METHOD FOR AUTOMATICALLY IDENTIFYING PARKING AND / OR NON-PARK AREAS

Description

[0001] The invention relates to a method for the automated identification of parking areas and / or non-parking areas as well as a device and a computer program. State of the art

[0002] The invention relates to a method for the automated identification of parking spaces and / or non-parking spaces. Further aspects of the invention relate to a computer program, a central computer device, and a device configured to carry out the method.

[0003] With the increasing level of vehicle automation, increasingly complex driver assistance systems are being deployed. Such driver assistance systems and functions, such as highly automated or fully automated driving, require a large number of sensors in the vehicle that enable precise detection of the vehicle's surroundings.

[0004] The recording of the vehicle environment includes, among other things, procedures for the automated identification of parking spaces and / or non-parking spaces.

[0005] In the automotive sector, various driver assistance systems are also used to assist the driver in performing various driving maneuvers. These include, for example, parking assistance systems that use sensors assigned to the vehicle to monitor the surroundings, identify possible parking spaces in the vicinity, and assist the driver in parking. Furthermore, driver assistance systems are known in the state of the art that assist the driver in locating suitable free parking spaces.

[0006] In the following, higher automation is understood to mean all those levels of automation which, in the sense of the Federal Highway Research Institute (BASt), correspond to automated longitudinal and lateral guidance with increasing system responsibility, e.g. highly and fully automated driving.

[0007] In the state of the art, various methods are known for determining distances of a vehicle to objects using distance-based sensors (e.g. ultrasonic, radar, laser, video, lidar sensors).

[0008] A transfer of parking space data to a server is known, for example, from DE 10 2004 062 021 A1, DE 10 2009 028 024 A1, DE 10 2008 028 550 A1, US2017 / 243486 A1, and DE 102015207804 A1.

[0009] For example, DE 10 2004 062 021 A1 discloses a system for utilizing available parking spaces. It involves road users determining the location and dimensions of available parking spaces while driving past and transmitting the collected data to a central location. This data is then made available by the central location to road users seeking parking spaces.

[0010] DE 10 2009 028 024 A1 discloses a parking guidance system and a navigation device for navigating a vehicle searching for a parking space to a free parking space. Information about available, free parking spaces is retrieved from vehicles in traffic and transmitted directly to the vehicle searching for a parking space or indirectly via a central control center.

[0011] It is an object of the present invention to provide an improved method for the automated identification of parking areas and / or non-parking areas. Disclosure of the invention

[0012] This object is achieved by means of the respective subject matter of the independent claims. Advantageous embodiments of the invention are the subject matter of the respective dependent subclaims.

[0013] According to one aspect of the invention, a method for automatically identifying parking areas and / or non-parking areas is provided, comprising the following steps: S1 Carrying out several drives, but at least two drives, through at least one road section with at least one measuring vehicle; S2 Capturing parking space data along the road section during each passage by means of at least one sensor of the at least one measuring vehicle, wherein the parking space data comprises parking space detections as well as respectively associated location and time information; S3 Accumulation of the parking space data collected during each passage into a data set; S4 Determination of a temporal change rate of parking space detections based on the data set; and S5 Determination of valid parking spaces and / or non-parking spaces along the road section based on the temporal change rate of various parking space detections,

[0014] In this way, in step S1, parking spaces between two vehicles are continuously measured and detected by the measuring vehicle, wherein in step S2, parking space data along the road section are recorded during each passage by means of at least one sensor, wherein the parking space data comprise parking space detections as well as respectively associated location and time information.

[0015] A further advantageous development of the method provides for the driving of the road section in different directions, with the parking space data along the road section being correlated during each drive in the different directions. In this way, for example, the determination of the parking space data can be advantageously carried out even more accurately based on the driving of the road section in opposite directions.

[0016] The parking space data are assigned to so-called bins, whereby a bin designates a defined section along the road section, and whereby the assignment of a parking space detection to a bin is carried out in particular on the basis of the location information assigned to the respective parking space detection.

[0017] Preferably, the length of the bins is freely configurable. For example, an average vehicle length of five meters can be selected. A further advantageous development of the method is advantageous in that the parking space data is determined for a defined length of the road section. In this way, temporal change rates of parking space detections can be created for selectively selected areas.

[0018] The temporal change rate of the parking space detections according to step S4 is determined jointly for each parking space detection summarized in a bin, so that as a result of this determination, each bin is assigned a temporal change rate which results from the temporal change rates of the parking space detections summarized in the bin.

[0019] The temporal rate of change represents the degree of temporal stability of area occupancy, such as parked cars, whereby for temporally constant events, as is typical for non-parking areas, the temporal rate of change is low and in the case of parking areas that are subject to frequent occupancy changes, the temporal rate of change is high.

[0020] For the further procedure, the invention advantageously provides that the step of determining valid parking areas and / or non-parking areas along the road section (S5) is carried out on the basis of the temporal change rate of the respective bins, wherein a respective bin is classified as a parking area or as a non-parking area on the basis of the temporal change rate of parking space detections assigned to it.

[0021] For example, the temporal rate of change can be normalized by the number of passages through a street. If the temporal rate of change is low, a parking space is detected for almost every passage in that bin. If the temporal rate of change is high, parking space detections occur only sporadically and irregularly. Accordingly, this method can be used to identify and separate non-parking areas (e.g., a courtyard entrance, a no-parking zone, or a tree-lined area) from parking areas.

[0022] In the invention, the parking space detections as well as the parking areas and / or non-parking areas are incorporated into a digital parking space map, wherein, for example, a courtyard entrance, a no-parking zone, a T-road intersection and / or a tree area can be registered as non-parking areas.

[0023] In a further embodiment of the invention, the method includes that the sensor used in step S2 to acquire parking space data operates according to a distance-based measuring method.

[0024] For the further procedure, one embodiment of the invention advantageously provides that the accumulation performed in step S3 takes place at least partially locally in the measuring vehicle and / or at least partially in a central server device. In this way, a large amount of historical data can be aggregated on the server device over the long term, representing a high level of timeliness and accuracy of the parking space data.

[0025] Advantageously, the determined temporal change rates of the parking space detections are used to select a parameterization of cluster-based methods for learning parking areas and non-parking areas.

[0026] A further subject of the invention is a device for the automated identification of parking areas and / or non-parking areas, comprising a measuring vehicle with at least one sensor, wherein data relating to a parking space along the road section can be detected by means of the sensor during each passage, further comprising a control device, wherein the control device is designed to carry out a method according to one of claims 1 to 5.

[0027] Preferably, the measuring vehicle has at least one communication device for transmitting the data to a server device.

[0028] In a particularly preferred embodiment, the sensor is designed as an ultrasonic sensor or radar sensor.

[0029] Furthermore, a computer program comprising program code means for carrying out the method according to one of claims 1 to 5, when the computer program is executed on a device for the automated identification of parking areas and / or non-parking areas, also forms an object of the invention.

[0030] Although the present invention is described below mainly in connection with passenger cars, it is not limited thereto, but can be used with any type of vehicle, commercial vehicle (truck) and / or passenger car (car).

[0031] Further features, possible applications, and advantages of the invention will become apparent from the following description of an exemplary embodiment of the invention, which is illustrated in the figure. It should be noted that the features illustrated are merely descriptive and can also be used in combination with features of other developments described above. They are not intended to limit the invention in any way. Drawings

[0032] The invention is explained in more detail below using a preferred embodiment, wherein the same reference numerals are used for the same features. The drawing is schematic and shows: Fig. 1 a schematic view of a road section with assigned parking space detections and their classification into bins.

[0033] Figure 1The upper section shows a road section 10 with a carriageway 12. At one edge of carriageway 5, there are parked vehicles in certain areas 12 and 14. In an area 16, a tree area is located at the edge of the roadway, and in an area 18, a courtyard entrance leads onto carriageway 5. While areas 12 and 14 are parking areas, areas 16 and 18 are non-parking areas.

[0034] The figure also shows a measuring vehicle 20, which has at least one sensor (not shown in detail) suitable for recording parking space data. The sensor can, in particular, be a sensor that operates according to a distance-based measurement method, such as a radar or ultrasonic sensor.

[0035] A step S1 of the method according to the invention includes that the measuring vehicle 20 carries out several passes, but at least two passes, through the road section 10.

[0036] In a step S2, parking space data along the road section 10 are collected for each passage, using, among other things, the at least one sensor. The parking space data includes parking space detections as well as associated location and time information. Location information can, for example, be a running coordinate running along the road section or GPS information. Time information can, for example, be the serial number of the passage of the measuring vehicle 20 through the road section 10 or a time at which the parking space detection was made.

[0037] The measurement of the road section by the sensor, i.e. the recording of the parking space data, is preferably carried out continuously or quasi-continuously.

[0038] The term "parking space detection" includes both the information that a parking space was detected at a specific location on the edge of road section 10 and the information that no parking space was detected there.

[0039] In step S3, the parking space data recorded during each passage are accumulated in a data set. Figure 1 This is symbolized in the middle area of the figure by a horizontal line 30, which represents a running coordinate of the road section 10, and points 35 arranged on it. In the figure, each point represents a parking space detection. The parking space detections shown result from several passing bys of one or more measuring vehicles 20. The parking space detections are incorporated into a digital map via the location information, which is why the parking space detections in Figure 1referred to as "map-matched" parking space detections. The data set can be stored in a suitable memory of the measuring vehicle 20 or in a central server computer.

[0040] A concentration of parking space detections can be seen in area 18, the courtyard entrance, as this area is generally clear and is detected almost every time a vehicle passes by. Due to positioning inaccuracies, which may be based on inaccurate GPS, for example, the parking space detections are spread out.

[0041] In areas 12 and 14, there is a medium concentration of parking space detections, since these areas, which are parking areas, are occupied by parked vehicles most of the time, but are occasionally free.

[0042] In area 16, i.e., the area of the tree area, a vanishingly small concentration of parking space detections can be observed. These parking space detections may, for example, result from very sporadic false detections due to device-related reasons.

[0043] In a step S4 of the invention, temporal change rates of parking space detections are determined based on the data set. For this purpose, clusters of parking space detections are advantageously sorted into so-called "bins" 40, 42, 44, 46, 48, 50, 52 based on the location information assigned to them, which is shown in the lower part of the Figure 1A bin 40, 42, 44, 46, 48, 50, 52 corresponds to a section of a specific length along road section 10. A bin 40, 42, 44, 46, 48, 50, 52 can, in principle, be freely defined and can, for example, have a length of 5 m, which corresponds to an average vehicle length. The assignment of a parking space detection to a bin can, for example, be based on the location information associated with the respective parking space detection.

[0044] In a further step S5 of the invention, valid parking spaces and / or non-parking spaces along the road section are determined based on the temporal change rate of various parking space detections.

[0045] The temporal rate of change is determined by the temporal stability of events. For regularly recurring events, the temporal rate of change is low. For sporadically occurring events, the temporal rate of change is high. The temporal rate of change is normalized by the number of passages through a road.

[0046] In the Figure 1 In the example shown, the temporal rate of change in bin 44 is low, since a parking space is detected during almost every passage through this bin. The temporal rate of change in bins 40, 42, 46, and 48 is high, since parking space detections occur only sporadically and irregularly. In bin 50, the temporal rate of change is low, since a parking space is detected during almost no passage through these bins. In bin 52, the temporal rate of change is zero, since no parking space detections occur in this bin.

[0047] In the example of Figure 1 Thus, the temporal change rate of the parking space detections is determined jointly according to step S4 for each parking space detection summarized in a bin, so that as a result of this determination, each bin is assigned a temporal change rate which results from the temporal change rates of the parking space detections summarized in the bin.

[0048] Now, by assigning bins 40, 42, 44, 46, 48, 50 and 52 to areas 12, 14, 16 and 18, existing non-parking areas, in this case areas 16 and 18, and currently free parking areas, in this case areas 12 and 14, can be determined and separated from each other.

[0049] According to one embodiment of the invention, in step S5, valid parking areas and / or non-parking areas along the road section are determined based on the temporal change rate of the respective bins, wherein a respective bin is classified as a parking area or as a non-parking area based on the temporal change rate of parking space detections assigned to it.

[0050] With this information, a digital parking space map can be generated with information about parking areas where vehicles are allowed to park, as well as non-parking areas where parking is not permitted.

[0051] Advantageously, the results of the temporal rate of change of parking space detections can be used to select the parameters of cluster-based methods for learning parking spaces and non-parking spaces. If the temporal rate of change between the bins is only small, the probability of detections between parking and non-parking spaces is approximately the same, and the parameters of a clustering method should be chosen accordingly conservatively (e.g., a high expected density for density-based methods). Conversely, if there are significant differences in the temporal rate of change, a high occupancy rate of the parking spaces can be assumed, and the clustering can be parameterized accordingly more broadly.

[0052] The invention is not limited to the described and illustrated embodiment. Rather, it also encompasses all expert developments within the scope of the invention defined by the patent claims.

[0053] In addition to the embodiments described and illustrated, further embodiments are conceivable, which may include further modifications and combinations of features.

Claims

1. Method for the automated identification of parking areas and / or non-parking areas, comprising the following steps: S1Performing a plurality of passes, but at least two passes, through at least one road section with at least one measuring vehicle (20);S2Capturing parking space data along the road section (10) during each pass by means of at least one sensor of the at least one measuring vehicle (20), wherein the parking space data comprise detected parking spaces as well as respectively assigned location and time information;S3Accumulating the parking space data captured during each pass in a data set;S4Determining a temporal rate of change of detected parking spaces based on the data set; andS5Determining valid parking areas and / or non-parking areas along the road section (10) based on the temporal rate of change of various detected parking spaces; wherein the parking space data are assigned to so-called bins (40, 42, 44, 46, 48, 50, 52), where a bin (40, 42, 44, 46, 48, 50, 52) indicates a defined section along the road section (10), and wherein a detected parking space is assigned to a bin (40, 42, 44, 46, 48, 50, 52) in particular based on the location information assigned to the respective detected parking space; and wherein the temporal rate of change of the detected parking spaces is determined jointly according to step S4 for detected parking spaces respectively combined in a bin (40, 42, 44, 46, 48, 50, 52), and so, as a result of this determination, each bin (40, 42, 44, 46, 48, 50, 52) is assigned a temporal rate of change that results from the temporal rates of change of the detected parking spaces combined in the bin (40, 42, 44, 46, 48, 50, 52), wherein the detected parking spaces as well as the parking areas and / or non-parking areas are incorporated in a digital parking space map.

2. Method according to Claim 1, characterized in that, in step S5, valid parking areas and / or non-parking areas along the road section (10) are determined based on the temporal rate of change of the respective bins (40, 42, 44, 46, 48, 50, 52), wherein a respective bin (40, 42, 44, 46, 48, 50, 52) is classified as a parking area or as a non-parking area on the basis of the temporal rate of change of detected parking spaces that is assigned to it.

3. Method according to one of the preceding claims, characterized in that the sensor used in step S2 to capture parking space data works according to a distance-based measuring method.

4. Method according to one of the preceding claims, characterized in that the accumulation carried out in step S3 takes place at least partially locally in the measuring vehicle (20) and / or at least partially in a central server device.

5. Method according to one of the preceding claims, characterized in that the determined temporal rates of change of the detected parking spaces are used to select a parameterization of cluster-based methods for learning parking areas and non-parking areas.

6. Apparatus for the automated identification of parking areas and / or non-parking areas, having a measuring vehicle (20) with at least one sensor, wherein data relating to a parking space along the road section (10) can be captured by means of the sensor during each pass, further having a control device, wherein the control apparatus is designed to carry out a method according to one of Claims 1 to 5.

7. Apparatus according to Claim 6, wherein the measuring vehicle (20) has at least one communication device for transmitting the data to a server device.

8. Apparatus according to Claim 6 or 7, wherein the sensor is designed as an ultrasonic sensor or radar sensor.

9. Computer program with program code means for carrying out the method according to one of Claims 1 to 5 when it is executed on an apparatus for the automated identification of parking areas and / or non-parking areas.