Intelligent parking management system based on multi-dimensional data analysis

Abstract

The invention discloses an intelligent parking management system based on multi-dimensional data analysis, and relates to the technical field of urban parking management, and the system can dynamically monitor the real-time state of a parking lot through a parking lot information collection module, can obtain the parking demands of a driver through a to-be-parked person collection unit, and can achieve the intelligent parking management of the driver. Comprise distance preference, cost budget and parking space types, and the most suitable parking lot is accurately matched by combining static and dynamic characteristics of the parking lot, so that the parking convenience is improved. And the correction module calculates a parking cost performance index of the parking lot, and performs dynamic adjustment according to a cost performance threshold value Z, so that the recommended parking lot is ensured to meet requirements, cost, parking convenience and parking space stability are considered, and an optimal choice is provided for a user. The system can reduce disordered wandering of vehicles in high-demand time periods or hot spot areas, improves reasonable configuration of parking resources, reduces traffic congestion caused by searching parking spaces, and improves the overall efficiency of urban parking management.

Description

Technical Field

[0001] The present invention relates to the technical field of XX, and specifically provides an intelligent parking management system based on multi-dimensional data analysis. Background Art

[0002] With the acceleration of urbanization and the continuous increase in the number of motor vehicles, parking difficulties have become a prominent problem in urban traffic management. Traditional parking lot management systems usually rely on manual operations or static data processing, and there are obvious deficiencies in dynamic demand and parking resource allocation. Especially during high-demand periods or in certain specific areas, traditional parking management methods often struggle to effectively respond to fluctuations in parking demand, resulting in waste of parking resources, traffic congestion, and time waste for drivers searching for parking spaces. The limitations of traditional systems are reflected in the following aspects:

[0003] The data of available parking spaces in the parking lot is not updated in a timely manner, the utilization rate of parking spaces in the parking lot is low, and vehicles often linger near the parking lot during peak parking demand periods, further exacerbating traffic congestion. Traditional parking lot management systems generally only provide static parking space information, unable to make dynamic recommendations based on real-time data, and it is difficult to provide personalized parking lot recommendations according to different parking demands and real-time situations. In many cases, drivers cannot obtain detailed information about the parking lot in real time, such as the remaining number of parking spaces, parking space types, charging standards, traffic conditions, etc., resulting in low effectiveness and timeliness of parking decisions. Summary of the Invention

[0004] In view of the deficiencies of the prior art, the present invention provides an intelligent parking management system based on multi-dimensional data analysis to solve the problems mentioned in the background art.

[0005] To achieve the above objectives, the present invention is realized through the following technical solutions: An intelligent parking management system based on multi-dimensional data analysis, including:

[0006] A virtual construction module, used to construct a parking lot and surrounding road scene model based on virtual reality technology, combine the basic information of the parking lot and the parking space distribution rules, simulate the vehicle parking process of multiple groups of parking lots, and generate a visual parking operation interface to display the recommended parking order and the optimal driving path;

[0007] A parking lot information collection module, used to monitor the real-time status of multiple parking lots, including the number of available parking spaces in the parking lot, the parking space turnover rate, the charging standard, the distribution of parking space types, the availability rate of charging parking spaces, and the congestion situation at the parking lot entrance and exit, and construct a static parking lot feature set Pcn and a dynamic parking lot feature set Pcr;

[0008] A parking person to be parked collection unit, used to collect the parking demand of the jth target parking person to be parked and construct a parking constraint demand set Pcm;

[0009] A sequential recommendation module, which is used to perform matrix matching according to the static parking lot feature set Pcn, the dynamic parking lot feature set Pcr, and the parking constraint requirement set Pcm, and generate a first recommended list according to the jth target person to park a vehicle.

[0010] A correction module, which is used to extract the dynamic parking lot feature set Pcr in the first recommended list, calculate and obtain the parking cost performance index X of the ith parking lot jb,i , and a cost performance threshold Z is preset. When X jb,i > Z, it is included in the second-priority recommended list, and is sorted from large to small according to X jb,i , and the highest value is preferentially recommended to the jth target person to park a vehicle.

[0011] Preferably, the parking lot information collection module includes a static feature collection unit and a dynamic feature collection unit;

[0012] The static feature collection unit is used to collect the static information of several parking lots, and the static information includes: the longitude and latitude position Loc of the parking lot, the total number of parking spaces N of the parking lot max , the parking restriction time T limtit , the upper limit P of the parking fee max , the lower limit P of the parking fee min , the monitoring coverage area S fg and the parking space type distribution set BL;

[0013] To construct the static parking lot feature set Pcn, the expression is as follows:

[0014] Pcm = {(Loc(x,y), N max , T limtit , P max , P min , S fg , BL) i |i = 1, 2,..., n|};

[0015] Among them, n represents the total number of parking lots;

[0016] The parking space type distribution set BL includes: the proportion of ordinary parking spaces, the proportion of VIP parking spaces, the proportion of charging parking spaces, the proportion of barrier-free parking spaces; the parking space type distribution set is expressed as: BL = {pt, vt, ct, zt}.

[0017] Preferably, the dynamic feature collection unit is used to collect the dynamic change information of the parking lot, and the dynamic change information includes: the remaining number of parking spaces N P , the parking space switching frequency Q pl , the average low-speed cruising duration T ds, Overtime Occupancy Ratio Z b , Average Return Rate H dt , Reverse Car Search Request Rate H fx and Road Congestion Coefficient D fx to construct a dynamic parking lot feature set Pcr;

[0018] Pcr = {(N P , Q pl , T ds , Z b , H dt , P min , H fx , D fx ) i | i = 1, 2,..., n |}; n represents the total number of parking lots.

[0019] Preferably, the remaining parking spaces N P are calculated and obtained through the following formula:

[0020] N P = (N max - N c );

[0021] where, N c is the total number of real-time parked spaces;

[0022] The parking space switching frequency Q pl is calculated and obtained through the following formula:

[0023]

[0024] where, N seitch represents the number of times a vehicle switches parking spaces during the monitoring period T, and T represents the monitoring period; the parking space switching frequency Q pl reflects the frequency of vehicle parking space changes and is related to the usage habits of the parking lot and the comfort of the parking spaces;

[0025] The average low-speed cruising duration T ds is calculated and obtained through the following formula:

[0026]

[0027] where, ∑T low-speed represents the total time of all vehicles driving at low speed in the parking lot or on the surrounding roads, and low-speed driving is identified as a speed lower than 15 km / h; N vehicle represents the total number of vehicles entering the parking lot during this period; the average low-speed cruising duration T ds is used to analyze the length of time for vehicles to search for parking spaces, so as to optimize the parking space distribution and navigation guidance;

[0028] The overtime occupancy ratio Z b is obtained by calculating through the following formula:

[0029]

[0030] where N overtime represents the number of parking times exceeding the set time limit. The set time is 3 - 6 hours. N occupied is the occupancy times of all parking spaces; The overtime occupancy ratio Z b is used to analyze the phenomenon of long-term occupancy of parking spaces.

[0031] Preferably, the average return rate H dt is obtained by calculating through the following formula:

[0032]

[0033] where N return represents the number of vehicles that leave the parking area and then return within the first monitoring time. The first time is set within the time periods of 11:00 - 14:00 at noon and 17:00 - 20:00 in the evening; The average return rate H dt represents the level of business district attraction of this parking lot. Whether the vehicle owner enters for shopping, dining or handling affairs multiple times in a short period. When the same vehicle frequently returns, combined with the reverse parking search request rate, it is judged whether there is a user who returns because of misremembering the parking lot or parking space;

[0034] The reverse parking search request rate H fx is obtained by calculating through the following formula:

[0035]

[0036] where represents the total number of times the direction parking search function is used within the monitoring period T. N total represents the total number of times the direction parking search function is used within the monitoring period T.

[0037] Preferably, the road congestion coefficient D fx is obtained in the following way:

[0038]

[0039] In the formula, C entry represents the sum of the passing times of each vehicle on average at the parking lot entrance and exit; N vehicles represents the number of vehicles passing through the four roads in the southeast, northwest, east and west directions around the parking lot within a unit time. L road represents the total length of the four roads in the east, west, south and north directions around the parking lot. V speed represents the average road vehicle speed of the roads around the parking lot. T grennIndicates the green light time of the signal lights on the four roads, east, west, south, and north, around the parking lot, T total Indicates the complete green light cycle time of the signal lights on the four roads, east, west, south, and north, around the parking lot, L queue Indicates the total queuing length at the parking lot entrance and exit, V peak Indicates the vehicle flow during the real-time period in the business district within 1 km of the parking lot; V normal Indicates the vehicle flow during the peak period in the business district within 1 km of the parking lot; a1, a2, a3, a4, a5, and a6 respectively represent the weight coefficients, and the sum of the weights is 1.

[0040] Preferably, the to-be-parked person acquisition unit is used to acquire the parking demand of the j-th target to-be-parked person, including the parking space type demand Requested Tytpe j , the cost range demand R in , the parking duration demand Requested Tytpe, the distance demand d in and the queuing acceptance duration pd in ; construct a parking constraint demand set Pcm, and the expression is as follows:

[0041] Pcm = {(Requested Tytpe, R in , Requested Tytpe, d in , pd in ) j |j = 1, 2,..., m|};

[0042] Among them, m represents the total number of target to-be-parked persons.

[0043] Preferably, the sequential recommendation module includes a matching unit and a to-be-recommended unit;

[0044] The matching unit is used to perform matrix matching according to the static parking lot feature set Pcn, the dynamic parking lot feature set Pcr, and the parking constraint demand set Pcm, and calculate and obtain the geographical location matching cost Cost loc,ij of the i-th parking lot and the j-th target to-be-parked person, the parking space type matching cost Cost type,ij and the remaining parking space matching cost Cost remaining,ij through the following formula:

[0045] Cost loc,ij = diseance(Loc i , Lac j );

[0046] Cost type,ij = |Requested Tytpe j-Available Tytpe i |

[0047] Cost remaining,ij =|N Pi -requested j |

[0048] where Loc i is the latitude and longitude position of the i-th parking lot, and Lac j is the current position of the j-th target person waiting to park; Requested Tytpe j is the parking space type requirement of the j-th target person waiting to park, and Requested Tytpe j is the parking space type provided by the i-th parking lot; N Pi is the remaining parking spaces in the i-th parking lot, and requested j is the parking space demand of the j-th target person waiting to park;

[0049] And match the cost Cost loc,ij between the i-th parking lot and the j-th target person waiting to park, the parking space type matching cost Cost type,ij and the remaining parking space matching cost Cost remaining,ij , after dimensionless processing, construct the matching matrix coefficient M ij between the i-th parking lot and the j-th target person waiting to park. The matching matrix coefficient M ij represents the matching cost between the i-th parking lot and the j-th target person waiting to park, specifically as follows:

[0050] M ij = a7 * Cost loc,ij + a8 * Cost type,ij + a9 * Cost remaining,ij ;

[0051] where a7, a8, and a9 are weight coefficients;

[0052] By constructing the matrix M, minimize the total matching cost:

[0053]

[0054] where X ij is a binary variable matching coefficient, indicating whether the i-th parking lot matches the j-th target person waiting to park:

[0055]

[0056] The unit to be recommended is used to traverse all parking lots. For each parking lot, for the j-th target person waiting to park, all parking lots with Xij = 1 are screened out, and these parking lots are added to the first recommended list.

[0057] Preferably, the correction module includes an extraction unit and an evaluation unit;

[0058] The extraction unit is used to extract the parking space switching frequency Q of the i-th parking lot in the dynamic parking lot feature set Pcr pl , the average low-speed cruising duration T ds , the overtime occupancy ratio Z b , the average return rate H dt , the reverse car search request rate H fx and the road congestion coefficient D fx , after dimensionless processing, the parking cost performance index X of the i-th parking lot is calculated through the following formula jb,i ,

[0059] X jb,i = e1*Q pl +(1-(a2*T ds + a3*Z b + a4*H dt + a5*H fx + a6*D fx ));

[0060] In the formula, e1, e2, e3, e4, e5, and e6 respectively represent the weight coefficients of the parking space switching frequency Q of the i-th parking lot pl , the average low-speed cruising duration T ds , the overtime occupancy ratio Z b , the average return rate H dt , the reverse car search request rate H fx and the road congestion coefficient D fx ; and the sum of the weights is 1;

[0061] The evaluation unit is used to preset the cost performance threshold Z, and compare the parking cost performance index X of the i-th parking lot jb,i with the cost performance threshold Z to obtain the evaluation result, including:

[0062] When the parking cost performance index X of the i-th parking lot jb,i > the cost performance threshold Z, it means that the resource utilization rate of the i-th parking lot is qualified and is classified into the second-priority recommended list;

[0063] When the parking cost performance index X of the i-th parking lot jb,i ≤ the cost performance threshold Z, it means that the resource utilization rate of the i-th parking lot is unqualified.

[0064] Preferably, the correction module further includes a priority recommendation unit; the priority recommendation unit is used to extract the parking cost performance index X of the i-th parking lot in the second priority recommendation list jb,i , and sort them from largest to smallest, and preferentially recommend the highest value to the j-th target person waiting for parking, and push the corresponding navigation instruction to the j-th target person waiting for parking to realize intelligent path pushing.

[0065] The present invention provides a smart parking management system based on multi-dimensional data analysis. It has the following beneficial effects:

[0066] (1) For the smart parking management system based on multi-dimensional data analysis, through the fusion of static + dynamic data, this system combines the real-time status of the parking lot (number of available parking spaces, turnover rate, charging standard, etc.) with the parking needs of users (parking space type, charging needs, etc.) to accurately match the best parking lot, effectively improving the accuracy of the recommendation.

[0067] (2) For the smart parking management system based on multi-dimensional data analysis, it adopts a parking cost performance index calculation method to quantify the utilization of parking lot resources and ensure that the recommended parking lot has a higher cost performance. The matrix matching algorithm combines multi-dimensional feature data to ensure that the system recommends the parking position that best meets the user's needs and reduces the waste of parking time caused by inefficient recommendations.

[0068] (3) For the smart parking management system based on multi-dimensional data analysis, the system continuously updates the parking cost performance index by real-time monitoring of parking lot data, and dynamically adjusts the recommendation order according to the cost performance threshold to ensure the reasonable utilization of parking resources. The intelligent sorting and recommendation mechanism arranges the parking lots from high to low according to the parking cost performance index, and preferentially guides users into efficient parking lots to reduce the waste of parking resources. Description of the Drawings

[0069] Figure 1 It is a schematic flowchart of a smart parking management system based on multi-dimensional data analysis of the present invention. Detailed Embodiments

[0070] Next, the technical solutions in the embodiments of the present invention will be clearly and completely described in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative efforts shall fall within the protection scope of the present invention.

[0071] Embodiment 1

[0072] Please refer to Figure 1 , the present invention provides a smart parking management system based on multi-dimensional data analysis, including:

[0073] A virtual construction module is used to construct a parking lot and surrounding road scene model based on virtual reality technology, combine the basic information of the parking lot and the parking space distribution rules, simulate the vehicle parking process of multiple groups of parking lots, and generate a visual parking operation interface to display the recommended parking order and the optimal driving route.

[0074] A parking lot information collection module is used to monitor the real-time status of multiple parking lots, including the number of available parking spaces, the parking space turnover rate, the charging standard, the distribution of parking space types, the availability rate of charging parking spaces, and the congestion situation at the parking lot entrances and exits, and construct a static parking lot feature set Pcn and a dynamic parking lot feature set Pcr.

[0075] A parking person to be parked collection unit is used to collect the parking requirements of the jth target parking person to be parked and construct a parking constraint requirement set Pcm.

[0076] An order recommendation module is used to perform matrix matching according to the static parking lot feature set Pcn, the dynamic parking lot feature set Pcr, and the parking constraint requirement set Pcm, and generate a first list of recommended items according to the jth target parking person to be parked.

[0077] A correction module is used to extract the dynamic parking lot feature set Pcr in the first list of recommended items, calculate and obtain the parking cost performance index X of the ith parking lot jb,i , and preset a cost performance threshold Z. When X jb,i > Z, it is included in the second priority recommendation list, and is sorted from large to small according to X jb,i and the highest value is preferentially recommended to the jth target parking person to be parked.

[0078] In this embodiment, by simulating the parking lot and surrounding road scenes through the virtual construction module and combining the basic information of the parking lot and the parking space distribution rules, it is possible to provide personalized parking order recommendations and optimal driving routes for drivers, reducing the time cost of finding parking spaces. The parking lot information collection module can dynamically monitor the real-time status of the parking lot, including the parking space turnover rate, the number of available parking spaces, the congestion situation at the entrances and exits, etc., enabling the system to adjust the recommendation strategy based on the latest data and improve the utilization rate of parking spaces. Through the parking person to be parked collection unit, the system can obtain the parking requirements of the driver, including distance preference, cost budget, parking space type (such as charging parking space, VIP parking space, etc.), and accurately match the most suitable parking lot in combination with the static and dynamic characteristics of the parking lot, improving parking convenience.

[0079] The correction module calculates the parking cost performance index of the parking lot, and makes dynamic adjustments according to the cost performance threshold Z to ensure that the recommended parking lot takes into account cost, parking convenience and parking space stability while meeting the needs, providing the best choice for users. The system can reduce the random wandering of vehicles during peak demand periods or in hot spots, improve the rational allocation of parking resources, reduce traffic congestion caused by searching for parking spaces, and improve the overall efficiency of urban parking management.

[0080] Embodiment 2

[0081] This embodiment is an explanatory description based on Embodiment 1. Please refer to Figure 1 , specifically, the parking lot information collection module includes a static feature collection unit and a dynamic feature collection unit;

[0082] The static feature collection unit is used to collect the static information of a number of parking lots. The static information includes: the longitude and latitude position Loc of the parking lot, the total number of parking spaces N max , the parking restriction time T limtit , the upper limit of parking fee P max , the lower limit of parking fee P min , the monitoring coverage area S fg and the parking space type distribution set BL;

[0083] To construct the static parking lot feature set Pcn, the expression is as follows:

[0084] Pcm = {(Loc(x,y), N max , T limtit , P max , P min , S fg , BL) i |i = 1, 2,..., n|};

[0085] Among them, n represents the total number of parking lots;

[0086] The parking space type distribution set BL includes: the proportion of ordinary parking spaces, the proportion of VIP parking spaces, the proportion of charging parking spaces, and the proportion of barrier-free parking spaces; the parking space type distribution set is expressed as: BL = {pt, vt, ct, zt}.

[0087] The following is a data example chart of the static parking lot feature set Pcn:

[0088]

[0089]

[0090] In this embodiment, through the parking space type distribution set BL (ordinary, VIP, charging, barrier-free), the system can accurately match the appropriate parking space type according to the user's needs, improving the parking experience of users with special needs (such as new energy vehicles and disabled people). The present invention constructs a static parking lot feature set Pcn through the static feature acquisition unit, realizing accurate modeling of parking resources and providing an efficient and reasonable decision-making basis for intelligent parking management.

[0091] Embodiment 3

[0092] This embodiment is an explanatory description based on Embodiment 1. Please refer to Figure 1 , specifically, the dynamic feature acquisition unit is used to collect the dynamic change information of the parking lot. The dynamic change information includes: the remaining parking spaces N P , the parking space switching frequency Q pl , the average low-speed cruising duration T ds , the over-occupancy ratio Z b , the average turning-back rate H dt , the reverse car-finding request rate H fx and the road congestion coefficient D fx to construct a dynamic parking lot feature set Pcr;

[0093] Pcr = {(N P , Q pl , T ds , Z b , H dt , P min , H fx , D fx ) i |i = 1, 2,..., n|}; n represents the total number of parking lots.

[0094] Embodiment 2. This embodiment is an explanatory description based on Embodiment 1. Please refer to Figure 1 , specifically, the remaining parking spaces N P are obtained through the following formula:

[0095] N P = (N max - N c );

[0096] wherein, N c is the total number of real-time parked parking spaces; obtained through sensing devices such as parking lot cameras, intelligent ground locks, and geomagnetic sensors;

[0097] The parking space switching frequency Q pl is obtained through the following formula:

[0098]

[0099] Among them, N switch represents the number of times a vehicle switches parking spaces within the monitoring period T, where T represents the monitoring period; the parking space switching frequency Q pl reflects the frequency of a vehicle changing parking spaces and is related to the usage habits of the parking lot and the comfort of the parking spaces; the license plate recognition camera records the entry and exit times of vehicles in each parking space and calculates the number of position changes. RFID (Radio Frequency Identification) records the time points when a vehicle enters and leaves a parking space;

[0100] The average low-speed cruising duration T ds is obtained by calculating through the following formula:

[0101]

[0102] Among them, ∑T low-speed represents the total time of all vehicles driving at low speed in the parking lot or on the surrounding roads, and low-speed driving is identified as a speed lower than 15 km / h; N vehicle represents the total number of vehicles entering the parking lot during this time period; the average low-speed cruising duration T ds is used to analyze the length of time for a vehicle to search for a parking space in order to optimize the parking space distribution and navigation guidance; the cameras in the parking lot and on the surrounding roads record the vehicle speed data; the GPS data collects the driving trajectories of vehicles through the navigation software.

[0103] The overtime occupancy ratio Z b is obtained by calculating through the following formula:

[0104]

[0105] Among them, N overtime represents the number of parking times exceeding the set time limit, and the set time is 3 - 6 hours, N ocupied is the occupancy times of all parking spaces; the overtime occupancy ratio Z b is used to analyze the phenomenon of long-term occupancy of parking spaces.

[0106] The average return rate H dt is obtained by calculating through the following formula:

[0107]

[0108] Among them, N return represents the number of vehicles that leave the parking area and then return within the first monitoring time, and the first time is set within the time periods of 11:00 - 14:00 at noon and 17:00 - 20:00 in the evening; the average return rate H dtIndicates the level of business district attraction of the parking lot. Vehicle owners enter for shopping, dining, or handling affairs multiple times within a short period. When the same vehicle frequently turns back, combined with the reverse car-finding request rate, it is determined whether there are users returning due to misremembering the parking lot or parking space. The parking management system records the entry time of each vehicle and calculates the parking duration. The license plate recognition camera automatically counts the vehicles that exceed the parking time limit.

[0109] Reverse car-finding request rate H fx Obtained by calculating through the following formula:

[0110]

[0111] Where, Represents the total number of times the direction car-finding function is used within the monitoring period T, N total Represents the total number of times the direction car-finding function is used within the monitoring period T. The cameras at the parking lot entrance and exit, combined with the license plate recognition system, count whether the same vehicle returns to the parking lot within a short period. The parking lot mobile application (APP) records the number of times users use the reverse car-finding function. The intelligent car-finding terminal records the situation of users querying parking spaces (such as the code-scanning car-finding machine).

[0112] The road congestion coefficient D fx The acquisition method is:

[0113]

[0114] In the formula, C entry Represents the sum of the passing times of each vehicle on average at the parking lot entrance and exit; N vehicles Represents the number of vehicles passing through the four roads (east, south, west, and north) around the parking lot within a unit time, L road Represents the total length of the four roads (east, south, west, and north) around the parking lot, V speed Represents the average road vehicle speed of the roads around the parking lot; T grenn Represents the green light time of the traffic lights on the four roads (east, south, west, and north) around the parking lot, T total Represents the complete green light cycle time of the traffic lights on the four roads (east, south, west, and north) around the parking lot, L queue Represents the total length of the queue at the parking lot entrance and exit, V peak Represents the vehicle flow during the real-time period of the business district within 1KM near the parking lot; V normal Represents the vehicle flow during the peak period of the business district within 1KM near the parking lot; a1, a2, a3, a4, a5, and a6 respectively represent weight coefficients, and the sum of the weights is 1.

[0115] The cameras at the parking lot entrance and exit record the entry and exit times of vehicles and calculate the passing efficiency.

[0116] Road monitoring cameras + AI video analysis to count the number and density of road vehicles.

[0117] The traffic signal control system provides signal light duration data.

[0118] Gaode / Baidu / Tencent Map APIs to obtain real-time traffic data around parking lots.

[0119] The following is a data example chart of the dynamic parking lot feature set Pcr:

[0120]

[0121] In this embodiment, through the real-time calculation of the remaining parking spaces, the available number of parking spaces in the parking lot can be accurately monitored, providing users with the latest parking space information, reducing the time for blindly searching for parking spaces, and improving parking efficiency. Combining with the parking space switching frequency, the turnover rate of parking spaces can be analyzed, which helps the parking lot optimize the parking space scheduling and improve the utilization rate of parking spaces. The average low-speed cruising duration reflects the time when vehicles slowly search for parking spaces in the parking lot or surrounding roads. By optimizing the parking space guiding strategy (such as dynamically adjusting guiding signs, intelligent navigation recommendations, etc.), the time for vehicle owners to find parking spaces can be effectively shortened, and the operation efficiency of the parking lot can be improved. The over-occupation ratio can identify the situation of over-occupied parking spaces, helping the parking management party adjust parking strategies (such as dynamically adjusting charging standards, optimizing parking space usage rules), reducing the situation of long-term occupation of parking spaces, and improving the turnover rate of parking spaces. The average return rate can reflect the attractiveness of the business district. When users enter the parking lot multiple times in a short period, it indicates that the commercial activities in this area are relatively active. Managers can use this data to optimize parking pricing strategies, set short-term free parking discounts, etc., to increase parking revenue and the activity of the business district. The reverse car-finding request rate can measure the difficulty for users to find their cars. If this value is too high, it means there are problems such as chaotic signs and unclear parking space numbers in the parking lot. The management party can optimize the parking space numbers, add intelligent car-finding systems or provide clearer parking navigation to improve the user parking experience. The road congestion coefficient is calculated based on factors such as the passing time at the parking lot entrance and exit, the traffic flow on surrounding roads, and the signal light duration. It can be used to predict the congestion situation of the parking lot and surrounding roads, assist urban traffic management departments in optimizing signal light timing, adjusting entrance and exit traffic control measures, and improving road traffic efficiency. By comprehensively analyzing Pcn (static feature set) and Pcr (dynamic feature set), data support can be provided for the intelligent parking guidance system, realizing intelligent parking space recommendation based on big data, improving parking matching efficiency, and reducing traffic congestion caused by vehicles searching for parking spaces.

[0122] Example 4

[0123] This embodiment is an explanatory description based on Embodiment 1. Please refer to Figure 1, specifically, the to-be-parked person collection unit is used to collect the parking demands of the j-th target to-be-parked person, including the parking space type demand Requested Tytpe j , the fee range demand R in , the parking duration demand Requested Tytpe, the distance demand d in and the queuing acceptance duration pd in ; construct a parking constraint demand set Pcm, and the expression is as follows:

[0124] Pcm = {(Requested Tytpe, R in , Requested Tytpe, d in , pd in ) j | j = 1, 2,..., m |};

[0125] Among them, m represents the total number of target to-be-parked persons.

[0126]

[0127] In this embodiment, by collecting parking demands (parking space type, fee range, parking duration, distance, and queuing acceptance duration) through the to-be-parked person collection unit and constructing the parking constraint demand set Pcm, the personalized demands of different users can be accurately identified, and a more suitable parking plan can be provided.

[0128] For example, give priority to meeting the VIP parking space demands of VIP users, ensure the supply of barrier-free parking spaces for the disabled, and improve the parking convenience of special-needs groups. By analyzing parking demand data, perform intelligent matching and optimized scheduling of parking space resources, reduce parking resource waste, and maximize the revenue of the parking lot.

[0129] For example, when the demand for charging parking spaces is high, dynamically adjust the priority of idle charging parking spaces to improve the utilization rate of charging parking spaces.

[0130] The collected queuing acceptance duration parameter can be used to optimize queuing scheduling, so that the user waiting time is kept within an acceptable range and avoid user loss caused by long waiting.

[0131] For example, give priority to arranging idle parking spaces for users with a short queuing acceptance time (such as VIP parking space users) to ensure a good experience for high-value users.

[0132] Embodiment 5

[0133] This embodiment is an explanatory description based on Embodiment 1. Please refer to Figure 1 , specifically, the sequential recommendation module includes a matching unit and a to-be-recommended unit;

[0134] The matching unit is used to perform matrix matching according to the static parking lot feature set Pcn, the dynamic parking lot feature set Pcr, and the parking constraint requirement set Pcm, and calculate and obtain the geographical location matching cost Cost between the ith parking lot and the jth target person waiting to park through the following formula loc,ij , the parking space type matching cost Cost type,ij and the remaining parking space matching cost Cost remaining,ij :

[0135] Cost loc,ij =diseance(Loc i ,Lac j );

[0136] Cost type,ij =|Requested Tytpe j -Available Tytpe i |

[0137] Cost remaining,ij =|N Pi -requested j |

[0138] In the formula, Loc i is the latitude and longitude position of the ith parking lot, and Lac j is the current position of the jth target person waiting to park; Requested Tytpe j is the parking space type requirement of the jth target person waiting to park, and Requested Tytpe j is the parking space type provided by the ith parking lot; N Pi is the remaining parking spaces of the ith parking lot, and requested j is the parking space requirement of the jth target person waiting to park;

[0139] And according to the geographical location matching cost Cost loc,ij between the ith parking lot and the jth target person waiting to park, the parking space type matching cost Cost type,ij and the remaining parking space matching cost Cost remaining,ij , after dimensionless processing, the matching matrix coefficient M ij between the ith parking lot and the jth target person waiting to park is constructed. The matching matrix coefficient M ij represents the matching cost between the ith parking lot and the jth target person waiting to park, which is specifically as follows:

[0140] M ij =a7*Cost loc,ij +a8*Cost type,ij+a9*Cost remaining,ij ;

[0141] Wherein, a7, a8, and a9 represent weight coefficients;

[0142] By constructing matrix M, the total matching cost is minimized:

[0143]

[0144] Where X ij is a binary variable matching coefficient, indicating whether the i-th parking lot matches the j-th target person waiting to park:

[0145]

[0146] The to-be-recommended unit is used to traverse all parking lots. For each parking lot, for the j-th target person waiting to park, all parking lots where X ij = 1 are screened out, and these parking lots are added to the first to-be-recommended list.

[0147] In this embodiment, by the matching unit constructing a parking matching matrix, integrating the geographical location matching cost, the parking space type matching cost, and the remaining parking space matching cost, optimal matching is achieved, and the efficiency of users finding suitable parking spaces is improved.

[0148] For example, for a target person waiting to park, the nearest and demand-compliant parking spaces are preferentially recommended, avoiding detours or parking difficulties caused by inaccurate matching. Through the matrix matching method, combining the static parking lot feature set Pcn (parking lot location, parking space type), the dynamic parking lot feature set Pcr, and the parking constraint demand set Pcm, the accuracy of the matching is ensured.

[0149] For example, users with special needs (such as the need for barrier-free parking spaces or charging parking spaces) are preferentially satisfied, improving the fairness and utilization rate of parking resources.

[0150] By setting the weight coefficients a7, a8, and a9 to adjust the matching priority, the matching rules are made flexible and can adapt to different user scenarios. For example, during peak hours, the geographical location weight can be increased to preferentially recommend the nearest parking spaces; during off-peak hours, the parking space type matching weight can be increased to meet the personalized needs of users. By using the method of minimizing the total matching cost, the matching of the entire parking system is made more efficient, resource waste is reduced, and the intelligent level of parking scheduling is improved.

[0151] Embodiment 6

[0152] This embodiment is an explanatory description based on Embodiment 1. Please refer to Figure 1 , specifically, the correction module includes an extraction unit and an evaluation unit;

[0153] The extraction unit is used to extract the parking space switching frequency Q of the i-th parking lot in the dynamic parking lot feature set Pcr pl , the average low-speed cruising duration T ds , the overtime occupancy ratio Z b , the average turning-back rate H dt , the reverse car-finding request rate H fx and the road congestion coefficient D fx , perform dimensionless processing, and calculate the parking cost performance index X of the i-th parking lot through the following formula jb,i ,

[0154] X jb,i = e1 * Q pl +(1 - (a2 * T ds + a3 * Z b + a4 * H dt + a5 * H fx + a6 * D fx ));

[0155] In the formula, e1, e2, e3, e4, e5, and e6 respectively represent the weight coefficients of the parking space switching frequency Q of the i-th parking lot pl , the average low-speed cruising duration T ds , the overtime occupancy ratio Z b , the average turning-back rate H dt , the reverse car-finding request rate H fx and the road congestion coefficient D fx ; and the sum of the weights is 1;

[0156] The following is an example chart of the calculation data of the parking cost performance index:

[0157]

[0158] In this embodiment, through the extraction unit, six key parameters (parking space switching frequency, average low-speed cruising duration, overtime occupancy ratio, average turning-back rate, reverse car-finding request rate, road congestion coefficient) are extracted from the dynamic parking lot feature set Pc, and a parking cost performance index is constructed to realize the dynamic evaluation of the parking lot. For example, a parking lot with a high parking space switching frequency may be more suitable for short-term parking, while a parking lot with a high overtime occupancy ratio may not be suitable for short-term parking needs. By calculating the parking cost performance index, users can quickly understand the cost performance of different parking lots, reduce inefficient parking behaviors such as long-term cruising and repeated car-finding caused by wrong choices, and improve parking efficiency.

[0159] For example, the parking lot with the highest cost-performance index (such as Parking Lot No. 005 with an index of 0.88) is more suitable for users to choose, thus reducing the time waste caused by inefficient parking. Through the calculation of the parking cost-performance index, the management system can adjust the parking recommendation strategy specifically, guide vehicles to divert to parking lots with high cost-performance, avoid overloading or inefficient use of certain parking lots, and improve the overall utilization rate.

[0160] For example, when the reverse car-finding request rate of a certain parking lot is relatively high, it indicates that users have difficulty finding their cars in this area. The recommendation priority can be reduced to guide users to a better parking lot.

[0161] Embodiment 7

[0162] This embodiment is an explanatory description carried out in Embodiment 6. Please refer to Figure 1 , specifically, the evaluation unit is used to preset a cost-performance threshold Z, and compare the parking cost-performance index X jb,i of the i-th parking lot with the cost-performance threshold Z to obtain an evaluation result, including:

[0163] When the parking cost-performance index X jb,i of the i-th parking lot > the cost-performance threshold Z, it means that the resource utilization rate of the i-th parking lot is qualified and is classified into the second-priority recommendation list;

[0164] When the parking cost-performance index X jb,i of the i-th parking lot ≤ the cost-performance threshold Z, it means that the resource utilization rate of the i-th parking lot is unqualified.

[0165] Specifically, the correction module further includes a priority recommendation unit; the priority recommendation unit is used to extract the parking cost-performance index X jb,i of the i-th parking lot in the second-priority recommendation list, sort it from largest to smallest, and preferentially recommend the highest value to the j-th target person waiting to park, and push the corresponding navigation instruction to the j-th target person waiting to park to achieve intelligent path pushing.

[0166] The following is an example chart of the second-priority recommendation list:

[0167] Priority Parking Lot Number Cost Performance Index Recommendation Status 1 005 0.88 Highly Recommended 2 002 0.85 Recommended 3 003 0.80 Recommended 4 004 0.78 Alternative 5 001 0.72 Alternative

[0168] In this embodiment, through the evaluation unit, the parking cost-performance index of each parking lot is compared with the preset cost-performance threshold Z to ensure that the recommended parking lots have a high resource utilization rate, thereby optimizing the allocation of parking resources.

[0169] For example, if the cost-performance threshold is set to Z = 0.75, parking lots with a cost-performance index ≤ 0.75 will not be recommended, reducing the likelihood of users parking in inefficient lots. · Through the priority recommendation unit, all qualified parking lots are sorted and recommended according to the cost-performance index from high to low, ensuring that users can obtain the optimal parking lot first.

[0170] For example, in the recommendation list, Parking Lot 005 (cost-performance index 0.88) is marked as highly recommended. In contrast, Parking Lot 004 (cost-performance index 0.78) is only an alternative, ensuring the quality of the recommended parking spaces.

[0171] Since the system has completed intelligent screening and sorting, users do not need to compare the conditions of parking lots one by one and can directly select according to the recommendation level, improving the parking decision-making efficiency.

[0172] For example, users only need to focus on parking lots at the "highly recommended" or "recommended" level, avoiding wasting time in inefficient parking lots.

[0173] Since the parking cost-performance index is based on real-time dynamic data (such as the frequency of parking space switching, the ratio of overtime occupancy, the road congestion coefficient, etc.), the ranking of parking lots can be automatically adjusted according to environmental changes, thus optimizing parking management.

[0174] For example, during peak hours, a parking lot with a high road congestion coefficient (such as Parking Lot 001) may be downgraded, while during off-peak hours, the ranking of this parking lot may rise, improving the system's adaptability.

[0175] Through the intelligent route push function, the system can automatically push navigation instructions for the best parking lot to users according to the priority recommendation results, reducing the trouble of users looking for parking spaces in unfamiliar areas and improving the parking experience.

[0176] For example, when a user receives Parking Lot 005 with a "highly recommended" label, the system will immediately provide the best navigation route to help the user reach the destination efficiently.

[0177] By scientifically matching parking lots with relatively high cost-performance and preferentially guiding users to park in them, the system can improve the overall utilization rate of parking lots and reduce the waste of parking resources.

[0178] For example, high-cost-performance parking lots (such as No. 005) are highly recommended, making their resources fully utilized, while low-cost-performance parking lots (cost-performance index below 0.75) are optimized for management, improving the overall operation efficiency.

[0179] The setting of the threshold size is for the convenience of comparison. Regarding the size of the threshold, it depends on the amount of sample data and the number of base quantities set by those skilled in the art for each group of sample data; as long as the proportional relationship between the parameters and the quantified values is not affected.

[0180] The above formulas are all obtained by collecting a large amount of data for software simulation and selecting a formula close to the true value. The coefficients in the formula are set by those skilled in the art according to the actual situation. As mentioned above, the above is only a preferred specific implementation manner of the present invention, but the protection scope of the present invention is not limited thereto. Any person skilled in the art within the technical scope disclosed by the present invention, according to the technical solution of the present invention and its inventive concept, makes equivalent substitutions or changes, and should be covered by the protection scope of the present invention.

Claims

1. A smart parking management system based on multi-dimensional data analysis, characterized in that, Including: A virtual construction module for constructing a parking lot and surrounding road scene model based on virtual reality technology, combining the basic information of the parking lot and the parking space distribution rules, simulating the vehicle parking process of multiple groups of parking lots, and generating a visual parking operation interface to display the recommended parking order and the optimal driving route; A parking lot information collection module for monitoring the real-time status of multiple parking lots, including the number of available parking spaces in the parking lot, the parking space turnover rate, the charging standard, the parking space type distribution, the availability rate of charging parking spaces, and the congestion situation at the parking lot entrances and exits, and constructing a static parking lot feature set Pcn and a dynamic parking lot feature set Pcr; A parking person to be parked collection unit for collecting the parking requirements of the jth target parking person to be parked and constructing a parking constraint requirement set Pcm; A sequence recommendation module for performing matrix matching according to the static parking lot feature set Pcn, the dynamic parking lot feature set Pcr, and the parking constraint requirement set Pcm, and generating a first recommended list according to the jth target parking person to be parked; A correction module, which is used to extract the dynamic parking lot feature set Pcr in the first recommended list, calculate and obtain the parking cost performance index X of the i-th parking lot jb,i , and preset a cost performance threshold Z. When X jb,i > Z, it is included in the second-priority recommended list, and sorted from large to small according to X jb,i , and the highest value is preferentially recommended and sent to the j-th target person waiting for parking.

2. The intelligent parking management system based on multi-dimensional data analysis according to claim 1, wherein The parking lot information collection module includes a static feature collection unit and a dynamic feature collection unit; The static feature acquisition unit is used to acquire the static information of several parking lots, and the static information includes: the longitude and latitude position Loc of the parking lot, the total number of parking spaces N in the parking lot max , the parking restriction time T limtit , the upper limit P of the parking fee max , the lower limit P of the parking fee min , the monitoring coverage area S fg and the set BL of the distribution of parking space types; To construct a static parking lot feature set Pcn, the expression is as follows: Pcm = {(Loc(x,y), N max , T limtit , P max , P min , S fg , BL) i | i = 1, 2,..., n |}; Where n represents the total number of parking lots; The parking space type distribution set BL includes: the proportion of ordinary parking spaces, the proportion of VIP parking spaces, the proportion of charging parking spaces, and the proportion of barrier-free parking spaces; the parking space type distribution set is expressed as: BL = {pt, vt, ct, zt}.

3. The intelligent parking management system based on multi-dimensional data analysis according to claim 2, characterized in that, The dynamic feature acquisition unit is used to acquire the dynamic change information of the parking lot, and the dynamic change information includes: the remaining number of parking spaces N P , the parking space switching frequency Q pl , the average low-speed cruising duration T ds , the overtime occupancy ratio Z b , the average turning-back rate H dt , the reverse parking search request rate H fx and the road congestion coefficient D fx to construct a dynamic parking lot feature set Pcr; Pcr = {(N P , Q pl , T ds , Z b , H dt , P min , H fx , D fx ) i | i = 1, 2,..., n |}; n represents the total number of parking lots.

4. A smart parking management system based on multi-dimensional data analysis according to claim 3, characterized in that, The remaining parking spaces N P are calculated and obtained by the following formula: N P = (N max - N c ); Among them, N c is the total number of real-time parked spaces; The parking space switching frequency Q pl is calculated and obtained through the following formula: Among them, N switch represents the number of times the vehicle switches parking spaces during the monitoring period T, where T represents the monitoring period; the parking space switching frequency Q pl reflects the frequency of the vehicle changing parking spaces and is related to the usage habits of the parking lot and the comfort of the parking spaces; The average low-speed cruising duration T ds is calculated by the following formula: Among them, ∑T low-speed represents the total time for all vehicles to drive at low speed in the parking lot or on the surrounding roads, and low-speed driving is recognized as a speed lower than 15 km / h; N vehicle represents the total number of vehicles entering the parking lot during this time period; the average low-speed cruising duration T ds is used to analyze the time taken for vehicles to search for parking spaces, so as to optimize the parking space distribution and navigation guidance; The timeout occupancy ratio Z b is calculated and obtained through the following formula: Among them, N overtime represents the number of parking times exceeding the set time limit. The set time is 3 - 6 hours. N occupied is the number of occupancy times of all parking spaces; the over - time occupancy ratio Z b is used to analyze the phenomenon of long - term occupancy of parking spaces.

5. The intelligent parking management system based on multi-dimensional data analysis according to claim 4, characterized in that, The average return rate H dt is calculated by the following formula: Among them, N return represents the number of vehicles that leave the parking area and then return during the first monitoring time. The first time is set within the time periods of 11:00 - 14:00 at noon and 17:00 - 20:00 in the evening; the average return rate H dt represents the level of the business district attraction of the parking lot. The car owners enter for shopping, dining or handling affairs multiple times in a short period. When the same vehicle frequently returns, combined with the reverse car-finding request rate, it is judged whether there are users who return because they misremember the parking lot or parking space; Reverse car-finding request rate H fx Obtained by calculating through the following formula: Among them, represents the total number of times the direction-based car finding function is used within the monitoring period T, N total represents the total number of times the direction-based car finding function is used within the monitoring period T.

6. The intelligent parking management system based on multi-dimensional data analysis according to claim 5, characterized in that, The road congestion coefficient D fx is obtained as follows: Where C entry represents the sum of the passing times of each vehicle on average at the entrance and exit of the parking lot; N vehicles represents the number of vehicles passing through the four roads (east, south, west, and north) around the parking lot within a unit time, L road represents the total length of the four roads (east, south, west, and north) around the parking lot, V speed represents the average road speed of the vehicles on the roads around the parking lot; T grenn represents the green light time of the traffic lights on the four roads (east, south, west, and north) around the parking lot, T total represents the complete green light cycle time of the traffic lights on the four roads (east, south, west, and north) around the parking lot, L queue represents the total length of the queue at the entrance and exit of the parking lot, V peak represents the vehicle flow during the real-time period in the business district within 1 km near the parking lot; V normal represents the vehicle flow during the peak period in the business district within 1 km near the parking lot; a1, a2, a3, a4, a5, and a6 respectively represent the weight coefficients, and the sum of the weights is 1.

7. An intelligent parking management system based on multi-dimensional data analysis according to claim 1, characterized in that The to-be-parked person collection unit is used to collect the parking needs of the j-th target to-be-parked person, including the parking space type requirement RequestedTytpe j , the cost range requirement R in , the parking duration requirement RequestedTytpe, the distance requirement d in and the queuing acceptance duration pd in ; construct a parking constraint requirement set Pcm, and the expression is as follows: Pcm = {(RequestedTytpe, R in , Requested Tytpe, d in , pd in ) j | j = 1, 2,..., m |}; Where m represents the total number of target parking persons to be parked.

8. A smart parking management system based on multi-dimensional data analysis according to claim 1, characterized in that The sequence recommendation module includes a matching unit and a recommended unit; The matching unit is used to perform matrix matching according to the static parking lot feature set Pcn, the dynamic parking lot feature set Pcr, and the parking constraint requirement set Pcm, and calculate and obtain the geographical location matching cost Cost between the ith parking lot and the jth target person to be parked through the following formula loc,ij , the parking space type matching cost Cost type,ij , and the remaining parking space matching cost Cost remaining,ij : Cost loc,ij = diseance(Loc i , Lac j ); Cost type,ij = |RequestedTytpe j - Available Tytpe i | Where Loc i is the latitude and longitude position of the i-th parking lot, and Lac j is the current position of the j-th target person waiting to park; RequestedTytpe j is the parking space type requirement of the j-th target person waiting to park, and Requested Tytpe j is the parking space type provided by the i-th parking lot; is the remaining parking spaces of the i-th parking lot, and requested j is the parking space requirement of the j-th target person waiting to park; And match the cost Cost according to the geographical locations of the i-th parking lot and the j-th target person waiting to park loc,ij , the cost Cost of matching the parking space type type,ij and the cost Cost of matching the remaining parking spaces remaining,ij . After dimensionless processing, construct the matching matrix coefficient M between the i-th parking lot and the j-th target person waiting to park ij . The matching matrix coefficient M ij represents the matching cost between the i-th parking lot and the j-th target person waiting to park, which is specifically as follows: M ij = a7 * Cost loc,ij + a8 * Cost type,ij + a9 * Cost remaining,ij ; In the formula, a7, a8, and a9 represent weight coefficients; By constructing a matrix M, the total matching cost is minimized: where X ij is a binary variable matching coefficient, indicating whether the i-th parking lot matches the j-th target person waiting to park: The recommended unit is used to traverse all parking lots. For each parking lot, for the jth target parking person to be parked, all parking lots with Xij = 1 are screened out, and these parking lots are added to the first recommended list.

9. The intelligent parking management system based on multi-dimensional data analysis according to claim 1, wherein, The correction module includes an extraction unit and an evaluation unit; The extraction unit is used to extract the parking space switching frequency Q of the i-th parking lot in the dynamic parking lot feature set Pcr pl , the average low-speed cruising duration T ds , the overtime occupancy ratio Z b , the average turning-back rate H dt , the reverse parking search request rate H fx and the road congestion coefficient D fx , dimensionless processing, and the parking cost performance index X of the i-th parking lot is calculated through the following formula jb,i , X jb,i = e1 * Q pl + (1 - (a2 * T ds + a3 * Z b + a4 * H dt + a5 * H fx + a6 * D fx )); Wherein, e1, e2, e3, e4, e5, and e6 respectively represent the switching frequency Q of the parking spaces in the i-th parking lot pl , the average low-speed cruising duration T ds , the over-occupation ratio Z b , the average turning-back rate H dt , the reverse car-finding request rate H fx and the road congestion coefficient D fx ; and the sum of the weight coefficients is 1; The evaluation unit is used to preset a cost performance threshold Z and compare the parking cost performance index X of the i-th parking lot jb,i with the cost performance threshold Z to obtain an evaluation result, including: When the parking cost performance index X of the i-th parking lot jb,i > the cost performance threshold Z, it means that the resource utilization rate of the i-th parking lot is qualified and is classified into the second-priority recommendation list; When the parking cost performance index X of the i-th parking lot jb,i ≤ the cost performance threshold Z, it means that the resource utilization rate of the i-th parking lot is unqualified.

10. A smart parking management system based on multi-dimensional data analysis according to claim 9, characterized in that, The correction module further includes a priority recommendation unit; the priority recommendation unit is used to extract the parking cost performance index X of the i-th parking lot in the second priority recommendation list jb,i , and sort them from largest to smallest, and preferentially recommend the highest value to the j-th target person waiting for parking, and push the corresponding navigation instruction to the j-th target person waiting for parking to achieve intelligent path pushing.

Images