Parking lot complex charging management system based on artificial intelligence

Abstract

The invention relates to the technical field of parking management, in particular to a parking lot complex charging management system based on artificial intelligence, and the system comprises a data collection module which is used for collecting the parking data of a user, and the parking data comprises the vehicle entering and leaving time, the vehicle type and the position of a parking space of each user; the behavior recognition module is used for generating a user behavior portrait corresponding to each user according to the parking data of the user; the portrait analysis module is used for analyzing the user behavior portrait, determining a parking strategy combination corresponding to the user behavior portrait, and determining differentiated earnings of the user under different behavior decisions and time sequence characteristics under the differentiated earnings according to the obtained parking strategy combination; and intelligence and refinement of charging management of the parking lot are realized.

Description

Technical Field

[0001] The present invention relates to the technical field of parking management, and in particular to a complex parking lot billing management system based on artificial intelligence. Background Art

[0002] The main reasons for urban parking problems are insufficient supply of parking space resources and difficulty in allocating parking space resources. Especially the latter, in hot spots in cities, the supply of parking spaces is relatively fixed, but the demand for parking spaces is very "tidal" and unplanned, which brings great challenges to the setting, control, and operation of parking lots.

[0003] For example, Chinese patent publication No. CN113808287A discloses a management method and system for time-based billing in parking lots by area, which includes the following steps: Step S1, drawing a user inherent portrait of the target user; Step S2, drawing a user demand portrait of the target user; Step S3, integrating the user inherent portrait and the user demand portrait to create a user parking portrait, and formulating a special parking service for the target user based on the target parking portrait.

[0004] The existing technology creates a user parking profile by setting a user demand profile and a user inherent profile to identify the characteristics of different areas when the user parks, so as to manage the current parking. However, when determining these characteristics, it is also necessary to determine the corresponding changes in the parking lot corresponding to the user when parking, as well as the differentiation corresponding to each parking space. For example, when differentiated charging management is implemented in the parking lot, these parking spaces cannot be adapted to the relevant situation of the current parking lot only through the weight of the user's parking profile. At the same time, dynamic adjustments are also required to avoid parking space allocation errors due to different parking lots or profile settings. Summary of the invention

[0005] In order to solve the above technical problems, the technical solution adopted by the present invention is: a complex parking lot billing management system based on artificial intelligence, including: a data acquisition module for collecting users' parking data, the parking data including the time of entry and exit of each user's vehicle, the vehicle type and the location of the parking space.

[0006] The behavior recognition module is used to generate a user behavior profile corresponding to each user based on the user's parking data.

[0007] The portrait analysis module is used to analyze the user behavior portrait, determine the parking strategy combination corresponding to the user behavior portrait, and determine the differentiated benefits of the user under different behavioral decisions and the time series characteristics under the differentiated benefits based on the obtained parking strategy combination.

[0008] The regional analysis module is used to determine the parking demand, turnover area and corresponding parking constraints of each time series according to the acquired time series characteristics; and determine the parking strategy selected by the user when parking according to the set parking constraints.

[0009] The parking adjustment module is used to determine the similarity measure between the parking strategy selected by the user when parking and the user behavior profile, and adjust the user guidance according to the difference in the similarity measure between the parking strategy combination and the user behavior profile when the user parks.

[0010] The beneficial effects of the present invention are as follows: 1. Through the behavior recognition module and user portrait analysis, the system can dynamically adjust the billing strategy according to the user's parking behavior and preferences, and provide personalized billing plans, thereby improving billing flexibility and user satisfaction.

[0011] 2. The regional analysis module provided in the present invention can intelligently allocate parking resources according to parking demands at different time sequences and constraints of turnover areas, effectively alleviate parking lot congestion problems, and improve parking efficiency.

[0012] 3. The present invention achieves refined management of parking lot revenue by calculating the differentiated revenue of users under different behavioral decisions and adjusting vehicle guidance according to revenue weights, which helps to improve the operational efficiency of the parking lot.

[0013] 4. Through intelligent parking strategy adjustment and user guidance, the system of the present invention can provide users with a more convenient and efficient parking experience, reducing the time and cost of users looking for parking spaces. BRIEF DESCRIPTION OF THE DRAWINGS

[0014] The present invention is further described below in conjunction with the accompanying drawings and embodiments.

[0015] Figure 1 It is a system diagram of a complex parking lot billing management system based on artificial intelligence.

[0016] Figure 2 It is a structural diagram of the user behavior portrait of the complex parking lot billing management system based on artificial intelligence.

[0017] Figure 3 It is a structural diagram of the regional analysis module of the complex parking lot billing management system based on artificial intelligence. DETAILED DESCRIPTION

[0018] The embodiments of the present invention are described in detail below. The embodiments described below are exemplary and are only used to explain the present invention, and should not be construed as limiting the present invention. If no specific techniques or conditions are specified in the embodiments, the techniques or conditions described in the literature in the art or the product specifications are used.

[0019] See also Figure 1 , Figure 2 , Figure 3 The complex parking lot billing management system based on artificial intelligence includes: data acquisition module, behavior recognition module, portrait analysis module, regional analysis module, and parking adjustment module; the data acquisition module sends the collected data to the behavior recognition module, and the behavior recognition module produces relevant user behavior portraits and outputs the data to the portrait analysis module and the regional analysis module. The portrait analysis module is used to determine the current user behavior preference and determine the differentiated benefits of the current strategy combination. The regional analysis module determines the changes in the relevant demand and the corresponding parking lot during the current parking to determine the parking constraints. The output results of the portrait analysis module and the regional analysis module are input into the parking adjustment module, and the parking adjustment module finally realizes the adjustment of parking.

[0020] The data collection module is used to collect the user's parking data, which includes the time when each user's vehicle enters and leaves, the vehicle type and the location of the parking space.

[0021] The behavior recognition module is used to generate a user behavior profile corresponding to each user based on the user's parking data.

[0022] The portrait analysis module is used to analyze the user behavior portrait, determine the parking strategy combination corresponding to the user behavior portrait, and determine the differentiated benefits of the user under different behavioral decisions and the time series characteristics under the differentiated benefits based on the obtained parking strategy combination.

[0023] The regional analysis module is used to determine the parking demand, turnover area and corresponding parking constraints of each time series according to the acquired time series characteristics; and determine the parking strategy selected by the user when parking according to the set parking constraints.

[0024] The parking adjustment module is used to determine the similarity measure between the parking strategy selected by the user when parking and the user behavior profile, and adjust the user guidance according to the difference in the similarity measure between the parking strategy combination and the user behavior profile when the user parks.

[0025] At this time, parking data is used to determine the user's needs at this time; to determine the user's parking behavior pattern under the user behavior portrait, to determine the user's behavior during peak hours and preferred parking spaces, so as to obtain the user's time preference and location preference when parking; and based on these preferences, find the parking strategy corresponding to the preference, and perform differentiation analysis on the parking strategy to obtain the parking method that is most relevant to the current parking lot setting, thereby guiding the user's parking.

[0026] At this time, in order to determine the accuracy of the current collected data, high-position and low-position cameras, geomagnetic sensors, curb machines, inspection vehicles and other equipment are set up in the parking lot to determine the current parking situation of vehicles in the parking lot, and based on the data obtained, the existence of vehicles in the parking lot, turnover rate, utilization of each parking space, etc. are judged.

[0027] High-position cameras can provide a wider field of view and are suitable for monitoring multiple parking spaces; low-position cameras provide more detailed vehicle information, such as license plate numbers, etc.; geomagnetic sensors are the core components of the geomagnetic induction parking fee collection system and are installed above the parking space. They detect the entry and exit of vehicles by sensing changes in the geomagnetic field on the underground parking space; when a vehicle enters or leaves a parking space, the geomagnetic sensor can accurately sense it and transmit the signal to the data processor for recording and billing; a curb machine is a device installed next to a parking space to identify vehicle entry and exit and license plate information; it can realize vehicle detection, license plate recognition, and support unlicensed vehicle detection; patrol cars are usually driven by parking lot managers, who patrol the parking lot regularly to monitor the use of parking spaces; patrol cars may be equipped with cameras and sensors for real-time detection and recording of the status of parking spaces; through patrol cars, managers can promptly discover and deal with problems such as illegal parking and occupying multiple parking spaces.

[0028] The method for generating a user behavior profile includes determining the user's parking time preference, location preference, parking duration preference and parking fee preference according to the acquired parking data, and generating the user behavior profile based on the user's parking time preference and location preference.

[0029] The time preference is expressed as obtaining the number of times the user parks and the duration of parking to determine the user's parking time preference; the user's parking time preference includes daily time preference, weekly time preference and long-term preference.

[0030] The daily time period preference is expressed as follows: the number and duration of users' parking in different time periods are counted to determine the difference in users' parking demands in different time periods; the proportion of parking times and parking duration in each time period are calculated to obtain the difference in users' parking demands in different time periods; the average value of the difference in users' parking demands in different time periods is used as the current output daily time period preference.

[0031] The weekly time period preference is expressed as follows: the number of times users park and the market on different days are counted to determine the differences in users' parking needs on different dates; each day is divided into multiple day types according to weekdays, rest days, holidays and corresponding specific times, and the proportion of parking times and parking duration of each day type is calculated to determine the differences in users' parking needs on different dates.

[0032] For long-term preference expression, the differences in users' parking demands in different cycles are counted to determine the changes in the differences in users' parking demands in consecutive cycles; by combining the daily time period preferences and weekly time period preferences according to the set cycles, the change rate of the number of parking times and parking duration in each cycle is determined, and the moving average of the number of parking times and parking duration in adjacent cycles is used as the change in the differences in the current user's parking demands in consecutive cycles.

[0033] The above-mentioned cycles are set in the form of 15 days, 1 month, 3 months, half a year, and 1 year; in order to make the currently set cycle able to cope with more scenarios, when setting the preset cycle, the behavior change rate of the user behavior indicator in the current cycle and the user behavior indicator in the previous cycle is calculated. The user behavior indicator represents the average value of the corresponding data in the user behavior portrait, and the behavior change rate is obtained by calculating the difference between the user behavior indicators of the current cycle and the previous cycle, and calculating the ratio of the difference to the user behavior indicator of the previous cycle. At this time, it can reflect the user's behavior changes in different cycles. If the behavior change amount obtained at this time exceeds the preset change amount threshold, the surrounding will be shortened to half. If three consecutive cycles are lower than half of the preset change amount threshold, the cycle will be doubled to adjust the current cycle for detecting user behavior.

[0034] For the user's parking location preference, the user's parking space is obtained, and the regional characteristics of the parking space in the parking lot are determined. The regional characteristics include the distance characteristics between the current parking space and the parking lot entrance, exit and elevator entrance, as well as the convenience characteristics corresponding to the current parking space; based on the regional characteristics of the parking lot, the location preference is obtained.

[0035] The distance feature is calculated by normalizing the distance between the current parking space and the parking lot entrance, exit, and elevator entrance, and taking the weighted sum of the normalized distance values ​​as the output distance feature: Where D represents the standardized distance of the current parking space, d represents the distance from the current parking space to the parking lot entrance, exit or elevator entrance, and d max Indicates the maximum distance from the current parking space to the parking lot entrance, exit and elevator entrance, d min Indicates the minimum distance from the current parking space to the parking lot entrance, exit and elevator entrance.

[0036] The minimum distance is the shortest walking distance from this parking space that can reach the entrance, exit or elevator directly or through the shortest path; the maximum distance is the longest walking distance among all possible paths that can reach the entrance, exit or elevator from this parking space; this does not mean taking a detour, but means that the layout of the parking lot makes certain key locations naturally farther away from the parking space; suppose a parking space is located on one side of the parking lot, and the parking lot has two exits, one on the same side but closer, and the other on the opposite side and farther away. In this case, the distance from the parking space to the closer exit will be the minimum distance to the exit, and the distance to the opposite exit will be the maximum distance.

[0037] The distance feature obtained at this time can represent the relative position of the current vehicle with respect to the corresponding exit, entrance and elevator entrance, so as to facilitate the determination of the corresponding position of the current user.

[0038] The convenience feature is represented by the ratio of the size of the current parking space to the surrounding space, and the distance feature and the convenience feature are output as the location preference. The convenience feature set at this time is used to determine whether the current user's parking space is convenient for exiting. When the ratio of the size of the parking space to the surrounding space is small, it means that the current parking space distribution is relatively crowded and congestion incidents are prone to occur, and it is not convenient for larger vehicles to park. When the ratio of the size of the parking space to the surrounding space is large, it means that the parking space is easy to pass and larger vehicles are also convenient to park.

[0039] The parking time preference indicates the average parking time of the user and the corresponding parking type for each parking. The parking types include short stop, medium stop, long stop, etc., to determine the parking time preferred by the user when parking.

[0040] Parking fee preference indicates the type of charge and price used for each parking, to determine the current user's preference for parking fees; charging types include time-based charging, per-time charging, monthly / quarterly / annual charging, time-based charging, tiered charging, membership-based charging, and dynamic charging, to determine which method the user prefers to settle the corresponding fees.

[0041] Time-based charging: The fee is calculated based on the length of time you park. Common methods include charging by minutes, hours or days.

[0042] Pay Per Use: A fixed fee is charged each time you enter the parking lot, regardless of the length of time you park.

[0043] Monthly / quarterly / annual charges: Users can choose to purchase unlimited parking services within a certain period of time, which is suitable for users who often park at the same place.

[0044] Time-of-day charging: Different fees are charged according to different time periods of the day (such as peak hours and off-peak hours).

[0045] Tiered charges: The charges are low in the first few hours and gradually increase as time goes by.

[0046] Membership fees: Enjoy discounts or other additional services by purchasing membership.

[0047] Dynamic charging: The charging standard is adjusted according to the real-time occupancy rate of the parking lot, increasing the fee when parking spaces are tight and reducing the fee when parking spaces are sufficient.

[0048] The final user behavior portrait is the values ​​of the corresponding data of time preference, location preference, parking duration preference and parking fee preference, as well as the corresponding mean, to determine the main behavior of the current user when parking.

[0049] By analyzing users' parking habits in detail, we can accurately understand key information such as when, where, how long users park and how much they are willing to pay, providing a basis for personalized services; based on user behavior portraits, parking lots can allocate resources more effectively, such as adding staff guidance during peak hours, or adding parking spaces in areas preferred by users, improving customer satisfaction and parking lot operation efficiency; user behavior portraits can serve as an important input for intelligent parking systems, helping to develop smarter parking recommendation, reservation and navigation systems, and improving user experience.

[0050] The processing method for analyzing the user behavior profile includes: determining the inherent characteristics existing in the current area and the time series characteristics of the current area according to the user behavior profile; and selecting the corresponding parking strategy combination from the parking strategy library based on the acquired time series characteristics and inherent characteristics.

[0051] The parking strategy combination is represented as a set of parking strategies related to the current user behavior profile, and there are at least two parking strategies. The inherent characteristics of the user are represented as general characteristics analyzed from the user behavior profile. The time series characteristics are represented as characteristics that appear in a specific time period.

[0052] Then the method to obtain inherent features and time series features is to input the user behavior profile into the time series decomposition model, use the random components identified in the time series decomposition model as inherent features, and use the trend components and seasonal components identified in the time series decomposition model as time series features.

[0053] Therefore, the implementation of the time series decomposition model is as follows.

[0054] Among them, Y t represents the index value of the observed user behavior profile at time point t, T t Represents the index value of the trend component at time point t, T t-1represents the index value of the trend component at time point t-1, S t represents the index value of the seasonal component at time point t, R t represents the index value of the random component at time point t, t represents the number of time points, k represents the time window size, and Y i represents the index value of the user behavior profile at the i-th time point in the time window, α represents the smoothing coefficient, m represents the number of time points in the seasonal cycle, and Y j represents the index value of the user behavior profile at the jth time point in the seasonal cycle, T j Represents the value of the trend component at the j-th time point in the seasonal cycle; i=t-k+1,t-k+2,,,t, j=1,2,,,m, and the currently set seasonal cycle belongs to the set corresponding to the set time point t.

[0055] At this time, the inherent features and time series features obtained by decomposing the user behavior profile are compared with the parking strategy library according to the user's preferred time to obtain a parking strategy combination.

[0056] At this time, the inherent characteristics identified can be expressed as basic information such as the user's gender, age, occupation, etc., the user's long-term parking preferences, such as preference for indoor parking lots or outdoor parking lots, and the user's average parking time. This time may be relatively stable in different time periods. At this time, the inherent characteristics will not change significantly over time.

[0057] The identified time series features can be expressed as the number of users' parking times gradually increasing on weekdays and gradually decreasing on weekends. During certain holidays each year, users' parking behaviors will change significantly, such as an increase in the number of parking times or a longer parking duration. At this time, the identified time series features are easily affected by time, and corresponding trend changes will occur at different times, which can be identified by the time series decomposition model.

[0058] At this time, the implementation method of selecting the corresponding parking strategy combination from the parking strategy library based on the acquired time series features and inherent features is expressed as follows: the cosine similarity between each inherent feature and time series feature and the parking strategy library is calculated, and the solution with the greatest similarity is identified as the output parking strategy combination; the parking strategy combination selected at this time includes the parking strategy with the greatest cosine similarity to the currently set time series feature, the parking strategy with the greatest cosine similarity to the inherent feature, and the parking strategy with the greatest product value of cosine similarity to the time series feature and the inherent feature, so as to facilitate verification of whether each set parking strategy can maximize the final parking benefit.

[0059] Therefore, the method of determining the differentiated benefits of users under different behavioral decisions is to determine the change value of parking space benefits under the currently set parking strategy combination, mark the parking strategy combination that can produce changes in benefits at this time, and record the time series characteristics of generating differentiated benefits; this part of the identified content is used as the output result of the portrait analysis module.

[0060] Specifically, the charging information of the current user is obtained, it is determined whether the current charging information matches the preset charging information, and the parking fees of the vehicles in the parking lot are verified according to the set charging standards. The charging information is divided into multiple target arrays according to different parking fees, and the data of the target array is sorted from large to small. The charging information in the target array that is different from the preset charging information is extracted, and the parking strategy combination corresponding to the extracted charging information is output as the differentiated benefits of the user under different behavioral decisions.

[0061] The charging information obtained at this time indicates whether the current user has historical data in the parking lot, and the charging type corresponding to the current user, to obtain the user-related charging situation. At this time, the charging information is modified and adjusted based on the corresponding situation in the historical data. According to the differentiated benefits obtained at this time and the subsequent calculation of user benefits, users can be guided and controlled according to the different benefits of parking spaces.

[0062] For example, there are parking strategies A and B at this time. When using parking strategy A to deal with seasonal parking peaks, the parking fee income of members near the exit or the parking space set will increase, but the income generated by using parking strategy B will be different from the income generated by parking strategy A. At this time, these parking strategies will be selected, and the time series characteristics of the parking strategies at this time will be determined. These time series characteristics are the time series characteristics identified in the user portrait. At the same time, these time series characteristics are compared to see whether the corresponding parking spaces in the relevant areas will have corresponding trend changes, so as to reversely find out what restrictions are there on the parking strategies that are seriously adopted, and adjust the parking strategies according to these restrictions, so as to obtain the solution that best matches the current parking strategy income to adjust the overall income; at this time, the system can intelligently allocate parking resources according to the parking needs of different time series and the constraints of the turnover area, effectively alleviate the congestion problem in the parking lot, and improve parking efficiency.

[0063] The specific implementation method of the regional analysis module includes, firstly, determining the parking demand under different time sequences according to the acquired time sequence characteristics, and the parking demand set at this time is expressed as the number of parking spaces and the relative parking area required in a specific time period.

[0064] Then, according to the geographical location and number of vehicles in the parking lot, the parking lot is divided into different turnover areas. Each turnover area is responsible for meeting parking needs within a certain range and ensuring that vehicles can enter and exit the parking lot quickly.

[0065] According to the turnover area and parking demand, the parking constraints existing in each turnover area are determined.

[0066] Finally, the parking strategy with the largest cosine similarity with the parking constraint condition is selected as the parking strategy selected by the user when parking, so as to realize the analysis of the parking area.

[0067] The parking constraints are expressed as the capacity limit of the parking lot, changes in demand in different time periods, vehicle types, charging standards and payment methods, environmental protection requirements, user satisfaction and convenience, etc. For example, parking lot A is often full during peak hours on weekdays, and parking lot B has a high vacancy rate during weekends. This means that the parking lot has certain conditions in different time periods. The parking strategy selected at this time needs to correspond to the corresponding conditions so that the final parking strategy can meet the parking needs of the current parking lot.

[0068] That is, the parking constraints in each turnover area can be determined according to the average parking time of each turnover area. The turnover areas are divided into first-class areas, second-class areas, third-class areas and fourth-class areas. Vehicles in the first-class areas have complete entry and exit data and charging information, vehicles in the second-class areas have entry data but no exit data and corresponding charging information, vehicles in the third-class areas have exit data and charging information but no entry data, and vehicles in the fourth-class areas have no data on the relevant parking lots. The parking constraints are determined according to the corresponding types of the turnover areas.

[0069] The first type of area is the normal parking area, which is usually located in the core area of ​​the parking lot for easy management and monitoring. The design and management of these areas focus on efficiency and convenience to ensure that vehicles can enter and exit smoothly. The second type of area is the parking area to be verified, which may be located deeper or in a less conspicuous position in the parking lot to reduce the impact on normal operations. These areas may have special signs or monitoring measures. The third type of area is the temporary parking area, which may be located near the entrance of the parking lot or in a location that is easy to monitor. Temporary parking signs will be set up in these areas. The fourth type of area is the unregistered vehicle area, which may be located at the edge or hidden location of the parking lot. These areas may have isolation measures to prevent unauthorized vehicles from entering at will.

[0070] At this time, according to the identified vehicle-related data, the area where the corresponding vehicle is parked is divided into four areas according to the matching of vehicle-related information. Different supervision measures are taken for these four areas to control parking management.

[0071] The average parking time in the first type of area is obtained by comparing the average vehicle entrance and exit time, and the parking spaces in the first type of area are sorted according to the parking time of the vehicle in each parking space; for the second type of area, the average parking time is obtained by comparing the patrol time of the second type of area with the recorded entrance time, and the relevant areas are patrolled according to the average parking time. The average parking time in the third type of area is calculated according to the charging records; for the fourth type of area, the average parking time of the corresponding vehicle is determined by the interval time of each patrol.

[0072] Then, according to the four category areas set in the turnover area, the corresponding parking constraints are generated in turn, and the corresponding parking strategy is determined according to the parking constraints generated at this time.

[0073] At this time, the method for selecting relevant parking constraints in the four set areas is to first search from the pre-set condition library according to the type of the area, divide the retrieved data according to the average parking time at this time, and identify each segmented data with the current actual parking space and position to select the parking constraint corresponding to the parking space. For example, if the parking space is set with a parking time, or a vehicle type restriction, and related fee restrictions, then the parking strategy needs to be matched with the parking constraint to determine the current parking strategy that meets the conditions set in the parking lot to complete the guidance of the user's vehicle.

[0074] In the parking adjustment module, the similarity measurement between the parking strategy and the user behavior profile is performed by calculating the Pearson correlation coefficient between the parking strategy and the user behavior profile, and the calculated Pearson correlation coefficient is used as the similarity measurement.

[0075] According to the different similarity measures between the parking strategy combination and the user behavior profile when the user parks, the method of adjusting the user guidance is as shown below.

[0076] When verification requires charging, the circulation exchange rate of each parking space within a fixed time, as well as the representative benefits existing under the premise of the circulation exchange rate.

[0077] According to the different representations of benefits, the benefit loss and benefit tendency of each parking space at different times are calculated; based on the benefit loss and benefit tendency, the benefit weight of each parking space is calculated, and according to the benefit weight of each parking space, the guidance of vehicles is adjusted.

[0078] The circulation exchange rate refers to the number of times a parking space is used by different users within a fixed period of time; the higher the circulation exchange rate, the higher the utilization rate of the parking space.

[0079] Representational revenue refers to the total revenue generated by each parking space within a fixed period of time; representational revenue is expressed as the product of circulation exchange rate, average parking time and parking fee.

[0080] Revenue loss refers to the reduction in revenue caused by changes in the circulation exchange rate if the user chooses to park at different time periods. Revenue loss is expressed as the change in revenue caused by changes in the circulation exchange rate.

[0081] The profit tendency refers to the ratio of the current parking space's representative profit to the highest representative profit in different time periods.

[0082] The profit weight is expressed as the profit loss and profit tendency after normalization. The profit weight is obtained by subtracting the profit loss from the set standard value and multiplying it by the profit tendency. At this time, the set standard value is based on the ratio of the current parking space usage time divided by the total time.

[0083] According to the calculated revenue weight, the vehicle guidance strategy can be adjusted to give priority to guiding users to parking spaces with higher revenue weights, thereby increasing the overall revenue of the parking lot; at this time, the parking space with the highest revenue weight in the parking measurement will be selected as the parking space for the current vehicle guidance; the system calculates the differentiated revenue of users under different behavioral decisions and adjusts vehicle guidance according to the revenue weight, thereby achieving refined management of parking lot revenue and helping to improve the operational efficiency of the parking lot.

[0084] Although the embodiments of the present invention have been shown and described above, it is to be understood that the above embodiments are exemplary and are not to be construed as limitations on the present invention. A person skilled in the art may make changes, modifications, substitutions and variations to the above embodiments within the scope of the present invention and they are still covered by the protection scope of the present invention.

Claims

1. Complex parking lot billing management system based on artificial intelligence, It is characterized in that include: A data collection module, used to collect users' parking data, including the time when each user's vehicle enters and leaves, the type of vehicle and the location of the parking space; A behavior recognition module is used to generate a user behavior profile corresponding to each user based on the user's parking data; The portrait analysis module is used to analyze the user behavior portrait, determine the parking strategy combination corresponding to the user behavior portrait, and determine the differentiated benefits of the user under different behavior decisions and the time series characteristics under the differentiated benefits based on the obtained parking strategy combination; The regional analysis module is used to determine the parking demand, turnover area and corresponding parking constraints of each time series according to the acquired time series characteristics; and determine the parking strategy selected by the user when parking according to the set parking constraints; The parking adjustment module is used to determine the similarity measure between the parking strategy selected by the user when parking and the user behavior profile, and adjust the user guidance according to the difference in the similarity measure between the parking strategy combination and the user behavior profile when the user parks.

2. According to the artificial intelligence-based parking complex billing management system of claim 1, It is characterized in that The method for generating a user behavior profile includes determining the user's parking time preference, location preference, parking duration preference and parking fee preference according to the acquired parking data, and generating the user behavior profile based on the user's parking time preference and location preference.

3. According to the artificial intelligence-based parking complex billing management system of claim 1, It is characterized in that The processing method for analyzing the user behavior profile includes: determining the inherent characteristics existing in the current area and the time series characteristics of the current area according to the user behavior profile; and selecting the corresponding parking strategy combination from the parking strategy library based on the acquired time series characteristics and inherent characteristics.

4. According to claim 3, the complex parking lot billing management system based on artificial intelligence, It is characterized in that The method for obtaining the inherent features and time series features is to input the user behavior profile into the time series decomposition model, use the random components identified in the time series decomposition model as the inherent features, and use the trend components and seasonal components identified in the time series decomposition model as the time series features; The implementation of the time series decomposition model is as follows: Y t =T t +S t +R t ; T t =αY t +(1-α)T t-1 ; Among them, Y t represents the index value of the observed user behavior profile at time point t, T t Represents the index value of the trend component at time point t, T t-1 represents the index value of the trend component at time point t-1, S t represents the index value of the seasonal component at time point t, R t represents the index value of the random component at time point t, t represents the number of time points, k represents the time window size, and Y i represents the index value of the user behavior profile at the i-th time point in the time window, α represents the smoothing coefficient, m represents the number of time points in the seasonal cycle, and Y j represents the index value of the user behavior profile at the jth time point in the seasonal cycle, T j Represents the value of the trend component at the j-th time point in the seasonal cycle; i=t-k+1,t-k+2,,,t, j=1,2,,,m.

5. According to claim 3, the complex parking lot billing management system based on artificial intelligence, It is characterized in that The implementation method of selecting the corresponding parking strategy combination from the parking strategy library based on the acquired time series features and inherent features is expressed as follows: calculating the cosine similarity between each inherent feature and time series feature and the parking strategy library, and identifying the solution with the greatest similarity as the output parking strategy combination; the selected parking strategy combination includes the parking strategy with the greatest cosine similarity to the currently set time series feature, the parking strategy with the greatest cosine similarity to the inherent feature, and the parking strategy with the greatest product value of the cosine similarity to the time series feature and the inherent feature.

6. According to claim 1, the complex parking lot billing management system based on artificial intelligence, It is characterized in that The implementation method of determining the differentiated benefits of users under different behavioral decisions is expressed as follows: obtaining the charging information of the current user, determining whether the current charging information matches the preset charging information, and verifying the parking fees of vehicles in the parking lot according to the set charging standards, dividing the charging information into multiple target arrays according to different parking fees, sorting the data of the target arrays from large to small, extracting the charging information in the target array that is different from the preset charging information, and outputting the parking strategy combination corresponding to the extracted charging information as the differentiated benefits of users under different behavioral decisions.

7. According to claim 1, the complex parking lot billing management system based on artificial intelligence, It is characterized in that The specific implementation of the regional analysis module includes: first, determining the parking demand under different time series according to the acquired time series characteristics; dividing the parking lot into different turnover areas according to the geographical location and number of vehicles in the parking lot; and determining the parking constraints existing in each turnover area according to the turnover area and parking demand; The parking strategy with the largest cosine similarity to the parking constraint condition is selected as the parking strategy selected by the user when parking.

8. According to claim 7, the complex parking lot billing management system based on artificial intelligence, It is characterized in that The parking constraints of each turnover area can be determined based on the average parking time of each turnover area. The turnover areas are divided into first-category areas, second-category areas, third-category areas and fourth-category areas. Vehicles in the first-category areas have complete entry and exit data and charging information, vehicles in the second-category areas have entry data but no exit data and corresponding charging information, vehicles in the third-category areas have exit data and charging information but no entry data, and vehicles in the fourth-category areas have no data on the relevant parking lots. Determine the parking constraints based on the type of the turnover area.

9. The complex parking lot billing management system based on artificial intelligence according to claim 1, It is characterized in that According to the different similarity measures between the parking strategy combination and the user behavior profile when the user parks, the user guidance is adjusted as follows: Verify the circulation exchange rate of each parking space within a fixed time when charging is required, and the representative benefits existing under the premise of the circulation exchange rate; According to the different representations of benefits, the benefit loss and benefit tendency of each parking space at different times are calculated; based on the benefit loss and benefit tendency, the benefit weight of each parking space is calculated, and according to the benefit weight of each parking space, the guidance of vehicles is adjusted.

10. The complex parking lot billing management system based on artificial intelligence according to claim 1, It is characterized in that In the parking adjustment module, the similarity measurement between the parking strategy and the user behavior profile is performed by calculating the Pearson correlation coefficient between the parking strategy and the user behavior profile, and the calculated Pearson correlation coefficient is used as the similarity measurement.

Images