Parking lot guidance system, device and method based on vehicle AI platform

Abstract

The present invention relates to the field of digital data processing technology, and more specifically to a parking space guidance system, device, and method based on a vehicle AI platform. The system comprises: dividing a parking lot into grids, obtaining parking space-related data for each grid, and obtaining urban weather and road condition data surrounding the parking lot; determining the unit occupancy impact of each grid in each time interval based on the correlation between various types of data in the parking space-related data and the changing trends of local data; and constructing a parking space availability index for each grid in each time interval based on the discreteness of the urban weather and road condition data; training an autoregressive moving average model based on the parking space availability indexes for each historical time interval; and predicting the parking space availability index for each grid in the next time interval based on the trained autoregressive moving average model. The grid with the highest predicted value is recommended to the driver. This system achieves parking space guidance, improves the accuracy of parking space availability probability assessment, and optimizes the customer's parking experience.

Description

Technical Field

[0001] The present application relates to the field of digital data processing technology, and specifically to a parking space guidance system, device, and method based on a vehicle AI platform. Background Art

[0002] As a key application of the Internet of Vehicles in urban traffic management, intelligent parking guidance has future development prospects including more accurate parking detection technology, intelligent navigation algorithms, real-time data sharing, and integration with autonomous driving technology. In addition, the development of intelligent parking guidance systems is expected to promote more efficient planning of urban space. Through data analysis and real-time monitoring, city managers can better understand parking needs and optimize parking resource allocation.

[0003] The use of data mining algorithms can enable the intelligent parking guidance system to better understand and adapt to the city's parking needs, thereby achieving more effective parking resource management and improving the overall efficiency of the transportation system. However, there are too many uncertainties in the city, such as time factors, rush hour factors, holiday factors, and weather factors. Various reasons will cause parking demand in different areas to change, which will affect the effectiveness of data mining and need to be improved. Summary of the Invention

[0004] In order to solve the above technical problems, the present invention provides a parking space guidance system, device and method based on a vehicle AI platform to solve the existing problems.

[0005] The parking space guidance method based on the vehicle AI platform of the present invention adopts the following technical solutions:

[0006] One embodiment of the present invention provides a parking space guidance method based on a vehicle AI platform, the method comprising the following steps:

[0007] The parking lot is divided into grids to obtain each grid, with a first preset time duration as a time interval, a statistical moment set every time a second preset time duration passes, and a third preset time duration as a cycle; within each time interval, parking space occupancy data for each grid is obtained based on the parking situation of all parking spaces in each grid at each statistical moment; parking intention data for each grid is obtained based on the area and facilities of non-parking spaces in each grid; urban road conditions and weather data around the parking lot are obtained; and the parking space occupancy data and parking intention data for each grid are used as parking space-related data for each grid;

[0008] In each period, the time series of various parking space-related data are divided into subsequences using the K-means clustering algorithm. The time period comprehensive impact factor of each grid is obtained based on the data changes in each subsequence and the correlation between parking data and parking intention data. The unit occupancy impact of each grid in each time interval is obtained based on the time period comprehensive impact factor and the numerical changes in each subsequence. The vehicle use impact index is obtained based on the discrete situation of various urban road conditions and weather data. The parking space vacancy index of each grid in each time interval is obtained based on the unit occupancy impact and the vehicle use impact index.

[0009] An autoregressive moving average model is trained based on the parking space vacancy index of each grid in each time interval. The parking space vacancy index of each grid in the next time interval is predicted based on the trained autoregressive moving average model, and the grid with the largest predicted value is recommended to the car owner, thus completing the guidance of the car owner's parking space selection.

[0010] Preferably, the parking space parking data of each grid is obtained according to the parking situation of all parking spaces at each statistical moment in each grid, specifically:

[0011] The parking data of each grid includes parking space occupancy rate, number of vehicles leaving, number of vehicles entering, average parking space occupancy time and average vehicle parking time;

[0012] For each grid, the parking space occupancy rate is as follows: obtaining the number of occupied parking spaces in the grid at each statistical moment, calculating the average of the number at all statistical moments; and taking the ratio of the average to the total number of parking spaces in the grid as the parking space occupancy rate of the grid;

[0013] The number of vehicles dispatched is as follows: the number of times each parking space changes from occupied to unoccupied is counted, which is recorded as the number of vehicles dispatched; the sum of the number of vehicles dispatched from all parking spaces is taken as the number of vehicles dispatched from the grid;

[0014] The number of cars entering the grid is calculated by counting the number of times each parking space changes from an unoccupied state to an occupied state, and recording the number of cars entering the grid; the sum of the number of cars entering all parking spaces is taken as the number of cars entering the grid;

[0015] The average parking space occupancy time is: obtain the total occupied time of each parking space; and take the average of the total occupied time of all parking spaces as the average parking space occupancy time of the grid;

[0016] The average parking time of a vehicle is as follows: the continuous occupation time of each parking space is obtained, and the average of the continuous occupation time of all parking spaces is used as the average parking time of the vehicle in the grid.

[0017] Preferably, the parking intention data of each grid is obtained based on the area and facilities of the non-parking spaces of each grid, specifically:

[0018] The parking intention data for each grid include the average vacant area and number of lanes;

[0019] For each grid, the average free area is calculated by: calculating the sum of the areas of all parking spaces that are not occupied at each statistical moment; calculating the average of the sum of the areas at all statistical moments; and taking the sum of the average and the area of ​​the non-parking spaces as the average free area of ​​the grid;

[0020] The number of channels is: the number of parking lot entrances, exits, elevators and corridor doors in the grid that are open is taken as the number of channels in the grid.

[0021] Preferably, the acquisition of urban road conditions and weather data around the parking lot is specifically as follows:

[0022] Obtain the minute-by-minute rainfall and total rainfall duration around the parking lot, as well as the minute-by-minute average sunshine intensity around the parking lot. Collect the minute-by-minute traffic volume at each intersection near the parking lot. The average of the minute-by-minute traffic volume at all intersections is used as the traffic volume around the parking lot.

[0023] The rainfall, total rainfall time, average sunshine intensity and traffic volume around the parking lot are used as urban road conditions and weather data.

[0024] Preferably, the comprehensive influencing factor of each grid period is obtained based on the data changes in each sub-sequence and the correlation between parking space data and parking intention data, specifically including:

[0025] Calculate the Pearson correlation coefficient between the time series of parking data of type i parking spaces and parking intention data of type j in each grid ; Arrange all subsequences of the time series of parking data of each type of parking space in each grid;

[0026] The expression for calculating the comprehensive impact factor of each grid period is:

[0027]

[0028]

[0029] in, is the comprehensive impact factor of the time period of the s-th grid, representing the overall impact of different time periods within a day and the public facilities conditions within the grid on the parking situation of the s-th grid; is the parking impact of the s-th grid; The number of types of parking data; is the number of types of parking intention data; K is the number of clusters obtained by clustering the parking data of each type of parking space; To find the maximum function; is the parking data of the i-th type of parking space The value of the last data point in the sequence, is the value of the first data point in the kth subsequence of parking data of the i-th type of parking space, is the value of the last data point in the kth subsequence of parking data of the i-th type of parking space, The parking data of the i-th type of parking space The value of the first data point in the sequence; is the number of all data points in the kth subsequence of parking data of the i-th type of parking space; is the average number of data points in the kth subsequence of all types of parking space data, Parking data for the i-th parking space and the parking intention data of category j Pearson correlation coefficient between the time series.

[0030] Preferably, obtaining the unit occupancy impact of each grid in each time interval according to the comprehensive impact factor of the time period and the value change in each sub-sequence specifically includes:

[0031] The first and last data points in each subsequence of each type of parking space-related data are taken as boundary points; the absolute value of the difference between the serial numbers of each data point in each subsequence and the nearest boundary point is recorded as the boundary distance of each data point;

[0032] In the time series of various parking space related data, the fitted straight line equation obtained by linear fitting all data points in the neighborhood of each data point is used as the fitted straight line equation of each data point;

[0033] The expression for calculating the unit occupancy impact of each grid in each time interval is:

[0034]

[0035] in, is the unit occupancy impact of the sth grid in the dth time interval, which expresses the degree to which the parking space occupancy of the sth grid in the dth time interval is affected by various situations. is the comprehensive impact factor of the time period of the s-th grid, is the number of types of parking space related data, is the value of the data related to the c-th type parking space in the d-th time interval of the grid, for The boundary distance of the corresponding data point, is the maximum value among all data related to the c-th type parking space. for The slope of the fitted line equation corresponding to the data points.

[0036] Preferably, the vehicle use impact index is obtained according to the discrete conditions of the road conditions and weather data of various cities, specifically including:

[0037] Obtain the dispersion coefficient of all data in the road conditions and weather data of various cities; the expression for calculating the vehicle use impact index based on the dispersion coefficient of the road conditions and weather data of various cities is:

[0038]

[0039] in, is the car use impact index, is the number of types of urban traffic and weather data, n is the number of time intervals in a cycle, is the dispersion coefficient of the f-th type of urban road conditions and weather data in the g-th time interval.

[0040] Preferably, obtaining the parking space vacancy index of each grid in each time interval according to the unit occupancy impact and the vehicle use impact index specifically includes:

[0041] Calculate the ratio of the unit occupancy impact to the vehicle use impact index of each grid in each time interval; and use the calculation result of an exponential function with a natural constant as the base and the inverse of the ratio as the exponent as the parking space vacancy index of each grid in each time interval.

[0042] The present invention also proposes a parking space guidance system based on a vehicle AI platform, comprising a memory, a processor, and a computer program stored in the memory and running on the processor. When the processor executes the computer program, it implements any step of the parking space guidance method based on the vehicle AI platform.

[0043] An embodiment of the present application also provides a parking space guidance device based on a vehicle AI platform, the device including a memory, a processor, and a computer program stored in the memory and executable on the processor. When the processor executes the computer program, the steps of the parking space guidance method based on the vehicle AI platform are implemented.

[0044] The present invention has at least the following beneficial effects:

[0045] The present invention proposes a parking space guidance method based on a vehicle AI platform. A time period comprehensive environmental factor is constructed based on parking space-related data combined with K-means clustering, which can analyze the degree of influence of the differences between the parking space environment and the surrounding environment of the parking lot in different time periods of a day on parking space occupancy. Then, a unit occupancy impact is constructed based on the linear fitting results of the data in the time window, which represents the comprehensive situation of the changes in parking space occupancy due to various influences within a time interval. A vehicle use influence index is constructed based on the discrete situation of urban road conditions and weather data, which further reflects its influence on the probability of car owners using their cars. Finally, a parking space vacancy index of each grid in a time interval is constructed by combining the vehicle use influence index and the unit occupancy impact. The ARIMA algorithm is trained according to the parking space vacancy index of each grid in each time interval in each historical period. The parking space vacancy index of each grid in the next time interval is predicted according to the trained autoregressive moving average model to obtain the grid with the largest predicted value. The grid with the largest predicted value is recommended to the car owner, thereby optimizing the customer's parking experience and enhancing the robustness of the algorithm. BRIEF DESCRIPTION OF THE DRAWINGS

[0046] In order to more clearly illustrate the technical solutions and advantages of the embodiments of the present invention or the prior art, the following briefly introduces the drawings required for use in the embodiments or the prior art descriptions. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative work.

[0047] Figure 1 This is a flow chart of the parking space guidance method based on the vehicle AI platform provided by the present invention;

[0048] Figure 2 A schematic diagram of the steps of a parking space guidance method based on a vehicle AI platform. DETAILED DESCRIPTION

[0049] To further illustrate the technical means and effectiveness of the present invention to achieve its intended purpose, the following, in conjunction with the accompanying drawings and preferred embodiments, describes in detail the specific implementation, structure, features, and effectiveness of the parking space guidance method based on the vehicle AI platform proposed in accordance with the present invention. In the following description, different references to "one embodiment" or "another embodiment" do not necessarily refer to the same embodiment. Furthermore, specific features, structures, or characteristics of one or more embodiments may be combined in any suitable manner.

[0050] Unless defined otherwise, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this invention belongs.

[0051] The specific scheme of the parking space guidance method based on the vehicle AI platform provided by the present invention is described in detail below with reference to the accompanying drawings.

[0052] An embodiment of the present invention provides a parking space guidance method based on a vehicle AI platform.

[0053] Specifically, the following parking guidance method based on the vehicle AI platform is provided. Figure 1 , the method comprises the following steps:

[0054] Step S001: divide the parking lot into grids, obtain parking space-related data for each grid, and obtain urban road conditions and weather data around the parking lot.

[0055] This embodiment aims to provide intelligent guidance for parking spaces in a parking lot based on historical parking data of the parking spaces and historical road conditions and weather data of the city surrounding the parking lot. This embodiment takes a shopping mall parking lot as an example for detailed description. First, this embodiment divides the shopping mall parking lot space into S grids of equal size. It should be noted that the value of S can be set by the implementer, and this embodiment does not impose any specific restrictions. Each grid contains multiple parking spaces and non-parking spaces. Non-parking space refers to the space occupied by everything that does not belong to the parking space, such as elevators, walkways, and pillars. The number of parking spaces in different grids is different, assuming that each parking space has the same shape and size; the area of ​​non-parking space in different grids is also different.

[0056] An infrared sensor is placed above the ground in each parking space. The infrared sensor collects information about the parking space occupancy status and records the information data collected by the infrared sensor in the parking lot management system. Determining whether a parking space is occupied using an infrared sensor is a well-known technique, and the specific process will not be described in detail here. The information data collected by the infrared sensor is whether each parking space is occupied at each moment. Based on the occupancy status of each parking space at each moment recorded in the parking lot management system, the parking space-related data for each grid is obtained, specifically:

[0057] Set the first preset duration Hours are used as a time interval, and the length of the time interval can be set by the implementer. In each time interval, for each grid:

[0058] (1) Every time the second preset time passes A statistical moment is set every minute. At each statistical moment, the number of occupied parking spaces in the grid is counted and determined as the number of occupied parking spaces at each statistical moment. The average number of occupied parking spaces at all statistical moments is divided by the total number of parking spaces in the grid, and the result is taken as the parking space occupancy rate of the grid.

[0059] (2) Within the time interval, count the number of times each parking space changes from occupied to unoccupied, take the first number as the number of vehicles leaving each parking space, and take the sum of the number of vehicles leaving all parking spaces as the number of vehicles leaving the grid;

[0060] (3) Within the time interval, count the number of times each parking space changes from the unoccupied state to the occupied state, take the second number as the number of cars entering each parking space, and take the sum of the number of cars entering all parking spaces as the number of cars entering the grid;

[0061] (4) Obtain the total duration of occupancy for each parking space within the time interval, record it as the total occupancy time of each parking space, take the average of the total occupancy time of all parking spaces within the time interval, and use this average as the average occupancy time of the parking spaces in the grid;

[0062] (5) Obtain the duration of each parking space being occupied, which is used as the parking duration of each vehicle parked in the parking space. Calculate the average parking duration of all vehicles parked in the parking spaces, and use this average as the average parking duration of the vehicles in the grid.

[0063] (6) Obtain the area occupied by non-parking spaces within the grid, calculate the sum of the areas of all unoccupied parking spaces at each statistical moment, and use this sum as the area of ​​vacant parking spaces at each statistical moment; and use the average of the vacant parking space areas at all statistical moments and the sum of the areas occupied by non-parking spaces as the average vacant area of ​​the grid;

[0064] (7) Obtain the number of parking lot entrances, exits, elevators, and corridor doors that are open in the grid, and use this number as the number of channels in the grid.

[0065] In this embodiment, the third preset time length Hours are considered a cycle, and the implementer can set the cycle length. The sequence of parking space occupancy rates for all time intervals of the sth grid within a cycle, in ascending time order, is used as the parking space occupancy rate sequence for the sth grid within the cycle. The sequence of vehicle outflows, vehicle inflows, average parking space occupancy time, average vehicle parking time, average free space area, and number of lanes for the sth grid within the cycle is obtained using the parking space occupancy rate sequence acquisition method.

[0066] Then collect the area around the parking lot The city's road conditions and weather data within meters, including the rainfall per minute, total rainfall time, average sunshine intensity, and traffic volume around the parking lot in each time interval. The rainfall and total rainfall time are obtained from the data recorded by the local meteorological bureau; the average sunshine intensity is collected by the light sensor; the traffic volume around the parking lot is the nearest neighboring area around the parking lot. The average value of traffic volume at the intersection. and The implementer can set the value of , and this embodiment does not impose any specific restrictions.

[0067] The rainfall sequence in each time interval is used as the rainfall sequence of the parking lot in the time interval. The total rainfall time series, average sunshine intensity series and traffic flow series around the parking lot in the time interval are obtained by the rainfall sequence acquisition method.

[0068] All of the above content is transmitted to the vehicle AI platform system through a computer program. The system can view parking lot information and related sensor data in real time and perform calculations. The main functions of the vehicle AI platform system include the following:

[0069] Data Integration and Convergence Center: The vehicle AI platform system is a unified receiving and storage platform for all collected data (including parking lot internal sensor data, urban road conditions, weather data, etc.).

[0070] Real-time monitoring and data analysis: The system can view various parking lot information in real time, such as the occupancy of each parking space, the occupancy rate of parking spaces in the grid, the number of vehicles entering and exiting the parking lot, and the average parking time of vehicles. It can also receive and process external weather and traffic data.

[0071] Computational Processing and Pattern Recognition: The vehicle AI platform system possesses powerful computing capabilities, enabling complex calculations and analysis of massive amounts of historical and real-time data. It generates various data series (such as parking space occupancy rates and vehicle dispatch rates) and further explores potential connections and patterns within these data, providing a basis for subsequent intelligent guidance.

[0072] Step S002: The unit occupancy impact of each grid in each time interval is obtained based on the correlation between various types of data in the parking space-related data and the local data change trend. The parking space vacancy index of each grid in each time interval is constructed in combination with the discrete situation between urban road conditions and weather data.

[0073] Among the collected parking space-related data, the five types of data, namely parking space occupancy rate, number of vehicles leaving, number of vehicles entering, average parking space occupancy time and average parking time of vehicles, reflect the overall vehicle parking situation in each grid. Through these five types of data, we can intuitively see the parking situation of each grid, and these five types of data are all determined as the grid's parking space parking data; and the two types of data, average vacant area and number of lanes, can show the popularity of each grid's parking space, and indirectly reflect the intensity of car owners' desire to park in each grid. These two types of data are both determined as the grid's parking intention data.

[0074] In the same cycle, for each grid, in order to analyze the correlation between various types of parking space parking data and various types of parking intention data, the i-th type of parking space parking data is used as Indicates that the j-th type of data in the parking intention data is used Representation; calculation The time series of data The Pearson correlation coefficient between the time series of the data is used to obtain the correlation between the time series of the two types of data. , correlation Can react Class data and The strength of the linear relationship between the class data, the larger the value, the greater the impact of the j-th class parking intention data on the i-th class parking space parking data, and vice versa, through the correlation The correlation between any type of parking intention data and any type of parking space parking data is obtained by the acquisition method.

[0075] For parking data of type i parking space The time series is divided into time series using the K-means clustering algorithm, and the K parameter is set. It should be noted that the value of K can be set by the implementer. Since a time interval in this embodiment is 1 day, in order to divide the data into five periods of vehicle parking, namely the early low-peak period, the early peak period, the middle low-peak period, the late peak period and the late low-peak period, the value of K is set to 5 in this embodiment, and the Manhattan distance is used as the measurement method to obtain five clusters. Arrange all the data in each cluster in chronological order, and the obtained sequence is used as the sub-sequence of each cluster. Arrange all the sub-sequences according to the early and late time, corresponding to the early low-peak period, the early peak period, the middle low-peak period, the late peak period and the late low-peak period respectively. Each sub-sequence of the parking data of the i-th type of parking space is obtained respectively. Construct the comprehensive influencing factor of the time period of the s-th grid. :

[0076]

[0077]

[0078] in, is the comprehensive impact factor of the time period of the s-th grid, representing the overall impact of different time periods within a day and the public facilities conditions within the grid on the parking situation of the s-th grid; is the parking impact of the s-th grid; The number of types of parking data; is the number of types of parking intention data; K is the number of clusters obtained by clustering the parking data of each type of parking space; To find the maximum function; is the parking data of the i-th type of parking space The value of the last data point in the sequence, is the value of the first data point in the kth subsequence of parking data of the i-th type of parking space, is the value of the last data point in the kth subsequence of parking data of the i-th type of parking space, The parking data of the i-th type of parking space The value of the first data point in the subsequence, where when k=1, let The value of is 0, when k=K, let The value of is 0; is the number of all data points in the kth subsequence of parking data of the i-th type of parking space; is the average number of data points in the kth subsequence of all types of parking space data, Parking data for the i-th parking space and the parking intention data of category j The Pearson correlation coefficient between the time series. It should be noted that when k=1 The value of k=K The value of can be set by the implementer, and this embodiment does not impose any specific restrictions.

[0079] For the overall environmental effect, the calculation is done by calculating the mean of the Pearson correlation coefficient between the time series of parking space data and parking intention data, which overall reflects the impact of the spaciousness of parking spaces and the convenience of surrounding public facilities on the parking of car owners in the area. The larger the value, the greater the impact of environmental factors on the parking intention of car owners in different areas, and the greater the parking impact; and the comprehensive influencing factor in the time period In is the environmental weight, By finding the maximum change between the kth subsequence and its left and right adjacent subsequences, we can reflect the change of parking space occupancy in the case of time period changes. The larger the value, the more sensitive the parking space occupancy is to the time period changes. The larger the time period, the response of different parking data to the time period will be analyzed. The weight as a function of the maximum value, The larger the value, the greater the difference in the number of data in the kth subsequence between the parking data of the i-th parking space and the parking data of other parking spaces. It means that the response of different parking data to the change of time periods is more complex, and the smaller the weight given to the maximum function.

[0080] Afterwards, the sub-sequences of each type of parking intention data are obtained by obtaining the sub-sequences of parking data for each type of parking space. The first data point and the last data point of each sub-sequence are used as the boundary points of each sub-sequence. In each sub-sequence, the serial number of each data point in the sub-sequence is obtained, and the absolute value of the difference between the serial number of each data point and the nearest boundary point is obtained as the boundary distance of each data point. When the data point is exactly a boundary point, the boundary distance of the data point is set to 1. The implementer can also set the value of the boundary distance according to the actual situation. This embodiment does not impose specific restrictions. In the time series of each type of parking space related data, each data point is constructed with each data point as the center. The neighborhood window is a window of the shape of the image. It should be noted that the value of P can be set by the implementer. In this embodiment, the value of P is set to 5. If there is no data in some positions in the window, it is filled with the mean of all existing data in the window. Then, all data points in the neighborhood window are linearly fitted by the least squares method to obtain the fitted straight line equation, which is used as the fitted straight line equation for each central data point. The independent variable of the fitted straight line equation is time, the dependent variable is the value of the data point, and the slope of each fitted straight line equation is calculated. The least squares method for fitting the linear equation and the calculation of the slope of the linear equation are well-known technologies, and the specific process will not be repeated here. Construct the unit occupancy effect of the sth grid in the dth time interval :

[0081]

[0082] in, is the unit occupancy impact of the sth grid in the dth time interval, which expresses the degree to which the parking space occupancy of the sth grid in the dth time interval is affected by various situations. is the comprehensive impact factor of the time period of the s-th grid, is the number of types of parking space related data, is the value of the data related to the c-th type parking space in the d-th time interval of the grid, for The boundary distance of the corresponding data point, is the maximum value among all data related to the c-th type parking space. For The slope of the fitted straight line equation for the data within the neighborhood window centered at .

[0083] In the calculation, through The shortest distance between the two boundary points of its subsequence right Make constraints, The larger it is, the greater the impact of the overall environment and time period on parking space occupancy. The bigger; The larger the The further away from the time period, the more The greater the constraint, The smaller; It represents the change of parking space occupancy. The larger the value, the greater the change of parking space occupancy. The bigger; It reflects the update rate of parking space occupancy changes around the dth time interval. The larger the value, the faster the parking space change update will occur around the dth time interval. The bigger.

[0084] Since urban road conditions and weather changes will affect the owner's car usage, calculate the discrete coefficient of the f-th type of data in the urban road conditions and weather data in the g-th time interval , where the dispersion coefficient is a well-known technology, and the specific process will not be repeated here. Constructing the car use impact index :

[0085]

[0086] in, The car use impact index reflects the comprehensive impact of road conditions and weather conditions on car owners' car use. is the number of types of urban traffic and weather data, n is the number of time intervals in a cycle, is the dispersion coefficient of the f-th type of urban road conditions and weather data in the g-th time interval.

[0087] exist In the calculation of represents the discrete degree of the road condition and weather data of the f-type city, that is, the chaotic situation of the external environment changes. The more chaotic the external environment changes, The bigger it is, the greater the impact on the car owner's car use. The bigger.

[0088] Furthermore, combined with the unit occupancy effect of the sth grid in the dth time interval and the car use impact index , construct the parking space vacancy index of the sth grid in the dth time interval :

[0089]

[0090] in, is the parking space vacancy index of the sth grid in the dth time interval, reflecting the probability that there are vacant parking spaces in the sth grid under the change of parking spaces in the dth time interval. is the car use impact index, is the unit occupancy impact of the sth grid in the dth time interval, is an exponential function with base e.

[0091] The larger the value is, the greater the impact of surrounding road conditions and weather on car use, and the higher the probability of vacant parking spaces. The bigger, The larger the value is, the more the parking space occupancy is affected by various situations, and the lower the probability of the parking space being vacant. The smaller.

[0092] The parking space vacancy index of each grid in each time interval is obtained by calculating the parking space vacancy index of the sth grid in the dth time interval.

[0093] Step S003: Train an autoregressive moving average model based on the parking space vacancy index of each historical time interval; predict the parking space vacancy index of each grid in the next time interval based on the trained autoregressive moving average model, and recommend the grid with the largest predicted value to the car owner to complete parking space guidance.

[0094] For each grid, the sequence of parking space vacancy indexes of all time intervals in each cycle in ascending time order is recorded as each sample sequence; the sample sequences of the first 30 cycles of the current time are combined into a set as a training set, and the training set is input into the autoregressive moving average model to train the model, wherein the autoregressive order, the difference order and the moving average order are set. It should be noted that the values ​​of the autoregressive order, the difference order and the moving average order can be set by the implementer. In this embodiment, the values ​​of the autoregressive order, the difference order and the moving average order are set to 2, 1 and 2 respectively.

[0095] For each grid, the sequence of parking space vacancy indices of all time intervals from the current moment of the current cycle to the same moment of the previous cycle in ascending time order is recorded as the parking space vacancy index sequence. The parking space vacancy index sequence at the current moment is input into the trained autoregressive moving average model to predict the parking space vacancy index of the next time interval at the current moment. The output of the model is the predicted value of the parking space vacancy index of the next time interval at the current moment.

[0096] Get the predicted value of parking space vacancy index of each grid in the next time interval, recommend the grid with the largest predicted value to the car owner, and guide the car owner to park in the recommended interval, thus completing the guidance of the car owner's parking space selection. The steps of the above method are shown in the figure below. Figure 2 shown.

[0097] In summary, the embodiment of the present invention combines parking space-related data with K-means clustering to construct a time period comprehensive environmental factor, which can analyze the degree of influence of the differences between the parking space environment and the surrounding environment of the parking lot in different time periods of a day on the parking space occupancy. Then, based on the linear fitting results of the data in the time window, a unit occupancy influence is constructed, which represents the comprehensive situation of the changes in the parking space occupancy due to various influences in a time interval. Based on the discrete situation of urban road conditions and weather data, a car use influence index is constructed to reflect its impact on the probability of car owners using their cars. Finally, the car use influence index and the unit occupancy influence are combined to construct the parking space vacancy index of each grid in a time interval. The ARIMA algorithm is trained according to the parking space vacancy index of each grid in each time interval in each historical period. The parking space vacancy index of each grid in the next time interval is predicted according to the trained autoregressive moving average model to obtain the grid with the largest predicted value. The grid with the largest predicted value is recommended to the car owner, thereby optimizing the customer's parking experience and enhancing the robustness of the algorithm.

[0098] Based on the same inventive concept as the above method, an embodiment of the present application also provides a parking space guidance system based on a vehicle AI platform, including a memory, a processor, and a computer program stored in the memory and running on the processor. When the processor executes the computer program, it implements the steps of any one of the above-mentioned parking space guidance methods based on the vehicle AI platform.

[0099] One embodiment of the present invention also provides a parking space guidance device based on a vehicle AI platform, which includes a memory, a processor, and a computer program, wherein the memory is used to store the corresponding computer program, and the processor is used to run the corresponding computer program. When the computer program runs in the processor, it can implement the parking space guidance method based on the vehicle AI platform described in steps S001-S003.

[0100] It should be noted that the order in which the embodiments of the present invention are described above is for illustrative purposes only and does not necessarily represent the superiority or inferiority of the embodiments. Furthermore, the foregoing descriptions of specific embodiments of this specification are provided. Furthermore, the processes depicted in the accompanying drawings do not necessarily require the specific order or sequential sequence shown to achieve the desired results. In certain embodiments, multitasking and parallel processing are also possible or may be advantageous.

[0101] The various embodiments in this specification are described in a progressive manner, and the same or similar parts between the various embodiments can be referred to each other. Each embodiment focuses on the differences from other embodiments.

[0102] The above-described embodiments are only used to illustrate the technical solutions of the present application, and not to limit them. Modifications to the technical solutions described in the aforementioned embodiments, or equivalent replacements of some of the technical features therein, do not cause the essence of the corresponding technical solutions to deviate from the scope of the technical solutions of the embodiments of the present application, and should all be included in the scope of protection of the present application.

Claims

1. A parking space guidance method based on a vehicle AI platform, characterized in that: The method comprises the following steps: The parking lot is divided into grids to obtain each grid, with a first preset time duration as a time interval, a statistical moment set every time a second preset time duration passes, and a third preset time duration as a cycle; within each time interval, parking space occupancy data for each grid is obtained based on the parking situation of all parking spaces in each grid at each statistical moment; parking intention data for each grid is obtained based on the area and facilities of non-parking spaces in each grid; urban road conditions and weather data around the parking lot are obtained; and the parking space occupancy data and parking intention data for each grid are used as parking space-related data for each grid; In each period, the time series of various parking space-related data are divided into subsequences using the K-means clustering algorithm. The time period comprehensive impact factor of each grid is obtained based on the data changes in each subsequence and the correlation between parking data and parking intention data. The unit occupancy impact of each grid in each time interval is obtained based on the time period comprehensive impact factor and the numerical changes in each subsequence. The vehicle use impact index is obtained based on the discrete situation of various urban road conditions and weather data. The parking space vacancy index of each grid in each time interval is obtained based on the unit occupancy impact and the vehicle use impact index. An autoregressive moving average model is trained based on the parking space vacancy index of each grid in each time interval. The trained autoregressive moving average model is used to predict the parking space vacancy index of each grid in the next time interval, and the grid with the largest predicted value is recommended to the driver, thus guiding the driver in parking space selection. The comprehensive impact factor of the time period satisfies the following formula: in, is the comprehensive impact factor of the time period of the s-th grid, representing the overall impact of different time periods within a day and the public facilities conditions within the grid on the parking situation of the s-th grid; is the parking impact of the s-th grid; The number of types of parking data; is the number of types of parking intention data; K is the number of clusters obtained by clustering the parking data of each type of parking space; To find the maximum function; is the parking data of the i-th type of parking space The value of the last data point in the sequence, is the value of the first data point in the kth subsequence of parking data of the i-th type of parking space, is the value of the last data point in the kth subsequence of parking data of the i-th type of parking space, The parking data of the i-th type of parking space The value of the first data point in the subsequence; is the number of all data points in the kth subsequence of parking data of the i-th type of parking space; is the average number of data points in the kth subsequence of all types of parking space data, Parking data for the i-th type of parking space and the parking intention data of category j Pearson correlation coefficient between the time series.

2. The parking space guidance method based on the vehicle AI platform according to claim 1, characterized in that: The parking space occupancy data of each grid is obtained based on the parking situation of all parking spaces at each statistical moment in each grid, specifically: The parking data of each grid includes parking space occupancy rate, number of vehicles leaving, number of vehicles entering, average parking space occupancy time and average vehicle parking time; For each grid, the parking space occupancy rate is as follows: obtaining the number of occupied parking spaces in the grid at each statistical moment, and calculating the average value of the number at all statistical moments; The ratio of the average value to the total number of parking spaces in the grid is used as the parking space occupancy rate of the grid; The number of vehicles dispatched is: the number of times each parking space changes from occupied to unoccupied is counted and recorded as the number of vehicles dispatched; The sum of the number of vehicles leaving all parking spaces is taken as the vehicle leaving capacity of the grid; The number of cars entering the grid is calculated by counting the number of times each parking space changes from an unoccupied state to an occupied state, and recording the number of cars entering the grid; the sum of the number of cars entering all parking spaces is taken as the number of cars entering the grid; The average parking space occupancy time is: the total time each parking space is occupied is obtained; The average of the total occupied time of all parking spaces is used as the average occupied time of parking spaces in the grid; The average parking time of a vehicle is as follows: the continuous occupation time of each parking space is obtained, and the average of the continuous occupation time of all parking spaces is used as the average parking time of the vehicle in the grid.

3. The parking space guidance method based on the vehicle AI platform according to claim 1, characterized in that: The parking intention data of each grid is obtained based on the area and facilities of the non-parking spaces of each grid, specifically: The parking intention data for each grid include the average vacant area and number of lanes; For each grid, the average free area is calculated by: calculating the sum of the areas of all parking spaces that are not occupied at each statistical moment; calculating the average of the sum of the areas at all statistical moments; and taking the sum of the average and the area of ​​the non-parking spaces as the average free area of ​​the grid; The number of channels is: the number of parking lot entrances, exits, elevators and corridor doors in the grid that are open is taken as the number of channels in the grid.

4. The parking space guidance method based on the vehicle AI platform according to claim 1, characterized in that: The acquisition of urban traffic and weather data around the parking lot is specifically as follows: Obtain the minute-by-minute rainfall and total rainfall duration around the parking lot, as well as the minute-by-minute average sunshine intensity around the parking lot; collect the minute-by-minute traffic volume at each intersection near the parking lot; and use the average of the minute-by-minute traffic volume at all intersections as the traffic volume around the parking lot. The rainfall, total rainfall time, average sunshine intensity and traffic volume around the parking lot are used as urban road conditions and weather data.

5. The parking space guidance method based on the vehicle AI platform according to claim 1, characterized in that: The unit occupancy impact of each grid in each time interval is obtained based on the comprehensive impact factor of the time period and the value changes in each sub-sequence, specifically including: The first and last data points in each subsequence of each type of parking space-related data are taken as boundary points; the absolute value of the difference between the serial numbers of each data point in each subsequence and the nearest boundary point is recorded as the boundary distance of each data point; In the time series of various parking space related data, the fitted straight line equation obtained by linear fitting all data points in the neighborhood of each data point is used as the fitted straight line equation of each data point; The expression for calculating the unit occupancy impact of each grid in each time interval is: in, is the unit occupancy impact of the sth grid in the dth time interval, which expresses the degree to which the parking space occupancy of the sth grid in the dth time interval is affected by various situations. is the comprehensive impact factor of the time period of the s-th grid, is the number of types of parking space related data, is the value of the data related to the c-th type parking space in the d-th time interval of the grid, for The boundary distance of the corresponding data point, is the maximum value among all data related to the c-th type parking space. for The slope of the fitted line equation corresponding to the data points.

6. The parking space guidance method based on the vehicle AI platform according to claim 1, characterized in that: The vehicle use impact index is obtained based on the discrete conditions of the road conditions and weather data of various cities, specifically including: Obtain the dispersion coefficient of all data in the road conditions and weather data of various cities; the expression for calculating the vehicle use impact index based on the dispersion coefficient of the road conditions and weather data of various cities is: in, is the car use impact index, is the number of types of urban traffic and weather data, n is the number of time intervals in a cycle, is the dispersion coefficient of the f-th type of urban road conditions and weather data in the g-th time interval.

7. The parking space guidance method based on the vehicle AI platform according to claim 1, characterized in that: The parking space vacancy index of each grid in each time interval is obtained based on the unit occupancy impact and vehicle use impact index, specifically including: Calculate the ratio of the unit occupancy impact to the vehicle use impact index of each grid in each time interval; and use the calculation result of an exponential function with a natural constant as the base and the inverse of the ratio as the exponent as the parking space vacancy index of each grid in each time interval.

8. A parking space guidance system based on a vehicle AI platform, comprising a memory, a processor, and a computer program stored in the memory and running on the processor, characterized in that: When the processor executes the computer program, the steps of the method according to any one of claims 1 to 7 are implemented.

9. A parking space guidance device based on a vehicle AI platform, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that: When the processor executes the computer program, the steps of the parking space guidance method based on the vehicle AI platform as described in any one of claims 1 to 7 are implemented.

Images