all occupied parking space training samples are divided into a second training set and a second test set, the second training set is sent to the parking space prediction model with initialized parameters for training, and then the second test set is input to the parking space prediction model to obtain a second test result; a second training condition is set to determine whether the second test result meets the second training condition, and if so, the trained parking space prediction model is output; otherwise, the parking space prediction model is continuously trained through the second training set.

[0024] Embodiment 2, a real-time vehicle-building collaborative communication data transmission and storage system, such as Figure 1 As shown, including: The car owner classification module is used to analyze the historical parking data of all car owners in the previous update cycle at the beginning of any update cycle. The historical parking data includes the entry time data and parking duration data, and select the car owners with regular behavior from Monday to Sunday; The vehicle flow prediction module is used to predict the vehicle flow in several time periods of the day according to the historical vehicle flow data of the vehicle owner with behavioral patterns, and obtain the first vehicle flow prediction data of each time period respectively; the historical vehicle flow data of ordinary vehicle owners is used as the input of the vehicle flow prediction model, and the second vehicle flow prediction data of each time period of the day is output; the first vehicle flow prediction data obtained is combined with the second vehicle flow prediction data to obtain the vehicle flow prediction data; The parking space occupancy prediction module is used to calculate and obtain the first parking space occupancy prediction data for each time period of the day according to the historical parking space occupancy data of all car owners with behavioral patterns; use the historical parking space occupancy data of ordinary car owners as the input of the parking space prediction model, and output the second parking space occupancy prediction data for each time period of the day; combine the obtained first parking space occupancy prediction data and the second parking space occupancy prediction data to obtain the parking space occupancy prediction data; The resource demand prediction module is used to calculate the resource demand index for each time period based on the acquired traffic flow prediction data and occupied parking space data; The resource allocation module is used to adjust the resources used for transmission and storage according to the resource demand index of each time period by using the linear interpolation method within the dispatchable range of the car building resources.

[0025] It should be understood that those skilled in the art can make improvements or changes based on the above description, and all such improvements and changes should fall within the scope of protection of the appended claims of the present invention. Parts not described in detail in this specification belong to the prior art known to those skilled in the art.

Claims

1. A real-time vehicle-building collaborative communication data transmission and storage method, characterized in that: include: At the beginning of any update cycle, the historical parking data of all car owners in the previous update cycle are screened at multiple levels, and the car owners with regular behavior from Monday to Sunday are screened out by comparing with the first preset threshold, the second preset threshold and the third preset threshold. The car owners other than the car owners with regular behavior are ordinary car owners; the historical parking data includes the entry time data and the parking duration data; At the beginning of any day, for all the car owners with behavioral patterns on that day, based on the historical traffic flow data of the car owners with behavioral patterns, the traffic flow of several time periods within the day is predicted, and the first traffic flow prediction data of each time period is obtained respectively; at the same time, based on the historical occupied parking space data of all the car owners with behavioral patterns, the first occupied parking space prediction data of each time period of the day is calculated and obtained respectively; The historical traffic flow data and historical occupied parking space data of ordinary car owners are used as inputs of the traffic flow prediction model and the parking space prediction model respectively, and the second traffic flow prediction data and the second occupied parking space prediction data of each time period of the day are output; For any time period, the first vehicle flow prediction data obtained for the time period is combined with the second vehicle flow prediction data to obtain vehicle flow prediction data; the first occupied parking space prediction data obtained for the time period is combined with the second occupied parking space prediction data to obtain occupied parking space prediction data; Calculate the resource demand index for each time period based on the acquired traffic flow forecast data and occupied parking space data; The resource demand index of each time period is applied, and within the dispatchable range of the building resources, the linear interpolation method is used to adjust the resources used for transmission and storage in each time period.

2. A real-time vehicle-building collaborative communication data transmission and storage method according to claim 1, characterized in that: Analyze the historical parking data of the car owner. The specific operations are as follows: At the beginning of any update cycle, the historical parking data of all car owners in the previous update cycle are obtained, and the historical parking data include the entry time data and the parking duration data; an update cycle includes several weeks, and the historical parking data of all weeks marked with "Monday" are extracted. For any car owner, it is determined whether the ratio of the number of times the car owner parked on Monday to the total number of weeks in the update cycle is greater than the first preset threshold. If not, no operation is performed; if so, the formula is used Calculate the standard deviation A of the parking time of the car owner on Monday, where: represents the i-th parking duration data, i=1, 2, ..., n; is the average parking time of the car owner on Monday; determine whether the parking time standard deviation A of the car owner is less than the second preset threshold value, if not, no operation; if yes, use the formula Calculate the standard deviation B of the car owner's entry time on Monday, where: Represents the i-th entry time data; is the average entry time of the car owner on Monday; determine whether the standard deviation B of the entry time of the car owner is less than the third preset threshold, if not, do nothing; if so, classify the car owner as a regular behavior owner on Monday of the current update cycle; for the historical parking data with week mark information of "Tuesday", "Wednesday", ..., "Sunday", and so on, filter out the regular behavior owners from Tuesday to Sunday of the current update cycle respectively.

3. A real-time vehicle-building collaborative communication data transmission and storage method according to claim 2, characterized in that: Based on the historical parking data of the car owner with behavioral patterns, the traffic flow is predicted for several time periods within the day. The specific operations are as follows: For all the car owners with behavioral patterns on the day, obtain the historical traffic flow data with the same week mark information as the day in the previous N weeks, and form all the obtained historical traffic flow data into a historical traffic flow data set; For any time period of the day, historical traffic flow data in the current time period is selected from the historical traffic flow data set, and the average value of the selected historical traffic flow data is calculated to obtain the first traffic flow prediction data of the current time period; According to the historical parking space occupancy data of all car owners with behavioral patterns, the first occupied parking space prediction data of each time period of the day is calculated and obtained respectively. The specific operation is as follows: For all the car owners with regular behavior on that day, obtain the historical parking space occupied data of the car owners with regular behavior in the previous N weeks with the same week mark information as that of that day, and form all the obtained historical parking space occupied data into a historical parking space occupied data set; For any time period of the day, historical occupied parking space data in the current time period is selected from the historical occupied parking space data set, and the average value of the selected historical occupied parking space data is calculated to obtain the first occupied parking space prediction data of the current time period.

4. A real-time vehicle-building collaborative communication data transmission and storage method according to claim 3, characterized in that: Based on the obtained traffic flow forecast data and occupied parking space data, the resource demand index for each time period is calculated. The specific operations are as follows: For any time period of the day, use the formula Calculate and obtain the initial resource demand index F for this time period, where T represents the traffic flow forecast data for this time period. Indicates the prediction data of occupied parking spaces in this time period, Indicates the total number of parking spaces in the parking lot. and are the weights of traffic flow and proportion of occupied parking spaces, respectively. > ; Get the highest initial resource demand index in the last update cycle and minimum initial resource requirement index , using the formula Calculate and obtain the resource demand index H for the time period.

5. A real-time vehicle-building collaborative communication data transmission and storage method according to claim 4, characterized in that: The historical traffic flow data and historical occupied parking space data of ordinary car owners are used as the input of the traffic flow prediction model and the parking space prediction model respectively. The specific operations are as follows: The historical traffic flow data and historical occupied parking space data of ordinary car owners in the previous 7 days are obtained respectively. For any time period of the day, the historical traffic flow data and historical occupied parking space data in the current time period are extracted from all the historical traffic flow data and historical occupied parking space data obtained respectively, and they are arranged in chronological order to form a traffic flow prediction set and an occupied parking space prediction set respectively. The traffic flow prediction set is used as the input of the traffic flow prediction model, and the second traffic flow prediction data in the time period of the day is output; the occupied parking space prediction set is used as the input of the parking space prediction model, and the second occupied parking space prediction data in the time period of the day is output.

6. A method for transmitting and storing real-time vehicle-building collaborative communication data according to claim 5, characterized in that: The traffic flow prediction model is established based on the LSTM model, including 7 traffic flow prediction LSTM units, 1 first fully connected layer and 1 first output layer. The traffic flow prediction LSTM unit is used to extract the temporal features in the traffic flow prediction set; the first fully connected layer is used to extract features from the outputs of all traffic flow prediction LSTM units; the first output layer is used to output the second traffic flow prediction data; For the training of the traffic flow prediction model, the specific operations are as follows: A plurality of traffic flow training samples are obtained, each of which stores 8 traffic flow data sorted by time, and the first 7 traffic flow data in the traffic flow training sample are used as inputs of the traffic flow prediction model, and the last traffic flow data in the traffic flow training sample is used as outputs of the traffic flow prediction model; all traffic flow training samples are divided into a first training set and a first test set, the first training set is sent to the parameter-initialized traffic flow prediction model for training, and the first test set is then input to the traffic flow prediction model to obtain a first test result; A first training condition is set, and it is determined whether the first test result satisfies the first training condition. If the first training condition is satisfied, the trained vehicle flow prediction model is output; otherwise, the vehicle flow prediction model is continuously trained using the first training set.

7. A method for transmitting and storing real-time vehicle-building collaborative communication data according to claim 6, characterized in that: The parking space prediction model is built based on the LSTM model, which includes 7 parking space prediction LSTM units, 1 second fully connected layer and 1 second output layer. The parking space prediction LSTM unit is used to extract the temporal features of the occupied parking space prediction set. The second fully connected layer is used to extract features from the outputs of all parking space prediction LSTM units; the second output layer is used to output the second occupied parking space prediction data; For the training of parking space prediction model, the specific operations are as follows: Obtain several occupied parking space training samples, each of which stores 8 occupied parking space data sorted by time, and use the first 7 occupied parking space data in the occupied parking space training sample as the input of the parking space prediction model, and use the last 1 occupied parking space data in the occupied parking space training sample as the output of the parking space prediction model; Dividing all occupied parking space training samples into a second training set and a second test set, sending the second training set to the parking space prediction model with initialized parameters for training, and then inputting the second test set to the parking space prediction model to obtain a second test result; A second training condition is set to determine whether the second test result satisfies the second training condition. If so, the trained parking space prediction model is output; otherwise, the parking space prediction model is continuously trained using the second training set.

8. A real-time vehicle-building collaborative communication data transmission and storage system, characterized in that: The system is applied to the real-time vehicle-building collaborative communication data transmission and storage method according to any one of claims 1 to 7, including: The car owner classification module is used to analyze the historical parking data of all car owners in the previous update cycle at the beginning of any update cycle. The historical parking data includes the entry time data and parking duration data, and select the car owners with regular behavior from Monday to Sunday; The vehicle flow prediction module is used to predict the vehicle flow in several time periods of the day according to the historical vehicle flow data of the vehicle owner with behavioral patterns, and obtain the first vehicle flow prediction data of each time period respectively; the historical vehicle flow data of ordinary vehicle owners is used as the input of the vehicle flow prediction model, and the second vehicle flow prediction data of each time period of the day is output; the first vehicle flow prediction data obtained is combined with the second vehicle flow prediction data to obtain the vehicle flow prediction data; The parking space occupancy prediction module is used to calculate and obtain the first parking space occupancy prediction data for each time period of the day according to the historical parking space occupancy data of all car owners with behavioral patterns; use the historical parking space occupancy data of ordinary car owners as the input of the parking space prediction model, and output the second parking space occupancy prediction data for each time period of the day; combine the obtained first parking space occupancy prediction data and the second parking space occupancy prediction data to obtain the parking space occupancy prediction data; The resource demand prediction module is used to calculate the resource demand index for each time period based on the acquired traffic flow prediction data and occupied parking space data; The resource allocation module is used to adjust the resources used for transmission and storage according to the resource demand index of each time period by using the linear interpolation method within the dispatchable range of the car building resources.

Images