Parking space real-time occupation sensing and predicting system

Abstract

The invention discloses a parking space real-time occupation sensing and predicting system, and belongs to the technical field of intelligent traffic. Comprising an event queue establishing module for establishing an event queue; the influence sub-value acquisition module is used for generating a dynamic event influence sub-value; the increase coefficient generation module is used for generating a predictive increase coefficient coupled with the real-time event information; the feedback correction module is used for performing feedback correction; and the result generation module is used for generating a real-time occupancy prediction result. According to the method, the association queue of the event and the city grid is established, the dynamic influence of the event is quantified, the prediction coefficient of the coupling real-time event is generated, the residual error feedback correction model is utilized, and the data is fused to realize the microscopic prediction of the single parking space, so that the effects of accurately sensing and predicting the parking space occupation and improving the prediction adaptability and precision are achieved; the problems that in the prior art, the influence of crowd gathering events is not fully considered, the prediction precision is not high, dynamic correction is lacked, and accurate prediction of a single berth is difficult to achieve are solved.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of intelligent transportation, and in particular to a parking space real-time occupancy perception and prediction system. BACKGROUND

[0002] With the acceleration of urbanization and the continuous growth of the number of motor vehicles, the contradiction between supply and demand of urban parking resources is increasingly prominent. How to efficiently use limited parking resources and improve the use efficiency of parking spaces has become an important issue to alleviate urban traffic congestion and improve travel experience. Real-time and accurate parking space occupancy perception and prediction can provide timely parking information guidance for vehicle owners and optimize parking resource allocation, which is one of the core functions of an intelligent parking management system.

[0003] Existing parking space prediction systems mostly predict by analyzing historical parking data and combining time period characteristics. Some systems introduce external factors such as weather for auxiliary correction.

[0004] In the prior art, the sudden impact of various events with crowd gathering effect (such as concerts, sports events, and large-scale exhibitions) on parking demand in the city is not fully considered, resulting in a large deviation between the prediction results and the actual situation during the event. At the same time, existing predictions mostly stay at the regional macro level and are difficult to achieve accurate prediction of a single parking space. Moreover, there is a lack of dynamic feedback correction mechanism, which cannot adjust the prediction model in real time according to the actual situation, resulting in low prediction accuracy and poor adaptability. SUMMARY

[0005] The embodiments of the present application provide a parking space real-time occupancy perception and prediction system, which solves the problem of low prediction accuracy and poor adaptability caused by insufficient consideration of the impact of crowd gathering events, prediction staying at the macro level, and lack of dynamic correction in the prior art, and achieves accurate prediction of parking spaces from the macro grid to the micro parking space, improving the accuracy and adaptability of the prediction.

[0006] The embodiments of the present application provide a parking space real-time occupancy perception and prediction system, which includes: an event queue establishment module configured to obtain historical event information and real-time event information of events with crowd gathering effect occurring in the city, and perform spatial mapping of the event information and the pre-divided city grid to establish an event queue associated with each grid;

[0007] An influence sub-value acquisition module is configured to obtain initial influence of the historical event information from the multi-source heterogeneous data in the historical event information, and generate dynamic event influence sub-values for each city grid.

[0008] The increasing coefficient generation module is configured to generate a predictive increasing coefficient coupled with the real-time event information by using a preset urban grid occupancy coefficient increasing model, and derive a prediction result based on the predictive increasing coefficient;

[0009] The feedback correction module is configured to monitor the actual occupancy rate of each urban grid in real time, correct the urban grid occupancy coefficient model and the dynamic event influence sub-value based on the residual error between the derived prediction result and the actual occupancy rate, and perform feedback correction;

[0010] The result generation module is configured to fuse the corrected predictive increasing coefficient and real-time state data of each parking space sensor in the parking lot, input the fused data into a microscopic prediction model, and generate a real-time occupancy prediction result specific to a single parking space.

[0011] Further, the step of spatially mapping the event information to the pre-divided urban grids and establishing an event queue associated with each grid includes:

[0012] Obtaining geographic coordinate data within a city, dividing the city area into a plurality of regular grid units according to a preset grid size, and assigning a unique grid identification code to each grid unit;

[0013] Based on the received historical event information and real-time event information, the geographic location coordinates of the events are extracted, and the grid unit to which each event belongs is determined through a coordinate matching algorithm;

[0014] An independent event queue data structure is created for each grid unit, and the events matched to the grid are stored in the corresponding queue in chronological order.

[0015] Further, the step of generating a dynamic event influence sub-value for each urban grid based on the multi-source heterogeneous data in the historical event information includes:

[0016] The multi-source heterogeneous data of the historical event information is collected and integrated, and preprocessed, and the multi-source heterogeneous data is converted into a feature vector;

[0017] Based on the converted feature vector, the actual parking demand impact generated by each historical event is obtained, and the baseline initial influence score of historical events of different types and different scales is obtained;

[0018] When a new event appears in the real-time event queue, the features of the new event are matched with the features of the historical events, and the matching successful historical event features are obtained;

[0019] The initial influence score corresponding to the matching successful historical event features is called, and the real-time dynamic features of the new event are combined to generate a new event influence sub-value, and the influence sub-value is assigned to each urban grid affected by the event.

[0020] Further, the preset urban grid occupancy coefficient increasing model expression is:

[0021]

[0022] In the formula, G i (t) is the predicted occupancy coefficient of the urban grid unit i at time t, B i (t) is the reference historical occupancy coefficient of the urban grid unit i at time t, N is the total number of activity events at time t whose influence can radiate to the grid unit i, S k is the dynamic event influence sub-value possessed by the kth event in the queue, W(d ik ,Δt k ) is a space-time weight decay function, d ik is the spatial distance between the center point of the event k and the center point of the grid unit i, Δt k is the time difference value between time t and the starting time of the event k.

[0023] Further, the step of deriving the prediction result based on the predicted increasing coefficient comprises:

[0024] For each urban grid that needs to be predicted, the predicted increasing coefficient value thereof is extracted;

[0025] From the pre-stored urban infrastructure database, a parking list associated with the unique identifier of the urban grid is queried and called to obtain the total of all parking resources in the urban grid area, and a preliminary predicted occupied parking number is obtained.

[0026] Based on the obtained preliminary predicted occupied parking number, the result is converted into a predicted occupancy rate;

[0027] The obtained predicted occupancy rate of the urban grid, together with the corresponding urban grid identifier and the predicted time point information, is stored as a data record.

[0028] Further, the step of combining the residual error between the derived prediction result and the actual occupancy rate by monitoring the actual occupancy rate of each urban grid in real time comprises:

[0029] Real-time state data from all parking sensors in each urban grid of the city is continuously received through the parking sensor network;

[0030] For each urban grid, the states of all the parking spaces inside it are aggregated in real time to obtain the number of occupied parking spaces at the current time, and the actual occupancy rate of the urban grid at the current time is obtained;

[0031] Based on the obtained predicted occupancy rate percentage, the actual occupancy rate is combined to obtain the residual error between the prediction result and the actual occupancy rate.

[0032] Further, the step of generating the real-time occupancy prediction result specific to the single parking space comprises:

[0033] obtaining a predictive increase coefficient of the feedback-corrected urban grid;

[0034] obtaining the occupancy of each parking space at the current time through the parking space sensor in the parking lot in the urban grid;

[0035] constructing a feature vector for each parking space, and fusing the corrected predictive increase coefficient with the micro feature vector of each parking space to obtain multi-dimensional input data containing demand pressure and parking space attributes;

[0036] transmitting the multi-dimensional input data to a preset prediction model, and outputting a probability value of the single parking space being occupied at a future specified time point.

[0037] Further, the influence score is obtained according to the following formula:

[0038]

[0039] In the formula, S is the reference initial influence score of the event k, is the M-dimensional normalized feature vector of the event k, v k,m is the normalized value of the mth feature, is the M-dimensional feature weight vector, ω m is the weight coefficient of the mth feature, and M is the total number of features.

[0040] Further, the influence sub-value is obtained according to the following formula:

[0041]

[0042] In the formula, S k (t) is the influence sub-value of the event k at time t, is the reference initial influence score of the event k, α is a social heat amplification coefficient, R k (t) is the real-time social heat of the event k at time t, R base is a heat normalization reference value, β is a coupling coefficient of aggregation intensity, Ψ k (t) is the cumulative aggregation intensity of the event k at time t, Θ gather is an aggregation effect determination threshold, Γ(Δt k ) is a time decay factor.

[0043] One or more technical solutions provided in the embodiments of the present application have at least the following technical effects or advantages:

[0044] 1. By establishing an event-driven, gridded dynamic prediction model, city-level macro events, such as concerts and sporting events, are coupled with micro-level parking space occupancy status, thereby achieving a precise response to sudden crowd gathering effects and effectively solving the problem of delayed parking demand prediction caused by ignoring the impact of real-time events in existing technologies.

[0045] 2. By spatially mapping event information with crowd gathering effects to urban grids and establishing a queue of related events, the spatial positioning of the impact of events on parking demand can be achieved, thereby improving the spatial correlation of parking space occupancy prediction and effectively solving the problem of insufficient prediction accuracy caused by the ambiguity of the event impact range in existing technologies.

[0046] 3. By extracting the initial impact of historical events from multi-source heterogeneous data and combining it with real-time event characteristics to generate dynamic event impact sub-values, dynamic quantitative assessment of the impact of events on parking demand is achieved, thereby enabling the prediction model to adapt to different types and scales of events. Attached Figure Description

[0047] Figure 1 This is a structural diagram of a real-time parking space occupancy perception and prediction system provided in an embodiment of this application. Detailed Implementation

[0048] This application provides a real-time parking space occupancy perception and prediction system, which solves the problems of insufficient consideration of the impact of crowd gathering events, low prediction accuracy, lack of dynamic correction, and difficulty in achieving accurate prediction of a single parking space in the prior art. By establishing a correlation queue between events and urban grids, quantifying the dynamic influence of events, generating prediction coefficients for coupled real-time events, using residual feedback to correct the model, and fusing data to achieve micro-prediction, the system achieves accurate perception and prediction of parking space occupancy, improving prediction adaptability and accuracy.

[0049] To better understand the above technical solutions, the following will provide a detailed explanation of the technical solutions in conjunction with the accompanying drawings and specific implementation methods.

[0050] like Figure 1 The diagram shown is a structural diagram of a real-time parking space occupancy perception and prediction system provided in an embodiment of this application. It includes: an event queue establishment module: used to acquire historical and real-time event information with crowd gathering effects occurring within the city, and spatially map the event information to pre-divided urban grids to establish event queues associated with each grid; the event queues are updated in real-time using real-time event information, and the event queues are used to provide a time-sequential event context for each urban grid to support initial influence calculation and real-time updating and correction of dynamic event influence sub-values.

[0051] The step of judging whether the event has a crowd gathering effect is: based on historical passenger flow, vehicle access and exit amount and TPI and other multi-source data obtained by physical equipment such as traffic flow detector, roadside radar and video traffic monitoring system, a comparison model is established, the real-time monitoring data during the current event is compared with the historical baseline data of the same place and time period, when the difference value exceeds the preset gathering threshold and the duration reaches the set period, it can be judged that the event has a crowd gathering effect.

[0052] Influence sub-value acquisition module: for obtaining the initial influence of historical event information from multi-source heterogeneous data in the historical event information, and generating dynamic event influence sub-values for each city grid in combination with historical parking behavior data of each city grid;

[0053] Increase coefficient generation module: for generating a predictive increase coefficient coupled with real-time event information through a preset city grid occupation coefficient increase model, and deriving a prediction result based on the predictive increase coefficient;

[0054] Feedback correction module: for real-time monitoring of the actual occupation rate of each city grid, and performing feedback correction on the city grid occupation coefficient model and the dynamic event influence sub-value in combination with the residual error between the derived prediction result and the actual occupation rate;

[0055] The actual occupation rate of each city grid can be obtained by the following steps: obtaining vehicle access and stay state data through sensing devices (such as geomagnetic sensors, video recognition cameras, ultrasonic detectors, etc.) arranged around parking lots or roads, and collecting and counting the collected occupation signals according to the spatially divided grid units, calculating the proportion of the number of occupied parking spaces in the total number of parking spaces in the grid in each monitoring period, which is the actual occupation rate of the grid. The occupation rate is updated and stored in real time through edge computing or cloud platform for subsequent crowd gathering effect analysis and occupation prediction model calling.

[0056] Result generation module: for fusing the corrected predictive increase coefficient and the real-time state data of each parking space sensor in the parking lot, and inputting into the microscopic prediction model to generate real-time occupation prediction results of a single parking space.

[0057] Further, the steps of spatially mapping the event information and the pre-divided city grids to establish an event queue associated with each grid include:

[0058] Obtain geographic coordinate data within the city range, divide the city area into a plurality of regular grid units according to a preset grid size, and assign a unique grid identification code to each grid unit;

[0059] Based on the received historical event information and real-time event information, the geographical position coordinates of the event occurrence are extracted, and the grid unit to which each event belongs is determined through a coordinate matching algorithm;

[0060] An independent event queue data structure is created for each grid unit, and the events matched to the grid are stored in the corresponding queue in chronological order;

[0061] Through the dynamic update of the event queue, new events are automatically added to the tail of the queue, and expired events are removed from the head of the queue, ensuring that the event information stored in each grid queue remains timely and complete.

[0062] Further, through the multi-source heterogeneous data in the historical event information, the initial influence of the historical event information is obtained, and the steps of generating a dynamic event influence sub-value for each city grid include:

[0063] Multi-source heterogeneous data of historical event information is collected and integrated, and preprocessed to convert multi-source heterogeneous data into feature vectors; the multi-source heterogeneous data includes official classification of events, media report heat, social network discussion volume, traffic control announcement, and historical parking occupancy data of the corresponding grid during the event occurrence period.

[0064] Through the cleaning and standardization process of multi-source heterogeneous data, different structured data such as text, numerical value, and time series are converted into unified and quantifiable feature vectors.

[0065] Based on the converted feature vectors, the actual parking demand impact generated by each historical event is obtained, and the benchmark initial influence score of different types and different sizes of historical events is obtained;

[0066] When a new event appears in the real-time event queue, the features of the new event are matched with the features of the historical events, and the matching successful historical event features are obtained;

[0067] The initial influence score corresponding to the matching successful historical event feature is called, and the real-time dynamic feature of the new event is combined to generate a new event influence sub-value, and the influence sub-value is assigned to each city grid affected by the event.

[0068] Further, the preset city grid occupancy coefficient increasing model expression is:

[0069]

[0070] In the formula, G i (t) is the final predicted occupancy coefficient of city grid unit i at time t in the future, B i(t) is the baseline historical occupancy factor of the urban grid cell i at time t, which is derived from the regular berth occupancy data without the influence of special events, N is the total number of active events at time t whose influence can radiate to grid cell i, S k is the dynamic event influence sub-value that the kth event in the queue has, which is used to quantify the driving ability of the independent event on parking demand, W(d ik , Δt k ) is a space-time weight decay function used to calculate the decay effect of the influence of event k over space and time, d ik is the spatial distance between the center point of event k and the center point of grid cell i, Δt k is the time difference between time t and the starting time of event k.

[0071] Further, the step of deriving the prediction result based on the predictive increase coefficient comprises:

[0072] For each urban grid that needs to be predicted, extract its predictive increase coefficient value;

[0073] From the pre-stored urban infrastructure database, query and call the berth list associated with the unique identifier of the urban grid, obtain the sum of all parking berth resources in the urban grid area, multiply the predictive increase coefficient of the urban grid by the actual berth sum to obtain the preliminary predicted occupied berth number by performing multiplication operation.

[0074] Based on the obtained preliminary predicted occupied berth number, further perform a division operation, divide the predicted occupied berth number by the berth sum, and convert the result into a predicted occupancy rate;

[0075] Store the obtained predicted occupancy rate of the urban grid together with the corresponding urban grid identifier and predicted time point information as a data record.

[0076] Further, the step of combining the residual error between the derived prediction result and the actual occupancy rate comprises:

[0077] Continuously receive real-time state data from all berth sensors in each urban grid of the city through the berth sensor network;

[0078] For each urban grid, aggregate the states of all berths inside it in real time, obtain the number of occupied berths at the current time, and obtain the actual occupancy rate of the urban grid at the current time by dividing the number of occupied berths by the total number of berths in the grid;

[0079] Based on the obtained predicted occupancy rate percentage, combine the actual occupancy rate, subtract the calculated actual occupancy rate from the predicted occupancy rate to obtain the residual error between the prediction result and the actual occupancy rate;

[0080] The residual between the prediction result and the actual occupancy rate is bound with the corresponding urban grid identifier and timestamp, and is stored in a time series database dedicated to recording errors, thereby providing data input for the subsequent feedback correction module.

[0081] Further, the step of generating real-time occupancy prediction results specific to a single parking space includes:

[0082] Obtaining the predictive increase coefficient of the urban grid after feedback correction;

[0083] Obtaining the occupancy of each parking space at the current time through the parking space sensors in the parking lot in the urban grid;

[0084] A feature vector is constructed for each parking space, which contains not only its current occupancy state, but also its static attributes, such as the specific location of the parking space in the parking lot, the degree of proximity to the entrance, and the historical usage frequency. The corrected predictive increase coefficient is data fused with the micro feature vector of each parking space to obtain multi-dimensional input data containing demand pressure and parking space attributes;

[0085] The multi-dimensional input data is transmitted to a preset prediction model, which independently calculates the input data of each parking space and outputs the probability value of the single parking space being occupied at a specified future time point, thereby realizing fine and differentiated prediction of the future occupancy state of each specific parking space.

[0086] Further, the influence score is obtained by the formula:

[0087]

[0088] wherein, is the reference initial influence score of event k, is the M-dimensional standardized feature vector of event k, v k,m is the standardized value of the mth feature, is the M-dimensional feature weight vector, ω m is the weight coefficient of the mth feature, and M is the total number of features.

[0089] Further, the influence sub-value is obtained by the formula:

[0090]

[0091] wherein, S k is the influence sub-value of event k at time t, is the reference initial influence score of event k, and α is the social heat amplification coefficient, R k(t) is the real-time social heat of event k at time t, R base is the heat normalization reference value, β is the aggregation intensity coupling coefficient, Ψ k (t) is the cumulative aggregation intensity of event k at time t, Θ gather is the aggregation effect determination threshold, Γ(Δt k ) is the time decay factor.

[0092] Those skilled in the art should understand that the above formulas are all dimensionless values, and the formulas are obtained by collecting a large amount of data to simulate the most recent real situation, and the preset parameters in the formulas are set by those skilled in the art according to the actual situation.

[0093] The above embodiments can be realized wholly or partially by software, hardware, firmware or any combination thereof. When realized by software, the above embodiments can be realized wholly or partially in the form of a computer program product.

[0094] Those skilled in the art can realize that the modules and algorithm steps of the examples described in combination with the embodiments disclosed herein can be realized by electronic hardware or a combination of computer software and electronic hardware. Whether the functions are realized by hardware or software depends on the specific application and design constraints of the technical solutions. Those skilled in the art can use different methods to realize the described functions for each specific application, but such implementation should not be considered beyond the scope of the present application.

[0095] In addition, the functional modules in each embodiment of the present application can be integrated in one processing module, or each module can exist physically alone, or two or more modules can be integrated in one module.

[0096] The above is merely a specific implementation of the present application, but the protection scope of the present application is not limited thereto. Any skilled person in the art can easily think of changes or replacements within the technical scope disclosed in the present application, which should be included in the protection scope of the present application. Therefore, the protection scope of the present application should be subject to the protection scope of the claims.

[0097] Finally, the above is only the preferred embodiment of the present application, and is not used to limit the present application. Any modification, equivalent replacement, improvement, etc. made within the spirit and principles of the present application should be included in the protection scope of the present application.

Claims

1. A real-time occupancy-aware prediction system for parking spaces, characterized in that, The method comprises the following steps: An event queue establishment module is used to obtain historical event information and real-time event information of events with crowd gathering effect occurring in a city, and to map the event information to a pre-divided city grid to establish an event queue associated with each grid; An influence sub-value acquisition module is used to obtain initial influence of the historical event information through multi-source heterogeneous data in the historical event information, and to generate a dynamic event influence sub-value for each city grid; An increase coefficient generation module is used to generate a predictive increase coefficient coupled with the real-time event information through a pre-set city grid occupancy coefficient increase model, and to derive a prediction result based on the predictive increase coefficient; A feedback correction module is used to monitor actual occupancy rates of each city grid in real time, to correct the city grid occupancy coefficient model and the dynamic event influence sub-value based on a residual error between the derived prediction result and the actual occupancy rate; A result generation module is used to fuse the corrected predictive increase coefficient with real-time state data of each parking space sensor in a parking lot, and to input the data into a microscopic prediction model to generate a real-time occupancy prediction result for a single parking space.

2. A real-time occupancy aware prediction system for parking spaces as claimed in claim 1 wherein, The step of mapping the event information to the pre-divided city grid to establish the event queue associated with each grid comprises the following steps: Geographic coordinate data in the city is obtained, and the city area is divided into a plurality of regular grid units according to a pre-set grid size, and each grid unit is assigned a unique grid identification code; Based on the received historical event information and real-time event information, geographic location coordinates of events are extracted, and a coordinate matching algorithm is used to determine the grid unit to which each event belongs; An independent event queue data structure is created for each grid unit, and events matched to the grid are stored in the corresponding queue in chronological order.

3. The real-time occupancy aware prediction system for parking spaces of claim 1, wherein, The step of obtaining the initial influence of the historical event information through multi-source heterogeneous data in the historical event information, and generating a dynamic event influence sub-value for each city grid comprises the following steps: Multi-source heterogeneous data of the historical event information is collected and integrated, and preprocessed to convert the multi-source heterogeneous data into a feature vector; Based on the converted feature vector, actual parking demand impact generated by each historical event is obtained, and a benchmark initial influence score of historical events of different types and different scales is obtained; When a new event appears in the real-time event queue, the features of the new event are matched with the features of historical events to obtain successfully matched historical event features; The initial influence score corresponding to the successfully matched historical event features is called, and a new event influence sub-value is generated by combining the real-time dynamic features of the new event, and the influence sub-value is assigned to each city grid affected by the event.

4. The real-time occupancy aware prediction system for parking spaces of claim 1, wherein, The pre-set city grid occupancy coefficient increase model expression is: In the formula, G i (t) is the predictive occupancy coefficient of the urban grid unit i at the future time t, B i (t) is the reference historical occupancy coefficient of the urban grid unit i at the time t, N is the total number of activity events at the time t, the influence of which can radiate to the grid unit i, S k is the dynamic event influence sub-value possessed by the kth event in the queue, W(d ik ,Δt k ) is a space-time weight decay function, d ik is the spatial distance between the center point of the event k and the center point of the grid unit i, Δt k is the time difference value between the time t and the starting time of the event k.

5. The real-time occupancy aware prediction system for parking spaces of claim 1, wherein, The step of deriving the prediction result based on the predictive increase coefficient comprises the following steps: For each city grid that needs to be predicted, the predictive increase coefficient value thereof is extracted; From a pre-stored city infrastructure database, a parking space list associated with the unique identifier of the city grid is queried and called to obtain the total number of all parking space resources in the city grid area, and a preliminary predicted occupied parking space number is obtained. based on the obtained preliminary predicted number of occupied parking spaces, converting the result into a predicted occupancy rate; storing the obtained predicted occupancy rate of the urban grid, together with the corresponding urban grid identifier and predicted time point information, as a data record.

6. The real-time occupancy aware prediction system for parking spaces of claim 1, wherein, The step of monitoring the actual occupancy rate of each urban grid in real time, combining the residual error between the derived prediction result and the actual occupancy rate, comprises: continuously receiving real-time state data from all parking space sensors in each urban grid of the city through the parking space sensor network; for each urban grid, aggregating the states of all parking spaces within it in real time to obtain the number of occupied parking spaces at the current time, and deriving the actual occupancy rate of the urban grid at the current time; based on the obtained predicted occupancy rate percentage, combining the actual occupancy rate, obtaining the residual error between the prediction result and the actual occupancy rate.

7. The real-time occupancy aware prediction system for parking spaces of claim 1, wherein, The step of generating real-time occupancy prediction results specific to a single parking space comprises: obtaining the predictive increase coefficient of the urban grid after feedback correction; obtaining the occupancy of each parking space at the current time through the parking space sensors in the parking lot in the urban grid; constructing a feature vector for each parking space, data fusing the corrected predictive increase coefficient with the micro feature vector of each parking space to obtain multi-dimensional input data containing demand pressure and parking space attributes; transmitting the multi-dimensional input data to a preset prediction model to output the probability value of the single parking space being occupied at a future specified time point.

8. A real-time occupancy aware prediction system for parking spaces as claimed in claim 3 wherein, The formula for obtaining the influence score is: wherein, is the baseline initial influence score for event k, is the M-dimensional normalized feature vector for event k, v k,m is the normalized value of the mth feature, is the M-dimensional feature weight vector, ω m is the weight coefficient of the mth feature, and M is the total number of features.

9. The real-time occupancy aware prediction system for parking spaces of claim 3, wherein, The formula for obtaining the influence sub-score is: In the formula, S k (t) is the influence sub-value of event k at time t, is the reference initial influence score of event k, and α is a social heat amplification coefficient, R k (t) is the real-time social heat of event k at time t, R base is a heat normalization reference value, β is a coupling coefficient of aggregation intensity, Ψ k (t) is the cumulative aggregation intensity of event k at time t, Θ gather is an aggregation effect determination threshold, Γ(Δt k ) is a time decay factor.

Images