CN111052199A

CN111052199A - Method for predicting parking area availability of street segments

Info

Publication number: CN111052199A
Application number: CN201880060251.5A
Authority: CN
Inventors: C.库恩哈; 陈若冰; P.迈尔
Original assignee: Robert Bosch GmbH
Current assignee: Robert Bosch GmbH
Priority date: 2017-07-18
Filing date: 2018-07-11
Publication date: 2020-04-21
Also published as: EP3655938A2; WO2019018168A2; WO2019018168A3

Abstract

A method for facilitating finding available parking areas for a street segment (120) comprises: receiving data (130) corresponding to parking areas located in a street segment (120), the data (130) comprising information ascertained by driving an ascertained vehicle through the street segment (120) and information received from a server; determining an instantaneous occupancy estimate (140) of the street segment (120) based on the received data (130); calculating a predicted occupancy estimate (160) using a timer sequence prediction model based on the instantaneous occupancy estimate (140); and generating a display representation of the calculated predicted occupancy estimate (160). The method comprises receiving the data (130) and determining the occupancy estimate (140), e.g. each time a vehicle is ascertained to be driving through a street.

Description

Method for predicting parking area availability of street segments

Cross Reference to Related Applications

The present application claims priority to U.S. patent application serial No.15/652,779 filed 2017, month 18, which is (a) continuation of U.S. patent application serial No.15/652,779 filed 2017, month 1, day 6, serial No.15/400,541, continuation of U.S. patent application serial No.15/400,541 filed 2015, month 9, day 11 and granted patent No.9,542,845 on 2017, month 1, day 10, and (b) continuation of U.S. patent application serial No.14/852,089 filed 2016, month 4, month 21, the contents of each of which are hereby incorporated by reference in their entirety.

Technical Field

The present invention relates generally to forecasting parking areas available for vehicles (vehicles), and more particularly to forecasting available parking areas for street segments based on historical occupancy estimates.

Background

Various methods are known in the relevant art to detect open parking areas for vehicles by means of distance-based sensors (e.g. ultrasonic sensors, laser sensors, radar sensors, stereo video cameras, etc.). Such processes are known, for example, from DE 102004062021 Al, DE 102009028024 a1 and DE 102008028550 Al.

Disclosure of Invention

Although the method for detecting open parking areas provides information of parking areas that are actually detected as being available at the present moment, the method does not provide a forecast of parking availability at a future time and does not provide information about availability without the present detection. That is, the methods discussed in the related art provide information about parking regions available at a particular time of detecting a parking region, but are not capable of predicting or predicting the availability of a parking region, for example, at a later point in time. Several disadvantages arise from the related method, for example as follows. First, if the driver uses the related method to decide where to park the driver's vehicle, the parking area may have become unavailable when the driver reaches the desired parking area. Second, by providing only available parking zones at a particular time of detecting a parking zone does not allow the driver to pre-plan the need to park the vehicle.

Example embodiments of the present application provide methods and systems for forecasting availability of parking areas for vehicles for street segments based on historical occupancy estimates.

According to an example embodiment of the present invention, a method for forecasting parking areas for streets includes: receiving data corresponding to a parking area located in a street segment, the data ascertained by driving an ascertaining vehicle through the street segment; determining, by a processing circuit, an instantaneous occupancy estimate for the street segment based on the received data; calculating, by the processing circuit, a predicted occupancy estimate based on the instantaneous occupancy estimate, the predicted occupancy estimate calculated using a time series prediction model; and displaying the calculated predicted occupancy estimate. In an example embodiment, the steps of receiving data and determining the instantaneous occupancy based on the received data are repeatedly performed each time at least one of an exploration vehicle and an additional exploration vehicle drives through the street segment.

In an example embodiment, the received data or otherwise obtained data includes: 1) a total number of unoccupied parking zones; 2) an estimated number of historically misdetected parking regions; and 3) the total number of parking areas located on the street segment.

In an example embodiment, the received data or otherwise obtained data includes: 1) the average length of the carrier; 2) a length of the determined unoccupied parking zone; 3) the estimated number of region lengths of historically misdetected parking regions; and 4) the total length of the street segment.

In an example embodiment, the received data or otherwise obtained data includes: 1) the length of the vehicle attempting to park; 2) a length of the determined unoccupied parking zone; 3) the estimated number of region lengths of historically misdetected parking regions; and 4) the total length of the street segment.

In an example embodiment, the predicted occupancy estimate is calculated using a seasonal autoregressive integrated moving average (SARIMA) model. In an example embodiment, the predicted occupancy estimate is visually displayed on a map using a color scale to visually render the occupancy level of the street segment.

In an example embodiment, the predicted occupancy estimate is modified based on external events that affect occupancy of the street segment. In an example embodiment, the confidence level of the predicted occupancy estimate is displayed.

An example embodiment of the invention relates to a server system for forecasting parking areas for street segments, the server comprising a database and a processing unit for forecasting parking areas for street segments, the processing unit performing the following: receiving data corresponding to a parking area located in a street segment, the data ascertained by driving an ascertaining vehicle through the street segment; determining an instantaneous occupancy estimate for the street segment based on the received data; calculating a predicted occupancy estimate based on the instantaneous occupancy estimate using a time series prediction model.

An example embodiment of the present invention is directed to a non-transitory computer readable medium having instructions stored thereon, the instructions being executable by a computer processor and when executed by the processor cause the processor to perform a method for forecasting parking areas for street segments, the method comprising: receiving data corresponding to parking areas located in a street segment, the data ascertained by an ascertaining vehicle driven along the street segment; determining, by a processor, an instantaneous occupancy estimate for the street segment based on the received data; calculating, by a processor and using a timer (timer) sequence prediction model, a predicted occupancy estimate based on the instantaneous occupancy estimate; and displaying the calculated predicted occupancy estimate.

These and other features, aspects, and advantages of the present invention are described in the following detailed description, which is to be read in connection with certain exemplary embodiments, and in view of the accompanying drawings, in which like characters represent like parts throughout the drawings. However, the detailed description and drawings merely describe and illustrate specific example embodiments of the invention, and are therefore not to be considered limiting of its scope, for the invention may admit to other equally effective embodiments.

Drawings

FIG. 1 is a flow chart depicting a method for forecasting parking areas for street segments in accordance with an exemplary embodiment of the present invention.

FIG. 2 is a representation of a function of occupancy determined for a street segment for a particular period of time according to an example embodiment of the present invention.

FIG. 3 is a representation of a function of predicted occupancy for a street segment for a particular period of time according to an example embodiment of the present invention.

4A-4E depict maps on which predicted occupancy for a plurality of street segments during a particular time period is displayed, according to an example embodiment of the present invention.

FIG. 5 is an illustration corresponding to a method of determining occupancy of a street segment in accordance with an example embodiment of the present invention.

FIG. 6 is an illustration corresponding to a method of determining occupancy of a street segment in accordance with an example embodiment of the present invention.

Detailed Description

Fig. 1 is a flow diagram of a method 100 for forecasting availability of parking areas for street segments based on historical occupancy estimates, according to an example embodiment.

At step 101, a street segment 120 is identified. The street segment 120 may be a street segment in which parking areas have been predefined, marked (i.e., delineated). The street segment 120 may alternatively be a street segment of a parking area that has not been predefined. At step 102, data 130 corresponding to a particular street segment is collected over a period of time. Data 130 is collected from various sensors located on vehicles traveling through street segment 120, and may include information relating to, among other things: the number of parking zones, e.g., predefined parking zones; the number of unoccupied parking zones; the number of occupied parking zones; any obstacles that may exist along the path of travel of the vehicle through the street segment 120; the length of the parking area; the length of the unoccupied parking zone; and the length of each detected obstacle. At step 103, an occupancy estimate 140 is calculated based on the collected data 130. In an example, the occupancy estimate 140 is determined based on a count occupancy estimate, a length occupancy estimate, or an automobile-based occupancy estimate, as described in detail below.

In an example embodiment,

steps

102 and 103 are performed in a loop such that after step 103 is completed, the method 100 may return to step 102 to collect data 130 for the street segment 120 at different points in time. The loop may continue in parallel with the execution of steps 105-108.

The data 130 obtained in loop 102 may be collected from one or more vehicles traveling down the same street segment. In this manner, data 130 is collected over a period of time to establish collection of data 130 over a particular period of time corresponding to a particular street segment. Further, each time data 130 is collected, a corresponding occupancy estimate 140 may be determined. Thus, the collection of both data 130 and corresponding occupancy estimates 140 may be determined for a particular street segment over a particular period of time. Based on this collected information, a relationship between occupancy of a particular street segment and a particular time period may be determined. Fig. 2 graphically illustrates one particular example of an occupancy estimate 140 determined over a particular time period, according to an example embodiment. The graph 200 includes a horizontal axis 201 corresponding to a particular time period. For example, the axis 201 shown in the figure corresponds to the time period starting from august of a particular year to october of the same year. The graph 200 further includes a vertical axis 202 corresponding to the occupancy estimate 140. For example, the axis 202 shown in the figure starts at 0.0, which corresponds to no occupancy, and ends at 1.0, which corresponds to a situation in which a street segment is fully occupied.

In an example embodiment, in a case where there are any gaps in the occupancy time series for a particular street segment, the determination of the occupancy estimate includes initially performing an estimate of missing data (imputation) to fill in the gaps in the occupancy estimate 140. Missing data may be the result of street segments not being accessed as frequently by vehicles as is needed for adequate data fill-in. For example, if the goal is to provide parking occupancy for a street on an hourly basis, data from at least one car driving through the street per hour would be required to provide the occupancy estimate. If there is an hour during which no cars visit the street, there is a missing point in the time series that would, for example, result in a gap in the graph shown in FIG. 2.

In some examples, the calculation of missing data is performed based on considering data at other times under the same street segment. In other examples, the calculation of missing data is performed based on considering data of other nearby streets at the same time.

For example, in an example embodiment where missing data is accounted for based on data at other times, a Bayesian Structured Time Series (BSTS) model is used to fill in the missing data. (see, e.g., in

"Bayesian structured time series" available at (C.). The method functions by: using moving windows to go forward and backward in time series, and fill in missing numbers with predictions from the BSTS modelAccording to the specification. For example, if there is 60 hours of data, but the 11 th hour is missing, the model may be trained over the first through tenth hours to forecast the occupancy for the eleventh hour, or over the twelfth through twenty-first hours to forecast the occupancy for the eleventh hour.

On the other hand, in an alternative example embodiment where data is set up based on data of neighboring streets, the following streets are used to fill in the missing data: with respect to the streets, the system includes information indicating them as being sufficiently close to the street for which the missing data exists such that there is an expected high correlation between subject streets and neighboring streets, the data of which are used to account for the missing data.

In an example embodiment, the missing data is accounted for by applying the Amelia process. (see, e.g., "AMELIA II: A Program for Missing Data" (2015) by Honaker et al, at

Is available). According to this example, the missing data is filled in with: "random missing" assumptions, and predictions of the time series of occupancy of streets with missing values using other streets via linear regression.

In an alternative example embodiment, missing data is accounted for by applying multi-variable accounting via chain equations (MICE), which is a bootstrap-based EM (expectation-maximization) algorithm that also assumes "random absence". (see, e.g., "mice: Multivariate Impurification by Chainated Equations in R" (2011) by Burren et al, supra

Is available).

In an alternative example embodiment, missing data is calculated using missfiest, a random forest based approach that does not require parameterization and has no assumptions about the functional form. (see, for example, Stekhoven, "Using the missForest Package" (2011) in

Is available).

Returning to FIG. 1, at step 105, the mode change detection 150 determines whether there are any anomalies present in a particular occupancy estimate 140. These anomalies may be due, for example, to external events that can affect occupancy estimation 140, as discussed below. At step 106, a predicted occupancy estimate 160 is calculated based on previously calculated historical occupancy estimates. For example, in the exemplary embodiment, predicted occupancy estimates 160 are calculated using various time series prediction algorithms, such as seasonal autoregressive integrated moving average (SARIMA) models and regression models. In this way, autocorrelation analysis is first performed to estimate trends and seasonality in historical occupancy estimates, which are then used to determine parameter values in a predictive algorithm. Next, the different model types and parameter settings are compared to determine the best model that provides the highest average accuracy across all forecast points.

Fig. 3 illustrates an example of a particular time series prediction model. In particular, fig. 3 illustrates a predicted occupancy estimate 160 generated using an autoregressive integrated moving average (ARIMA) model. Fig. 3 includes a graph 300, the graph 300 having a horizontal axis 301 corresponding to a particular time period and a vertical axis 302 corresponding to an occupancy, either an actual occupancy 303 or a predicted occupancy estimate 160. Further, the graph 300 shown in fig. 3 includes

confidence levels

305 and 306, which indicate different confidence levels associated with the outcome of a particular predictive model.

Returning to FIG. 1, at step 107, the predicted occupancy estimate 160 is displayed, for example, on a map. 4A-4E illustrate example embodiments of displays of predicted occupancy estimates 160 on various maps. For example, fig. 4A-4E illustrate

various street segments

401, 402, 403, and 404 and their corresponding predicted occupancy estimates 405, 406, 407, and 408, which are graphically illustrated as highlighted street segments. In the illustrated example, the predicted occupancy estimate 405 & 408 is superimposed on the street segment 401 & 404 and the occupancy level may be visually shown using a particular color scheme (although the highlighted street segments are shown in the figure as thick grey segments, they may instead be colors highlighted with an assigned color coding, with different segments highlighted with different colors). For example, a color scheme ranging from green to red may be used, where green indicates low occupancy, yellow indicates average occupancy, orange indicates above average occupancy, and red indicates high occupancy. For example, in FIG. 4A, the

street segments

401 and 404 may all have an average occupancy level, which may be illustrated by representing the predicted

occupancy estimate

405 and 408 in yellow (i.e., average occupancy level). Fig. 4A may, for example, indicate the occupancy of street segment 401-404 at midnight 12 AM on a particular day. The

street segments

401 and 404 as shown in FIG. 4B may have different occupancy estimates. For example,

sections

402 and 404 may be shown as being occupied more than

street sections

401 and 403; thus, the predicted

occupancy estimates

406 and 408 may be illustrated with orange, which indicates above average occupancy, and the predicted

occupancy estimates

405 and 407 may remain illustrated in yellow, which indicates average occupancy. Fig. 4B may, for example, indicate the occupancy of

street segment

401 and 404 at 6 AM on the same day as illustrated in fig. 4A. The

street segment

401 and 404 as shown in FIG. 4C may also have different occupancy estimates. For example, section 402-404 may be significantly more occupied than section 401; thus, the predicted occupancy estimate 406-408 may be illustrated with red, which indicates high occupancy, and the predicted occupancy estimate 405 may be illustrated with orange, which indicates above average occupancy. Fig. 4C may, for example, indicate occupancy of street segment 401-404 at 12 PM on the same day as illustrated in fig. 4A-4B. The

street segments

401 and 404 as shown in FIG. 4D may have the same occupancy estimate. For example,

section

401 and 404 may be significantly occupied; thus, the predicted occupancy estimation 405-408 may be illustrated with a red color, which indicates high occupancy. FIG. 4D may, for example, indicate occupancy of street segment 401-404 at 6 PM at noon on the same day as illustrated in FIGS. 4A-4C. The

street segment

401 and 404 as shown in FIG. 4E may also have different occupancy estimates. For example, zone 401-403 may be less occupied than zone 404; thus, the predicted occupancy estimate 405-407 may be illustrated with yellow indicating average occupancy, and the predicted occupancy estimate 408 may be illustrated with orange indicating above average occupancy. Fig. 4E may, for example, indicate occupancy of street segment 401-404 at midnight 12 AM on a day following the day illustrated in fig. 4A-4D.

Returning to FIG. 1, in one particular embodiment, at step 108, a confidence level 170 is also displayed, for example, by displaying a numerical value corresponding to the confidence level of the time series predictive model used to determine the occupancy prediction. The confidence level 170 corresponds to an assessment of the accuracy of the predicted occupancy estimate 160 calculated by various time series prediction algorithms, for example, as shown in fig. 3 by

confidence levels

305 and 306.

In one particular embodiment, the determination of occupancy estimates for street segments is calculated for street segments having defined parking regions, i.e., having predefined, marked (i.e., delineated) parking regions, such that a particular street segment has a corresponding integer corresponding to the total number of parking regions for that particular street segment. In this embodiment, the occupancy estimate may be determined based on: 1) a total number of detected unoccupied parking zones, 2) an estimated number of historically misdetected parking zones, and 3) a total number of detected parking zones. For example, FIG. 5 is a diagram depicting a street segment 500, the street segment 500 having a beginning segment 520 and an ending segment 530, and including defined

parking regions

501, 502, 503, 504, 505, 506, 507, and 508. The street segment 500 further includes a driveway 509 that is obstructing the parking area 504 (i.e., the vehicle cannot legally or physically park in the parking area 504). As the vehicle 510 drives down the street segment 500 in the direction 511, the vehicle 510 detects the presence of: occupied

parking zones

501, 502, 505, 507, and 508, respectively, are parked in parked

vehicles

512, 513, 514, 515, and 516 in the defined

parking zones

501, 502, 505, 507, and 508. The vehicle 510 also detects

unoccupied parking zones

503, 504, and 506.

As shown in fig. 5, 504 is a falsely detected parking area and corresponds to a blocked parking area, such as a driveway, a fire hydrant, a no parking zone, and so forth. To determine 504 that a parking area is a false detection, parking information for a particular street segment may be obtained over a period of time by a vehicle traveling through the street segment. In this way, the total number of parking areas and the total number of parked vehicles are obtained each time a vehicle travels through a particular street segment. If over time multiple vehicles detect a total number of parking zones equal to 10, then it is assumed that the street segment has a total of 10 parking zones. However, if no vehicle detects more than 9 parked vehicles over a predefined period of time, it may be assumed that one parking area of the particular street segment is a blocked parking area, i.e. a false detection. Thus, the particular street segment is identified as having a falsely detected parking area.

In an example, based on detected parking regions and falsely detected parking regions, a count occupancy estimate for the street segment 200 is calculated as follows:

occupancy estimation (counting)=

In which N is_detRepresents the total number of detected unoccupied parking zones, e.g.,

unoccupied parking zones

503, 504, and 506, as shown in FIG. 5; n is a radical of_falseRepresents an estimated number of historically misdetected parking regions, such as parking regions 504 obstructed by a driver's way 509 as shown in FIG. 5; and N is_totalRepresenting the total number of parking regions on a particular section of street, such as 501, 502, 503, 504, 505, 506, 507, and 508. Thus, the counted occupancy estimate for the street segment 500 as shown in FIG. 5 is 72% occupied.

In one particular embodiment, the determination of occupancy estimates is for street segments that do not have defined parking regions (i.e., parking regions that are not marked and/or not delineated). (Note that in an example embodiment, the system is configured to perform the determination for both types of street segments). In this embodiment, a length occupancy estimate may be used. The length occupancy estimate may be calculated based on: 1) an average length of the vehicle, 2) a length of the determined unoccupied parking zone, 3) a zone length of the estimated number of historically misdetected parking zones, and 4) a total length of the street segment. In this way, unoccupied parking zones of insufficient length for parking are excluded from the occupancy calculation based on the average length of the vehicle. For example, if the average length of the vehicle is predefined to be 15 feet, then an unoccupied zone having a length of 10 feet is disregarded and is not considered an unoccupied parking zone. In this way, it is ensured that each detected unoccupied parking zone has a length that is sufficiently large enough for a particular vehicle to be able to maneuver and park in the unoccupied parking zone. To achieve this result, minimum and maximum length thresholds may be used in determining whether the detected parking area is large enough for vehicle handling and parking. For example, FIG. 6 depicts a street segment 600, the street segment 600 having a beginning segment 620 and an ending segment 630, and including

parking regions

601, 602, 603, 604, 605, 606, 607, and 608 having respective lengths. The street segment 600 further includes a driveway 609 that is obstructing the parking area 604. As the vehicle 610 drives down the street segment 600 in the direction 611, the vehicle 610 detects the length of occupied

parking zones

601, 602, 605, 607, and 608, and the length of

unoccupied parking zones

603, 604, and 606. Further, the vehicle 610 detects the presence of parked

vehicles

612, 613, 614, 615, and 616 that are parked in the

parking areas

601, 602, 605, 607, and 608, respectively. In this example, the length of the parking area 606 is less than the selected average length of the vehicles, and thus, the parking area 606 and its length are ignored and not used to calculate the occupancy of the street segment.

Based on the foregoing, in an example embodiment, a needleThe length occupancy estimate for the street segment 600 is calculated as:account for By estimation (length)= 1–

Wherein

Represents the total length of detected unoccupied parking zones for a vehicle over a particular segment, excluding any length of unoccupied parking zones that is shorter than the average vehicle length, such as the sum of the lengths of

unoccupied parking zones

603 and 604 as shown in fig. 6;

a total length of area representing an estimated number of historically falsely detected parking areas for vehicles on a particular section of street, such as length 604 obstructed by a driver's lane 609 as shown in fig. 6; and L is_{length_total_avg}Is the total length of the street segment 600.

In an alternative example embodiment, the determination of occupancy estimates for street segments without defined parking zones is performed in an alternative manner similar to length occupancy estimates, but instead of using the average length of the vehicles, the actual length of the car attempting to park is used. Thus, the car-based occupancy estimate is calculated based on: 1) a length of the vehicle attempting to park, 2) a length of the determined unoccupied parking zone, 3) a zone length of the estimated number of historically misdetected parking zones, and 4) a total length of the street segment. In this way, based on the length of the actual car attempting to park, a parking area that is too small and unoccupied is identified and not considered for calculating the occupancy of the street segment. For example, if the length of the car attempting to park is 10 feet, an unoccupied parking zone having a length of 8 feet, for example, is ignored and is not considered as an unoccupied parking zone, but an unoccupied parking zone having a length of 11 feet is considered as an unoccupied parking zone. Vehicle-based occupancy estimation is computed, for exampleComprises the following steps:occupancy estimation (based on cars)= 1 –

Wherein

Represents the total length of detected unoccupied parking zones, excluding any length of unoccupied parking zones determined to have a parking length insufficient for a particular car;

a total length of area representing an estimated number of historically misdetected parking areas for vehicles over a particular section of street; and L is_{length_total_car}Representing the total length of the street segment.

In this way, an occupancy estimate based on the car is calculated, which is a more tailored occupancy estimate, since the unoccupied parking zones are selected to correspond to a specific length of the particular vehicle attempting to park.

Based on the foregoing, whenever a vehicle (which includes the requisite sensing, computing and communication device (s)) drives through a particular street segment, a corresponding occupancy estimate may be calculated. Thus, over time, each street segment may be associated with a collection of stored occupancy estimates. Based on the collected occupancy estimates, various time series prediction models may be used to calculate the predicted occupancy estimate, as discussed above.

In an example embodiment, when a predicted occupancy estimate is calculated for a particular street segment within a particular time period, the mode change detection 150 may determine whether there are any anomalies affecting a particular occupancy estimate 140. In this way, the predicted occupancy estimate may be examined to determine whether any anomalies (i.e., special or external events) exist during a particular time period of the predicted occupancy estimate for that particular street segment. For example, the external data may be analyzed to determine whether a particular time period during which the predicted occupancy estimate is calculated coincides with, for example, a public holiday, a public event, or some other event that would affect parking availability in a particular street segment during the particular time period. In this way, anomalies can negatively impact the ability of the time series prediction model to generate accurate predicted occupancy estimates. Therefore, it is advantageous to consider any of these potential events that coincide with the predicted occupancy estimate, so that the effects of external events can be accounted for and an improved occupancy estimate can be calculated.

Furthermore, it is advantageous for the mode change detection 150 to accurately predict the magnitude of the impact of an abnormal event on parking availability. The magnitude of the effect may be calculated based on newly collected data from vehicles traveling down a particular street segment during a particular external event combined in a bayesian framework combined with a data time period in which a similar external event occurred.

In one particular embodiment, the mode change detection 150 may determine whether any unforeseen external events are affecting parking occupancy when calculating a predicted occupancy estimate for a particular street segment within a particular time period. For example, a particular street segment may be undergoing repair or construction that prevents vehicles from parking in certain parking areas that would otherwise be available for parking. In this way, it is advantageous to accurately detect whether a particular street segment is experiencing any unforeseen external event, such as road construction, from collected data corresponding to the particular street segment, and to determine the magnitude of the impact of such an event on the predicted occupancy estimate. Non-parametric multi-change point analysis methods may be used to determine the presence of unforeseen external events and their corresponding effects. Furthermore, parameters of the non-parametric multi-change point algorithm, such as the minimum number of observations between change points, may be adjusted so that multiple change points may be detected without assuming any underlying distribution. When a change is detected, the mode change detection 150 may perform an analysis of the cause being performed, and if an unforeseen external event is determined to be a repetitive event, the presence and its corresponding impact on parking availability may be characterized as a special event, which increases the accuracy of the predicted occupancy estimate.

Example embodiments of the present invention are directed to processing circuitry, e.g., comprising one or more processors, which may be implemented using any conventional processing circuitry and devices, e.g., a Central Processing Unit (CPU) of a Personal Computer (PC) or other workstation processor, or a combination thereof, to execute code, e.g., provided on a non-transitory computer-readable medium comprising any conventional memory device, to perform any of the methods described herein, alone or in combination. The one or more processors may be embodied in a server or a user terminal, or a combination thereof. The user terminal may be embodied as, for example, a desktop computer, a laptop computer, a handheld device, a Personal Digital Assistant (PDA), a television set-top internet appliance, a mobile phone, a smart phone, etc., or as a combination of one or more of these. The memory device may include any conventional permanent and/or temporary memory circuit or combination thereof, a non-exhaustive list of which includes Random Access Memory (RAM), Read Only Memory (ROM), Compact Disc (CD), Digital Versatile Disc (DVD), and magnetic tape.

Example embodiments of the present invention are directed to a plurality of ascertained vehicles that perform detection along a street segment regarding a current parking area status. The plurality of ascertained vehicles may transmit the detected parking area status to a server. The server accumulates the detected parking zone states to create a predicted occupancy estimate based on the detected parking zone states. The server may transmit the predicted occupancy estimates to the plurality of vehicles, to a user terminal, such as a desktop computer, laptop computer, handheld device, Personal Digital Assistant (PDA), television set-top internet appliance, mobile phone, smart phone, etc., or to an additional server. The ascertained vehicle, user terminal or server may then use a display device to display the predicted occupancy estimate.

The predicted occupancy estimate does not necessarily mean predicted for the future, but the predicted occupancy estimate may also be an estimate of the current parking state along a street segment for which there is currently no actual information sensed, the predicted occupancy estimate being determined from historical information as described above. The predicted occupancy estimate may be sent to vehicles, including ascertained vehicles (i.e., vehicles that send information to a server regarding the current parking area status along the street segment), and also vehicles that have not and/or have not sent such information.

Example embodiments of the present invention are directed to one or more non-transitory computer-readable media, for example, as described above, on which are stored instructions executable by a processor and each method when executed by the processor performs the various methods described herein, individually or in combination, or sub-steps thereof, in isolation or in other combinations.

Example embodiments of the present invention are directed to methods, such as hardware components or machines transmitting instructions that are executable by a processor to perform the various methods described herein, either individually or in combination with each other, or in isolation or in other combinations of sub-steps thereof.

The above description is intended to be illustrative, and not restrictive. Those skilled in the art will appreciate from the foregoing description that: the present invention can be implemented in various forms, and various embodiments can be implemented alone or in combination. Therefore, while the embodiments of this invention have been described in connection with particular examples thereof, the true scope of the embodiments and/or methods of the invention should not be so limited since other modifications will become apparent to the skilled practitioner upon a study of the drawings, the specification and the following claims.

Claims

1. A method for assisting in finding available parking areas for street segments, the method comprising:

receiving, by a processing circuit, data corresponding to a parking area located in a street segment, the data ascertained by driving at least one ascertaining vehicle through the street segment;

determining, by the processing circuit, an instantaneous occupancy estimate for the street segment based on the received data;

calculating, by the processing circuit, a predicted occupancy estimate based on the instantaneous occupancy estimate, wherein the predicted occupancy estimate is calculated using a time series prediction model;

generating, by the processing circuit, a display representation of the calculated predicted occupancy estimate; and

the display representation is provided for display on a display device.

2. The method of claim 1, wherein said receiving data and determining an instantaneous occupancy estimate are performed each time at least one of said at least one ascertained vehicles drives through said street segment.

3. The method of claim 1, wherein the received data comprises: 1) a total number of unoccupied parking zones; 2) an estimated number of parking regions that have historically been misestimated; and 3) the total number of parking areas located on the street segment.

4. The method of claim 3, wherein:

the instantaneous occupancy estimation is by 1-

To be determined;

N_detrepresents the total number of unoccupied parking zones;

N_falserepresenting the number of estimated parking regions that have historically been misestimated; and

N_totalrepresenting the total number of parking areas located on the street segment.

5. The method of claim 1, wherein the received data comprises: 1) the average length of the carrier; 2) a length of the determined unoccupied parking zone; 3) estimated length of a parking region that has historically been misestimated; and 4) the total length of the street segment.

6. The method of claim 5, wherein:

the instantaneous occupancy estimation is by 1-

To be determined;

represents the total length of the detected unoccupied parking zones;

representing an estimated total length of parking area that has historically been misestimated; and

L_{length_total}representing the total length of the street segment.

7. The method of claim 1, wherein the received data comprises: 1) the length of the vehicle attempting to park; 2) a length of the determined unoccupied parking zone; 3) estimated length of a parking region that has historically been misestimated; and 4) the total length of the street segment.

8. The method of claim 7, wherein:

the instantaneous occupancy estimation is by 1-

To be determined;

indicating a detected failureThe total length of the occupied parking area;

L_{length_total}representing the total length of the street segment.

9. The method of claim 2, wherein the predicted occupancy estimate is calculated using a seasonal autoregressive integrated moving average (SARIMA) model.

10. The method of claim 2, wherein the display representation comprises a visual representation of the predicted occupancy estimate on a map that visually represents an occupancy level of a street segment using a color scale.

11. The method of claim 2, wherein the predicted occupancy estimate is modified based on external events affecting occupancy of the street segment.

12. The method of claim 2, wherein the displaying the representation comprises displaying a confidence level of the predicted occupancy estimate.

13. A server system for assisting in finding available parking areas for street segments, the server comprising:

a database; and

a processing unit, wherein the processing unit is configured to perform the following:

receiving data corresponding to a parking area located in a street segment, the data ascertained by driving at least one ascertaining vehicle through the street segment;

determining an instantaneous occupancy estimate for the street segment based on the received data;

calculating a predicted occupancy estimate based on the instantaneous occupancy estimate, wherein the predicted occupancy estimate is calculated using a time series prediction model;

generating a display representation of the calculated predicted occupancy estimate; and

the display representation is provided for display on a display device.

14. The server system of claim 13, wherein the receiving of data and the determining of an instantaneous occupancy estimate are performed repeatedly each time a vehicle is driven through the street segment.

15. The method of claim 13, wherein the received data comprises: 1) a total number of unoccupied parking zones; 2) an estimated number of parking regions that have historically been misestimated; and 3) the total number of parking areas located on the street segment.

16. The method of claim 13, wherein the received data comprises: 1) the average length of the carrier; 2) a length of the determined unoccupied parking zone; 3) estimated length of a parking region that has historically been misestimated; and 4) the total length of the street segment.

17. The method of claim 1, wherein the received data comprises: 1) the length of the vehicle attempting to park; 2) a length of the determined unoccupied parking zone; 3) estimated length of a parking region that has historically been misestimated; and 4) the total length of the street segment.

18. The method of claim 2, wherein the predicted occupancy estimate is calculated using a seasonal auto-regressive moving average (SARIMA) model.

19. The method of claim 2, wherein the display representation comprises a visual representation of the predicted occupancy estimate on a map that visually represents an occupancy level of a street segment using a color scale.

20. A non-transitory computer readable medium having stored thereon instructions executable by a computer processor and which when executed by the processor cause the processor to perform a method for facilitating finding available parking areas for street segments, the method comprising:

the display representation is provided for display on a display device.