Procedure for providing parking information on available parking spaces

Abstract

Procedure for providing parking information on available parking spaces, in which a) historical and current information on available, free parking spaces is determined for at least one street section, whereby at least one parking segment comprising one or more streets is determined from the information obtained, and statistical parameters on available parking spaces are generated for each parking segment from the information obtained; b) a model is created for each parking segment, in which the information determined for the parking segment is processed to represent a parking state of the respective parking segment as a probability distribution (P i ) to determine, characterized in that initial information about a request rate (λ) is determined, which indicates the number of requests for a parking space per time for a parking segment.

Description

[0001] The invention relates to a method for providing parking information on available parking spaces.

[0002] Parking information on available parking spaces is used, for example, by parking guidance systems and / or navigation devices to guide vehicles searching for parking. Modern urban systems operate on a simple principle: if the number of parking spaces and the flow of vehicles entering and exiting the area are known, the availability of free parking spaces can be easily determined. Appropriate signage on the access roads and dynamic updates of the parking information allow vehicles to be navigated to available spaces. Due to the inherent limitations of this principle, parking areas must be clearly defined, and vehicle entry and exit must be closely monitored at all times. This requires structural measures such as barriers or other access control systems.

[0003] Due to this limitation, navigation to only a small number of available parking spaces is possible. Typically, only parking garages or fenced parking areas can be integrated into a parking guidance system with the necessary structural modifications. The far greater number of on-street parking spaces or unfenced parking lots are disregarded, as the parking situation in public spaces is largely unknown. Only a few municipalities or traffic management centers provide information for specific areas.

[0004] To locate available parking spaces, especially in city centers and densely populated areas, it is desirable to identify parking spaces along specific streets. German patent DE 10 2009 028 024 A1 discloses the use of vehicles capable of detecting parking spaces, such as public transport vehicles like regularly scheduled buses or taxis, which are equipped with at least one sensor for parking space detection. The sensor technology can be optical and / or non-optical.

[0005] Furthermore, community-based applications are known where vehicle users, for example, enter information into an app when they leave a parking space. This information is then made available to other users of the service. The disadvantage here is that the information about available parking spaces is only as good as the information provided by the users.

[0006] Both of the described alternatives have the problem that information about the availability of a single parking space is very fast-moving, i.e., in areas with a lot of traffic searching for parking, where parking information would be helpful, a free parking space is usually occupied in a very short time.

[0007] In the subsequently published German patent application DE 10 2012 201 472 A1, the applicant further describes a method for providing parking information on available parking spaces, in which a knowledge database containing historical data is generated from information obtained about available, free parking spaces. The historical data comprises statistical data on available parking spaces for specified streets and / or specified times or periods. From the historical data and current information obtained at a given time for one or more selected streets from vehicles in traffic, a probability distribution of expected available parking spaces for the selected street(s) is determined. This probability distribution represents parking information on available parking spaces in the selected street(s).

[0008] A disadvantage of this method is that current information on parking lot occupancy must be measured to make an estimate and forecast using historical parameters. If no current estimate is available, a statement is calculated based solely on historical data. Similarly, if only samples or individual parking events are available, no statement can be made. This makes the method insufficiently accurate.

[0009] US Patent 2012 / 0161984A1 discloses a device for assisting in the search for on-street parking spaces. The device receives information about a geographic location for which parking assistance is desired and specifies an effective time for which assistance should be offered. The device then provides parking assistance for the specified location at the time of use. The device generates a route based on determining the probability that for each of the on-street parking areas, referred to as "parking segments" or "segments," at least one suitable on-street parking space is available in the segment.

[0010] The object of the present invention is to provide, based on the applicant's method, an improved method for providing parking information on available parking spaces.

[0011] This problem is solved by a method according to the features of claim 1, a computer program product according to the features of claim 12, and a system for providing parking information according to the features of claim 13. Advantageous embodiments are specified in the dependent claims.

[0012] The invention provides a method for providing parking information on available parking spaces. In particular, it provides a method for taking into account available parking spaces along streets, i.e., areas not delimited by barriers.

[0013] The process involves gathering historical and current information on available, vacant parking spaces for at least one street segment. From this information, at least one parking segment, encompassing one or more streets, is identified. For each parking segment, statistical parameters regarding available parking spaces are generated from the gathered information. This initial step can be performed using historical data.

[0014] Historical data can be stored in a knowledge base, where the historical data for predefined streets and / or predefined times or time periods includes statistical data on available parking spaces. For example, the knowledge base might record that, on a given street, at a specific time or time period, an average of y parking spaces are available out of a total of x available spaces. At a different time or time period, however, only z < y parking spaces are available on the same street. Thus, the historical knowledge base contains information on which parking spaces can, in principle, be used as parking spaces (so-called valid parking spaces or gaps), and also information on the average number of available parking spaces at specific times.The historical data therefore includes occupancy data and can also contain information about parking maneuvers (entering and exiting) or traffic searching for parking.

[0015] Furthermore, a model is generated for each parking segment, processing the information gathered for that segment to determine the parking status of that segment as a probability distribution. The determination of the probability distribution of expected free parking spaces is preferably performed by a central computer. The current information on available, free parking spaces is thus transmitted to the central computer by vehicles in traffic or stationary sensors in the relevant street sections of the parking segment.

[0016] The parking situation on one street affects adjacent streets. Using functions that describe an entire area, a so-called parking segment, takes this into account. Furthermore, it has the advantage that an assessment can be made even for streets where no information is currently available.

[0017] According to the invention, the first piece of information determined is a request rate (λ), which indicates the number of requests for a parking space per unit of time for a parking segment. This information thus represents the number of vehicles that wish to park in a parking segment within a specific unit of time and therefore send a request to the central computer. The request rate is also referred to as the arrival rate.

[0018] In one implementation of the method, an updated probability distribution of the parking status is determined from real-time information about free, available parking spaces in a parking segment. An "online" estimation of the parameters allows for adaptive learning in response to changes in parking habits.

[0019] In a further embodiment of the method, information about available, free parking spaces is measured and recorded from vehicles already in traffic. This can be done using sensors already present in vehicles, which may be based on optical and / or non-optical sensors. The use of a camera is particularly preferred. Side-facing cameras in vehicles, such as those designed to assist with parking maneuvers by detecting obstacles in vehicles, are especially suitable. Sensors originally intended for lane departure or lane change assist systems can also be used. Such sensors may be based on radar or other non-optical technologies.

[0020] Similarly, a roadside area can be captured by a vehicle's camera, generating a sequence of images which are then analyzed by the vehicle's computer to identify available parking spaces. It is advisable that only valid parking spaces are included in the probability calculation. A valid parking space is defined as one where a vehicle is legally permitted to park. Invalid parking spaces include, for example, intersections, fire lanes, and similar areas. Plausibility is verified using image processing and additional sensors, such as a digital map, whereby available parking spaces are automatically detected and validated while the vehicle is driving. For example, a camera mounted on the side of the vehicle can be used for this purpose.

[0021] In a further embodiment of the method, information about available, free parking spaces is measured by sensors arranged along the streets. Such sensors are known, for example, for monitoring parking spaces in parking garages or other demarcated parking areas.

[0022] It may also be possible for information about available, free parking spaces to be generated manually by users entering data into a device (e.g., smartphone, laptop, tablet, etc., or even a vehicle's user interface). For example, dedicated apps could be provided for this purpose, allowing users to report available parking spaces. A corresponding user entry could be made, for instance, when a user pulls their vehicle out of a parking space. This information would then be considered by the aforementioned parking computer as part of its processing of current information.

[0023] In a further development of the method, information about available, free parking spaces is automatically generated by a vehicle by transmitting corresponding information with each parking maneuver and each exit maneuver. Many vehicles today are equipped with systems that report to a central server when they are parked. These include vehicles from logistics or fleet companies, as well as car-sharing vehicles. The proposed method uses this information to make up-to-date statements when systems capable of measuring street occupancy do not yet have the necessary level of implementation.

[0024] Information about available, free parking spaces is expediently transmitted to a central computer that generates and / or manages the knowledge base. Such a central computer can be administered, for example, by a service provider that supplies parking information. This service provider could also be a vehicle manufacturer, for instance, who uses the information about available parking spaces to process it as part of their route navigation system.

[0025] In a further refinement, secondary information about vehicles parking in and / or exiting a parking space is collected, whereby an exit rate is calculated from the dwell times between parking and exiting a given vehicle. This exit rate can be processed, for example, in a queuing model, allowing for a prediction of changes in probability at a later time. Such a later time could be, for instance, the arrival time at a specific street within the context of a calculated route navigation. A prediction can, in principle, also be made based on the historical probability distribution. However, the more current the data, the better the prediction quality. Here, a probability is calculated for the current time based on historical and current information (i.e.,Vehicle is parked out) hit.

[0026] In a further iteration, third-party information is obtained regarding a fixed number of detected free parking spaces. This number can be determined using the aforementioned methods. A number of "0," for example, indicates that no parking space is available.

[0027] In a further refinement, each parking segment is assigned at least one basic function, which represents the parking exit rate and the request rate depending on a parking characteristic. Basic functions include, for example, classification as a residential area, shopping area, business district, or entertainment district. Each of the basic functions exhibits characteristic features regarding parking behavior, which can be taken into account in the model by corresponding values ​​for the request rate and the parking exit rate. For example, in an entertainment district, increased traffic searching for parking is observed during the night. During the day, however, this traffic may be negligible. Conversely, the situation is reversed in a parking segment classified as a shopping district or business district. In a residential area, there is relatively little traffic searching for parking during the day. A parking segment can only be assigned one basic function.Similarly, a parking segment can be assigned multiple basic functions. For example, a parking segment might simultaneously be a residential area and an entertainment district. The assignment of one or more basic functions can be done manually or automatically using map search functions. For this purpose, information about shops, businesses, bars, and restaurants contained in Google Maps® can be analyzed.

[0028] The modeling for each park segment is carried out according to a configuration of the procedure using a birth-death process (so-called BD process).

[0029] The invention further provides a computer program product that can be directly loaded into the internal memory of a digital computer or computer system and comprises software code sections with which the steps according to one of the preceding claims are performed when the product is running on the computer or computer system.

[0030] Ultimately, the invention creates a system for providing parking information on available parking spaces in at least one street. The system includes a) a first unit for determining information on available, free parking spaces for at least one street section, which is designed to determine at least one parking segment comprising one or more streets from the determined information and to generate statistical parameters on available parking spaces for each parking segment from the determined information; b) a second unit for generating a model for each parking segment, which is designed to process the information obtained for the parking segment in order to determine a parking state of the respective parking segment as a probability distribution.

[0031] According to the invention, initial information about a request rate is determined, which indicates the number of requests for a parking space per time for a parking segment.

[0032] The system has the same advantages as those explained above in connection with the method according to the invention.

[0033] Furthermore, the system may include additional means for implementing preferred configurations of the procedure.

[0034] The invention is explained in more detail below with reference to an exemplary embodiment shown in the drawing. The drawing shows: Fig. 1 a schematic representation of a system for carrying out the method according to the invention; Fig. 2 a schematic representation of the model underlying the method according to the invention; Fig. 3 a schematic representation of the system initialization Fig. 1; Fig. 4 a schematic representation of the system update from Fig. 1; Fig. 5 a schematic representation of the operation of the system Fig. 1; and Fig. 6 a probability distribution of expected free parking spaces for a parking segment and their change depending on an observed and reported exiting maneuver to the system.

[0035] Fig. Figure 1 shows a schematic representation of a system according to the invention for providing parking information on available parking spaces in one or more streets. The system comprises a central computer 10, which can be composed of one or more computers. The central computer 10 is managed, for example, by a service provider for the provision of parking information. The service provider could, for example, be a vehicle manufacturer. The task of the central computer 10 is to process information about available, free parking spaces, which is transmitted to the central computer, for example, by vehicles in traffic, but also by stationary sensor units, and to determine a probability distribution of expected free parking spaces. As will become apparent from the following description, the determination of the probability distribution is not carried out for a single specific street, but for a so-calledParking segment that covers several streets.

[0036] The central computer 10 includes a user interface 11, for example, in the form of a web server frontend, through which a user searching for an available parking space can submit a request to the central computer 10 and receive a corresponding response (communication path K1). When the description refers to a user's request or a user's response, this naturally means that communication takes place between the user's terminal device and the central computer 10. The terminal device can be a component of the user's vehicle or computer, or a portable device (e.g., a smartphone or tablet PC). The user interface 11 is connected to a processing unit 12 via a communication path K2. The processing unit 12 is used to process incoming and outgoing data and to prepare requested parking information.Furthermore, a data provider 14 is connected to the central computer 10 or its computing unit 12 via an interface not shown in detail, through which information about available, free parking spaces is transmitted to the central computer (communication path K4).

[0037] Furthermore, the central computer 10 includes a database 13, which can be accessed by the computing unit 12 (communication path K3). In addition to historical data, database 13 stores information about the current status of available parking spaces.

[0038] The model underlying the inventive method for providing parking information on available parking spaces is in Fig. 2 is shown in more detail. The model has a two-stage structure and can be implemented in the computing unit 12.

[0039] A so-called macromodel 20 calculates a map of parking segments from historical data on parking maneuvers, search traffic, and other parking information. For this purpose, contiguous or interconnected streets with similar characteristics regarding parking maneuvers, search traffic, and other parking information are grouped into a parking segment. A parking segment thus typically comprises several streets, although the number of streets can vary from segment to segment. A map composed of streets (i.e., a city map) is subdivided somewhat more coarsely using this approach. Instead of streets, the map or city map is divided into (parking) segments. In a subsequent step, the macromodel 20 determines the statistical parameters queuing rate λ(t) and average parking duration 1 / µ(t), also known as the exit rate, for each parking segment.

[0040] The request rate λ(t) represents the number of (user) requests per unit of time for an available parking space in the relevant parking segment. The average parking duration 1 / µ(t) represents the time elapsed between a parking entry and an exit.

[0041] Each parking segment can be assigned at least one so-called basic function, which represents the parking exit rate 1 / µ(t) and the request rate λ(t) as a function of a parking characteristic. Examples of basic functions are classifications such as residential, shopping, business, or entertainment districts. Each basic function exhibits characteristic features regarding parking search traffic, which can be taken into account in the model by corresponding values ​​for the request rate and the parking exit rate. For example, increased parking search traffic is observed in an entertainment district during the night. During the day, however, parking search traffic may be negligible. Conversely, the situation is different in a parking segment classified as a shopping or business district. In a residential area, there is relatively little parking search traffic during the day. Only one basic function can be assigned to a parking segment.Similarly, a parking segment can be assigned multiple basic functions. For example, a parking segment might simultaneously be a residential area and an entertainment district. The assignment of one or more basic functions can be done manually or automatically using map search functions. For this purpose, information about shops, businesses, bars, and restaurants contained in Google Maps® can be analyzed.

[0042] The micromodel 21 (µmodel) models each parking segment using a birth-death process (BD process 22) known to those skilled in the art and therefore not described in detail, and uses the statistical parameters obtained from the macromodel to estimate the new parking state in the form of a probability vector P.

[0043] If real-time data can be applied to a specific parking segment, the probabilities are further improved by means of a unit 23 for updating the micromodel (µmodel updater).

[0044] The result of the modeling is a probability distribution of available parking spaces for each parking segment, as shown, for example, in the diagrams of the Fig. Figure 6 is shown. In this figure, the probability vector P(t) is plotted against the number m of all theoretically available parking spaces, where m can take values ​​between 0 and 1. At time t1 (left diagram), the probability that four parking spaces are occupied is greatest in this example. The sum of all probabilities P i (i=0 to m) is 1.

[0045] Modeling familiar parking maneuvers can be done as follows. When parking out, the elements of the vector P are... i(t2) is shifted one position to the left at time t2 and renormalized to "1". This shift in the probability distribution of available parking spaces in a parking segment is exemplified in Fig. Figure 6 is shown. When parking, the elements of the vector P are used. i (t2) is shifted one unit to the right at time t2 in an analogous manner and renormalized to 1. Direct observations, such as "5 free parking spaces were discovered in this segment", automatically lead to a defined probability vector.

[0046] The developed macromodel enables improved prediction of the probability of available parking spaces by adaptively estimating the parameters of the request rate λ and the exit rate µ based on new measurements (input in block 21 of the micromodel). Over time, the parameters can thus be learned and adjusted to the characteristics of the parking segment based on specific criteria such as day of the week, time of day, etc., for example, when new segments are introduced or parking conditions change.

[0047] The three events that serve as input variables into the µ model (Block 21) are: parking rate λ p , Parking rate µ p and no parking space was found on a street. These variables are related to the parameters request rate λ and parking exit rate µ as follows: λp(t)=(1−Pn)λ(t) μp=μ∑iPi

[0048] Ultimately, the unknown parameters request rate λ and parking rate µ can be estimated by solving an optimization problem such as a maximum likelihood estimation.

[0049] Another approach allows the number of occupied parking spaces to be determined based on the ratio of parking pressure to parking density. Parking pressure can be estimated using the macro model and the parking rate λ. p The parking density is determined using a geographic map model. Such a map can be generated, for example, using so-called "kernel density estimation" based on previously used and thus learned parking spaces.

[0050] With reference to the Fig. The procedure is explained in more detail in sections 3 to 5 below. Fig. Figure 3 shows the initialization step, in which offline optimization of the parking segments and the statistical parameters of the request rate λ(t) and the average parking duration 1 / µ(t) takes place. For this purpose, historical data (parking and exiting operations, search traffic, parking space monitor, occupancy data) are collected. A parking space monitor is described in the aforementioned, subsequently published German patent application DE 10 2012 201 472 A1 of the applicant, the contents of which are incorporated by reference. The offline macro model calculates an initial set of statistical parameters as well as an initial map of the parking segments. The results of the calculations are stored in database 13. The system can then be started.

[0051] Fig. Figure 4 shows the step of updating the model during operation. First, the central computer 10 receives messages ("parking messages") from the data provider 14, containing updated information about available parking spaces. For example, a detected parking exit is transmitted to the central computer. The processing unit 12 determines the affected parking segment(s) from the database 13. The information is processed in real time in the macro model 20. This can be done, for example, using kernel density estimation in combination with a Kalman filter derivative. Subsequently, the micro model 21 updates the model, also in real time. For this purpose, a BD process 22 can be numerically integrated, for example. The update 23 can be performed using a shift operator. The results of the calculations are stored in the database 13.

[0052] Fig.Figure 5 shows the use of the model during operation. Here, a user request ("parking request") is received via user interface 1. This request is processed by the computing unit 12, which determines the parking segment(s) necessary to answer the request and retrieves the associated data from the database. The data is then evaluated in real time using the micromodel 21, in which the BD process 22 is numerically integrated. This evaluation is performed by the computing unit. Finally, the probability distribution of available parking spaces for the parking segment is sent back to the user via user interface 11. Reference symbol list 10 central computers 11 User Interface 12 computing units 13 Database 14 Data provider 20 Macro model 21 Micromodel (µmodel) 22 BD process 23 Update of the micromodel K1 communication path K2 communication path K3 communication path K4 communication path P(t) Probability of a free parking space m Number of free parking spaces

Claims

[1] Method for providing parking information on available parking spaces, wherein a) historical and current information on available, free parking spaces is determined for at least one street section, whereby at least one parking segment comprising one or more streets is determined from the information obtained, and statistical parameters on available parking spaces are generated for each parking segment from the information obtained; b) a model is created for each parking segment, in which the information determined for the parking segment is processed to represent a parking state of the respective parking segment as a probability distribution (P i ) to determine, characterized by , that initial information about a request rate (λ) is determined, which indicates the number of requests for a parking space per time for a parking segment. [2] Method according to claim 1, wherein an updated probability distribution (P) is derived from real-time information about free, available parking spaces in a parking segment. i ) the parking condition is determined. [3] Method according to claim 1 or 2, wherein the information on available, free parking spaces is recorded by means of measuring equipment for vehicles in traffic. [4] Method according to one of the preceding claims, wherein the information on available, free parking spaces is measured by sensors arranged along the streets. [5] Method according to one of the preceding claims, wherein the information on available, free parking spaces is generated manually by users entering it into a terminal device. [6] A method according to any of the preceding claims, wherein the information on available, free parking spaces is automatically generated by a vehicle by providing corresponding information during each parking and exit operation. [7] Method according to any of the preceding claims, wherein the information on available, free parking spaces is transmitted to a central computer (10) that generates and / or manages the knowledge database. [8] Method according to one of the preceding claims, wherein second information about the parking of vehicles in and / or the parking of vehicles out of a parking space is determined, wherein a parking rate (µ) is determined from the holding times between the parking and the parking of a respective vehicle. [9] Method according to one of the preceding claims, wherein third information about a fixed number of detected free parking spaces is obtained as information. [10] Method according to claim 9, wherein at least one basic function is assigned to each parking segment, which represents the exit rate (µ) and the request rate (λ) as a function of a parking characteristic. [11] Method according to any of the preceding claims, wherein the modeling for each park segment is carried out by a birth-death method. [12] Computer program product that can be loaded directly into the internal memory of a digital computer or computer system and comprises software code sections that perform the steps according to any of the preceding claims when the product is running on the computer or computer system. [13] System for providing parking information on available parking spaces in at least one street, comprising a) a first unit for determining historical and current information on available, free parking spaces for at least one street section, which is designed to determine at least one parking segment comprising one or more streets from the determined information and to generate statistical parameters on available parking spaces for each parking segment from the determined information; b) a second unit for generating a model for each parking segment, which is trained to process the information determined for the parking segment in order to determine a parking state of the respective parking segment as a probability distribution, characterized by, that the first unit is trained to determine initial information about a request rate (λ), which indicates the number of requests for a parking space per time for a parking segment. [14] System according to claim 13, comprising further means for carrying out the method according to any one of claims 2 to 11.

Images