Method for predicting the availability of feature-based localisation of a vehicle, and method for controlling a vehicle

Abstract

The invention relates to a method (100) for predicting the availability of feature-based localisation for a vehicle (401), comprising: receiving (101) map data (412) from a feature map (411) of a road traffic network; identifying (103), on the basis of the feature information from the feature map (411), features (413) for at least one lane (403) to be used by the vehicle (401); calculating (105) an availability value for the features (413) of the at least one lane (403) to be used by the vehicle (401) with respect to an availability criterion; and outputting (107) an availability indicator that comprises the availability value and relates to the features (413) of the lane (403) to be used.

Description

[0001] Description

[0002] title

[0003] Method for availability prediction of a feature-based localization of a vehicle and method for controlling a vehicle

[0004] The invention relates to a method for predicting the availability of a feature-based localization of a vehicle. The invention further relates to a method for route planning for an autonomously controllable vehicle. The invention further relates to a method for controlling a vehicle.

[0005] State of the art

[0006] Automated and highly automated driving are considered trends of the future. Map representations of the vehicle's surroundings are of utmost importance for the orientation of autonomously controlled vehicles. Vehicle localization is achieved primarily by comparing features with corresponding feature maps and objects located in the vehicle's surroundings detected by the environment sensors. If such localization is not possible because the corresponding objects are not detected by the environment sensors during the journey, autonomous control cannot occur. For autonomous vehicle control, it is therefore crucial to select routes where reliable detection of a sufficient number of objects can be expected for comparison with corresponding features on a feature map.

[0007] It is therefore an object of the invention to provide an improved method for availability prediction of a feature-based localization of a vehicle, an improved method for route planning for an autonomously controllable vehicle and an improved method for controlling a vehicle.

[0008] This problem is solved by the method for predicting the availability of a feature-based localization of a vehicle, the method for route planning for an autonomously controllable vehicle, and the method for controlling a vehicle of the independent claims. Advantageous embodiments are the subject of the subordinate claims.

[0009] According to one aspect of the invention, a method for predicting the availability of a feature-based localization of a vehicle is provided, comprising:

[0010] Receiving map data of a feature map of a road network, wherein the feature map comprises feature information of a plurality of roadways that can be driven by a vehicle, and wherein the features are configured to be detected by a vehicle driving on the roadway via sensor data from environmental sensors of the vehicle and to be used to localize the vehicle;

[0011] Determining features for at least one lane to be traveled by the vehicle based on the feature information of the feature map;

[0012] Calculating an availability value of the features of the at least one roadway to be traveled by the vehicle with respect to an availability criterion, wherein the availability value indicates a quantitative measure of the usability of the features for locating the vehicle when the vehicle travels the roadway; and

[0013] Output of availability information including the availability value regarding the characteristics of the roadway to be traveled.

[0014] This makes it possible to achieve the technical advantage of providing an improved method for predicting the availability of feature-based vehicle localization. The method according to the invention is used in particular for route planning for vehicles, for example autonomously controllable vehicles. For a route to be traveled by a vehicle, features of a feature map are thus determined for the respective lanes of the route, and availability values ​​are calculated for the respective features. The availability values ​​describe a quantitative measure of the usability of the features for localizing the vehicle when traveling along the lanes of the selected route. Features of the feature map can be used to localize the vehicle if the respective features can be detected by the vehicle's environmental sensors when the vehicle travels along the respective lane of the selected route.Features of the feature map correspond to corresponding objects within the environment. The objects can be arranged along the respective roadway the vehicle is to travel on. The objects can, for example, be arranged at the edge of the roadway to be traveled on. However, depending on the position of the vehicle on the roadway, the type of environmental sensors or environmental conditions such as weather conditions or traffic volume, not all objects can be viewed by the vehicle's environmental sensors at all times or each time the vehicle travels on the roadway and are therefore not available for determining the vehicle's position. After calculating the availability values ​​for the majority of features of the various roadways, corresponding availability information regarding the features of the roadways to be traveled on is output.Using the provided availability information, a corresponding route selection can be made based on the availability of the features or objects located along the individual lanes of the selected route. For example, based on the availability information, a route planning process can provide the driver with a route with the shortest travel time, as well as a route with the highest availability of the features. Availability describes a value, such as a percentage or a probability value, that indicates how many of the objects can be viewed or the probability of detecting the objects.

[0015] According to one embodiment, the availability criterion comprises a

[0016] Detectable features by the vehicle's environmental sensors based on a sensor type and / or a directional characteristic of the vehicle's environmental sensors, wherein calculating the availability value comprises: determining a sensor type and / or a directional characteristic of at least one environmental sensor of the vehicle; and determining the detectability of the features of the roadway to be traveled by the vehicle for the sensor type and / or the directional characteristic of the at least one environmental sensor.

[0017] This provides the technical advantage of enabling a precise calculation of the availability of the features stored in the feature map for locating a vehicle. For this purpose, the availability criterion used to calculate the availability value is defined as the detectability of the features by the environmental sensors of the vehicle traveling along the roadway, based on a sensor type and / or a directional characteristic of the vehicle's environmental sensors. The type of environmental sensors can include the configuration of the environmental sensors as camera sensors, radar sensors, or LIDAR sensors. According to the invention, the directional characteristic is to be understood as a corresponding direction-dependent sensitivity of the respective environmental sensor.Based on the sensor type and the directional characteristic, the detectability of the features arranged along the roadway the vehicle is traveling on can be calculated for each of the vehicle's environmental sensors. The detectability of a feature by the respective environmental sensor can thus depend on the position of the feature relative to the vehicle, in that the respective feature is fully, partially, or not at all mapped by the directional characteristic of the respective environmental sensor. Likewise, the detectability of a feature can be determined by the respective nature of the feature, in that, depending on the reflectivity of the feature, it can be detected well, poorly, or not at all by different types of environmental sensors.For example, objects that can be easily detected by a radar sensor may be virtually undetectable by a camera sensor, for example at night and in insufficient lighting, and thus cannot be used to locate the vehicle. According to one embodiment, the availability criterion includes the detectability of the features by the vehicle's environmental sensors, taking into account the traffic volume on the roadway to be traveled, wherein calculating the availability value includes:

[0018] Determining the traffic volume on the roadway to be traveled by the vehicle; and

[0019] Determining the detectability of the characteristics of the roadway to be traveled by the vehicle by calculating a coverage of the characteristics of the roadway by the traffic volume on the roadway.

[0020] This can achieve the technical advantage of further improving the calculation of the availability information for the individual features of the feature map. For this purpose, the availability criterion is further calculated based on the detectability of the features by the vehicle's environmental sensors, taking into account the traffic volume on the roadway the vehicle is traveling on. Since the individual features are usually arranged along the roadway the vehicle is traveling on, the detectability of the features by the vehicle's environmental sensors can be significantly impaired not only by the properties of the respective environmental sensors but also by other road vehicles located in the vehicle's vicinity.In the case of a high traffic volume, in which a correspondingly high number of other vehicles are arranged in the vicinity of the vehicle, each of which hinders the detection of the features arranged at the edge of the roadway by the vehicle's environmental sensors, there is a correspondingly low availability of the features of the respective roadway, since these cannot be detected or recorded by the environmental sensors due to the blocking other vehicles.

[0021] According to one embodiment, determining the traffic volume comprises: determining an average number and / or an average speed of vehicles per unit of time on the roadway to be traveled by the vehicle; and wherein determining the detectability comprises: calculating an average coverage of the features of the roadway based on the average number and / or speed of vehicles per unit of time.

[0022] This provides the technical advantage of enabling a precise prediction of the traffic volume for the respective lane to be traveled by the vehicle. This, in turn, contributes to an improved determination of the availability information regarding the respective lane. Since the method according to the invention is primarily intended for route planning, in order to determine the availability of the features of a lane to be traveled by the respective vehicle in the future, a traffic volume must be predicted that will likely prevail on the respective lane at the time the vehicle enters the lane.By calculating the average traffic volume and the resulting average coverage of the characteristics by the vehicles in the traffic volume, a precise and probable prediction of the expected traffic volume for any given point in the future can be achieved. The average traffic volume can be calculated based on the average number and average speed of vehicles on the respective roadway.

[0023] According to one embodiment, calculating the availability value comprises determining at least one characteristic of the features, wherein the at least one characteristic comprises: type of feature, number of features, number of features of a specific type or category, extent and / or size of the features, reflectivity of the features; recognition value of the features; and wherein the detectability is determined with respect to the at least one specific characteristic of the features, and a detectability-corrected characteristic is calculated. This provides the technical advantage of further specifying the calculation of the availability value or the availability information.The detectability of the majority of features of the respective roadway can be further quantified using the various characteristics of the features, such as the type of feature, the number of features, or the spatial extent of the features. The characteristics can be extracted from the detected features, in particular by executing an appropriately trained artificial intelligence. According to the invention, the characteristics are, in particular, statistical quantities that describe the set of features in their entirety.

[0024] According to one embodiment, calculating the availability value comprises: comparing the at least one characteristic of the features of the map representation with the corresponding detectability-corrected characteristic.

[0025] This can achieve the technical advantage of enabling further precision in the calculation of the availability value. For this purpose, a comparison of the characteristics of the features in the feature map is performed with a correspondingly detectable, corrected characteristic of the features. For this purpose, for example, the number of features within the feature map can be compared or related to the number of features predicted for a specific traffic volume and correspondingly reduced. Due to the corresponding traffic volume and the other vehicles in the vehicle's vicinity, not all features of the roadway can be detected by the vehicle.By comparing the features of the feature map with the number of features that can be detected by the vehicle's environmental sensors, calculated for the respective traffic volume, a quantitative value for the availability of the respective features of the feature map can be calculated.

[0026] According to one embodiment, the determination of the detectability is carried out for at least one previously known position of the vehicle on the roadway or for different previously known positions of the vehicle on different lanes of the roadway.

[0027] This provides the technical advantage of allowing the position-dependent detection of the features of the respective lane to be traveled. By taking position into account, the availability of the features in the feature map can be further refined. Taking different lanes of the lane to be traveled further refines the availability determination.

[0028] According to one embodiment, the determination of the detectability is carried out by a correspondingly trained artificial intelligence, wherein the artificial intelligence is trained on environmental sensor data from environmental sensors of at least one vehicle, and wherein the environmental sensor data were recorded during a plurality of journeys of the vehicle along the at least one roadway during a plurality of different traffic volumes.

[0029] This provides the technical advantage of enabling a precise, fast and reliable determination of detectability.

[0030] According to one embodiment, the features are characteristic objects arranged at an edge of the roadway and include: buildings, traffic signs, road markings.

[0031] This can achieve the technical advantage of enabling comprehensive consideration of various features. In particular, the characteristic objects are defined as all objects suitable for localization and detectable by radar, lidar, ultrasound, and video. They can include directional arrows, lane markings, especially broken lines, trees, other vegetation, bridges, distributed structures such as poles and posts, and roadway delineation elements such as paving stones, guardrails, and bollards. This list is not intended to be exhaustive. In addition, features can be detected and used that cannot be semantically named but can be extracted from the raw data of the environmental sensors by appropriately trained artificial intelligence.The artificial intelligence typically recognizes or detects key points in the environmental sensor data and assigns them a vector describing the type of point. These key points, which are unknown to the operator of the artificial intelligence and cannot be semantically identified, but which are nevertheless reliably detected by the corresponding artificial intelligence, can also serve as characteristic objects within the meaning of the invention.

[0032] According to a further aspect, a method for route planning for an autonomously controllable vehicle is provided, comprising:

[0033] Executing the method for availability prediction of a feature-based localization of a vehicle according to one of the preceding embodiments on a plurality of possible routes of the vehicle between a predetermined starting point and a predetermined end point; and

[0034] Determining a route of the vehicle with a highest predicted availability of a feature-based localization of the vehicle based on the performed availability prediction.

[0035] This makes it possible to provide the technical advantage of improved route planning for an autonomously controllable vehicle, where, in addition to a fastest route, a route can be provided on which the highest availability of the features required for positioning can be expected, and on which, thus, fully autonomous control of the vehicle is possible with the highest probability.

[0036] According to a further aspect, a method for controlling a vehicle is provided, comprising:

[0037] Executing the route planning procedure for an autonomously controllable vehicle; and

[0038] Controlling the vehicle based on the selected route. This allows the technical advantage of improved control of an autonomously controllable vehicle to be achieved, with the control system relying on improved route planning and improved availability prediction with the aforementioned technical advantages.

[0039] According to one embodiment, controlling the vehicle comprises executing maneuver planning, wherein the maneuver planning comprises a driving maneuver of the vehicle that enables optimal availability of the features of the features of the feature map.

[0040] This can achieve the technical advantage of improved vehicle control. In particular, based on the results of the availability determination, the vehicle can execute a driving maneuver that optimizes availability. For example, the driving maneuver can involve changing lanes on the roadway to be traveled if a difference in the availability of features for different lanes has been detected.

[0041] According to a further aspect, a computing unit is provided which is configured to execute the method for availability prediction of a feature-based localization of a vehicle according to one of the preceding embodiments and / or the method for route planning for an autonomously controllable vehicle and / or the method for controlling a vehicle.

[0042] According to a further aspect, a computer program product comprising instructions is provided which, when executed by a data processing unit, cause the data processing unit to execute the method for predicting the availability of a feature-based localization of a vehicle according to one of the preceding embodiments and / or the method for route planning for an autonomously controllable vehicle and / or the method for controlling a vehicle. Embodiments of the invention are explained with reference to the following drawings. In the drawings:

[0043] Fig. 1 schematic representation of a system for controlling a vehicle;

[0044] Fig. 2 is a flowchart of a method for availability prediction of a feature-based localization of a vehicle according to one embodiment;

[0045] Fig. 3 is a flowchart of a method for route planning for an autonomously controllable vehicle according to one embodiment;

[0046] Fig. 4 is a flowchart of a method for controlling a vehicle according to an embodiment; and

[0047] Fig. 5 is a schematic representation of a computer program product for executing the method for availability prediction of a feature-based localization of a vehicle and / or the method for route planning for an autonomously controllable vehicle and / or the method for controlling a vehicle.

[0048] Fig. 1 shows a schematic representation of a system 400 for controlling a vehicle 401.

[0049] Graphic A of Figure 1 shows a vehicle 401 traveling on a roadway 403. Various features 413 are arranged along the roadway 403. The features 413 are designed as various objects, for example buildings, vegetation, traffic signs, or the like, by means of which orientation of the vehicle 401 is enabled. Graphic A further shows a feature map 413. The feature map 413 shows the roadway 403 shown with the correspondingly arranged features 413. Graphic A further shows a computing unit 415 with an installed artificial intelligence 419. The computing unit 415 is configured to execute the inventive method for availability prediction of a feature-based localization of a vehicle 401.

[0050] The vehicle 401 further comprises at least one environment sensor 405, which enables detection of the surroundings of the vehicle 401. The vehicle 401 further comprises a computing unit 417. The computing unit 417 can be configured, for example, for autonomous or semi-automated control of the vehicle.

[0051] The computing unit 415 can be designed, for example, as an external server unit or as an external cloud server.

[0052] To implement the method according to the invention for predicting the availability of a feature-based localization of a vehicle 401, the computing unit 415 first receives the feature map 411 in the form of corresponding map data 412. Furthermore, the computing unit 415 is provided with sensor information data 407 from the environmental sensors 405 of the vehicle 401. The sensor information data 407 can, for example, include information regarding the sensor type of the environmental sensors or a directional characteristic of the respective environmental sensors 405 of the vehicle 401. Furthermore, the computing unit 415 can be provided with traffic volume data 409 regarding a roadway 403. The traffic volume data 409 can be current data regarding traffic volume in a specific roadway or a plurality of different roadways at a given point in time. Alternatively, the data can refer to past points in time.Alternatively, the data can describe average traffic volumes. The data 409 can, for example, be transmitted by other vehicles that were traveling on the respective roadway at the specified time. Alternatively, the data 409 can be archived data. The data 409 can further include the number and / or speeds of the vehicles. To calculate the availability prediction of a feature-based localization of a vehicle 401, the computing unit 415 calculates an availability value for a roadway 403 to be traveled by a vehicle 401 based on the map data 412 of the feature map 411 for the features 413 arranged along the roadway 403, taking into account an availability criterion.

[0053] According to one embodiment, the availability criterion of the individual features 413 can take into account the detectability of the features 413 by the environmental sensors 405 of the respective vehicle 401. For this purpose, the information from the sensor information data 407 regarding the sensor type and / or the directional characteristic 410 of the individual environmental sensors 405 can be taken into account. Based on the feature information of the feature map 411, a corresponding detectability of the respective feature 413 by the environmental sensors 405 can be calculated for each feature 413 of the feature map 411 arranged along the roadway 403 on which the vehicle 401 is to travel, according to the respective sensor type or directional characteristic 410 of the respective environmental sensors 405 of the vehicle 401. The detectability describes a quantifiable quantity according to which the respective features 413 can be detected by the environmental sensors 405 of the vehicle 401.This detectability may depend, for example, on the nature of the respective feature 413 and / or the nature of the environmental sensors. For example, features 413 may have advantageous reflectivity for LIDAR or radar sensors, but be difficult to detect for camera sensors, for example, due to low ambient light.

[0054] Alternatively or additionally, the detectability of the features 413 can be calculated taking into account the expected traffic volume on the respective lane 403. Based on corresponding data 409 of the current or average traffic volume that can be expected on the respective lane 403, the traffic volume for the time at which the vehicle 401 will travel on the corresponding lane 403 can be calculated. For the calculation, for example, an average speed or an average number of vehicles that can be expected on the lane 403 at the future time can be used to calculate an average coverage. The average coverage describes a coverage of the features 413 arranged at the edge of the lane 403 by the other vehicles that are on average arranged due to the traffic volume in the vicinity of the vehicle 401 when traveling on the lane 403.

[0055] Graphic B graphically illustrates such a calculation of the detectability of the individual features 413 for the environmental sensors 405 of a vehicle 401 traveling on the roadway 403. In the illustration shown, the vehicle 401 is located in the left lane 404 of the roadway 403. Three other vehicles of the expected traffic volume 409 are also shown in the right lane 404 of the roadway 403. Furthermore, Graphic B shows three directional characteristics 410 from three different environmental sensors 405 of the vehicle 401. One directional characteristic 410 is directed toward the front of the vehicle 401, while the two other directional characteristics 410 are directed toward the rear of the vehicle 401. In the illustration shown, the features 413 arranged on the right edge of the roadway 403 are covered by the other vehicles of the traffic volume 409 for the environment sensors 405 of the vehicle 401.This is illustrated in graph B by an average coverage 414. The average coverage 414 can, for example, be specified as a percentage. Of the six features 413 arranged on the left edge of the roadway 403, only five features 413 in the illustration shown lie in the areas of the directional characteristics 410. The feature 413 not covered by the directional characteristics 410 cannot therefore be detected by the vehicle 401 in the position P shown. According to the invention, the availability of the features 413 of the feature map 411 can be calculated position-dependently for different positions P of the vehicle 401 on the roadway 403. In graph B, therefore, for the position P of the vehicle 401, only five of the six features 413 on the left edge of the roadway 403 can be detected by the environment sensors 405 of the vehicle 401.According to one embodiment, characteristics of the features can be determined to calculate the availability values ​​of the various features 413. The characteristics can include, for example, the type of feature 413, the number of features 413 in the area of ​​the roadway 403 to be traveled, spatial extents or sizes of the features 413, or reflectivity of the features 413. To determine detectability, these characteristics of the features 413 can be taken into account by calculating corresponding detectability-corrected characteristics. In the example of the number of features 413 as a characteristic, the detectability-corrected number of features 413 can be calculated as the number of features 413 that can be detected by the environment sensors 405 for the respective sensor type or directional characteristic 410 of the environment sensors 405 or for the calculated predicted traffic volume 409.

[0056] The availability value of the individual features 413 calculated in this way for a selected lane 403 can thus be specified, for example, as a percentage of the number of features 413 of the lane 403. The corresponding availability information can thus indicate that, for a respective lane 403, a certain percentage of the features 413 of the feature map 411 are available for locating the vehicle 401.

[0057] For a corresponding route planning of a vehicle 401, the vehicle 401 or the driver can thus be provided with a route that has the highest availability of the features 413 of the calculated route based on calculated availability information of the various lanes 403 of the respective route, in addition to a calculated route with the shortest travel time.

[0058] The corresponding route data 421 can be provided in the respective vehicle 401 by the computing unit 415 configured as an external server unit. The external server unit 415 can thus be configured to calculate the availability of the features 413 or objects arranged on the respective lanes for vehicles 401 and corresponding routes or lanes 403 to be traveled, and to determine corresponding routes with maximized availability of the features 413. This information can be provided to the vehicles 401 communicating with the external server unit 415.

[0059] To calculate a route, according to the invention, a data connection, for example in the form of a wireless data connection, can be established between a vehicle 401 and the computing unit 415 designed as an external server unit. In this case, the sensor information data 407 of the environmental sensors 405 of the vehicle 401 can be transmitted to the computing unit 415. In addition to the data 409 regarding the average traffic volume, the computing unit 415 or the external server unit can calculate a corresponding route with maximum availability according to the inventive method for availability prediction and the method for route planning and provide it to the vehicle 401 via wireless communication.

[0060] Fig. 2 shows a flowchart of a method 100 for availability prediction of a feature-based localization of a vehicle 401 according to one embodiment.

[0061] According to the invention, in a first method step 101, map data 412 of a feature map 411 of a road network are received, wherein the feature map 411 comprises feature information of a plurality of features 413 of a plurality of roadways 403 that can be traveled by a vehicle 401, and wherein the features 413 are configured to be detected by a vehicle 401 traveling on the roadway 403 via sensor data 406 from environmental sensors 405 of the vehicle 401 and to be used to localize the vehicle 401. In a further method step 103, features 413 for at least one roadway 403 to be traveled by the vehicle 401 are determined based on the feature information of the feature map 411.

[0062] In a further method step 105, an availability value of the features 413 of the at least one roadway 403 to be traveled by the vehicle 401 is calculated with reference to an availability criterion, wherein the availability value indicates a quantitative measure of a usability of the features 413 for locating the vehicle 401 when the vehicle 401 travels the roadway 403.

[0063] For this purpose, in the embodiment shown, a sensor type and / or a directional characteristic 410 of the environmental sensors 405 of the vehicle 401 are determined in a further method step 109.

[0064] Furthermore, in a further method step 113, a traffic volume is determined.

[0065] For this purpose, in a method step 115, an average number and / or an average speed of vehicles on the roadway 403 to be traveled by the vehicle is determined.

[0066] In a further method step 117, an average coverage 414 of the features 413 by the vehicles of the traffic volume is calculated based on the average number or average speed.

[0067] Furthermore, in a method step 119, characteristics of the features 413 are determined. The characteristics may include, for example, the number, properties, or spatial extents of the features.

[0068] In a method step 111, based on the sensor type or

[0069] Directional characteristic and / or based on the calculated

[0070] Traffic volume or average coverage, the detectability of the individual features by the environment sensors 405 of the vehicle 401 is calculated.

[0071] In a further method step, at least one characteristic of the features 413 is compared with the detectability-corrected characteristic calculated in method step 119. This comparison can provide an availability value for the various characteristics in the form of a quantitative numerical value. This can be specified, for example, as a percentage.

[0072] In a further method step 107, availability information based on the availability value calculated in method step 105 is output with respect to the features 413 of the feature map 411 for the roadway 403 to be traveled by the vehicle 401.

[0073] Fig. 3 shows a flowchart of a method 200 for route planning for an autonomously controllable vehicle 401 according to one embodiment.

[0074] According to the invention, in a method step 201, the method 100 for availability prediction of a feature-based localization of a vehicle 401 is first executed on a plurality of possible routes of the vehicle 401 between a predetermined starting point and a predetermined end point.

[0075] In a further method step 203, a route 421 of the vehicle 401 with a highest predicted availability of a feature-based localization of the vehicle 401 is determined based on the executed availability prediction.

[0076] According to one embodiment, controlling the vehicle 301 can comprise executing maneuver planning. In particular, driving maneuvers can be planned and executed by the vehicle 301 that increase the availability of the features of the feature map for the roadway to be traveled by the vehicle 301. For example, such a driving maneuver can comprise changing to a specific lane of the roadway to be traveled, wherein an increased availability of the feature is to be expected for the specific lane. Alternatively, the controlling can comprise other driving maneuvers that can increase the availability of the features, for example a reduction in the planned speed or activating the vehicle headlights in order to be able to better view the objects arranged along the lane to be traveled using the vehicle's environmental sensors.

[0077] Fig. 4 shows a flowchart of a method 300 for controlling a vehicle 401 according to one embodiment.

[0078] According to the invention, in a method step 301, the method 200 for route planning for an autonomously controllable vehicle is first carried out.

[0079] In a further method step 303, the vehicle 401 is controlled based on the selected route.

[0080] FIG 3 shows a schematic representation of a computer program product 500, comprising instructions which, when the program is executed by a computing unit, cause the computing unit to execute the method 100 for predicting the availability of a feature-based localization of a vehicle 401 and / or the method for route planning for an autonomously controllable vehicle 401 and / or the method for controlling a vehicle 401.

[0081] In the embodiment shown, the computer program product 500 is stored on a storage medium 501. The storage medium 501 can be any storage medium known from the prior art.

Claims

Claims 1. Method (100) for predicting the availability of a feature-based localization of a vehicle (401), comprising: Receiving (101) map data (412) of a feature map (411) of a road network, wherein the feature map (411) comprises feature information of a plurality of features (413) of a plurality of carriageways (403) traversable by a vehicle (401), and wherein the features (413) are configured to be detected by a vehicle (401) traversing the carriageway (403) via sensor data (406) from environmental sensors (405) of the vehicle (401) and to be used for localizing the vehicle; Determining (103) features (413) for at least one carriageway (403) to be traveled on by the vehicle (401) based on the feature information of the feature map (411); Calculating (105) an availability value of the features (413) of the at least one carriageway (403) to be traversed by the vehicle (401) with respect to an availability criterion, wherein the availability value provides a quantitative measure of the usability of the features (413) for the localization of the vehicle (401) when traversing the carriageway (403) by the vehicle (401); and outputting (107) availability information comprising the availability value with respect to the features (413) of the carriageway (403) to be traversed.

2. Method (100) according to claim 1, wherein the availability criterion comprises the detectability of the features (413) by the environmental sensors (405) of the vehicle (401) based on a sensor type and / or a directional characteristic (410) of the environmental sensors (405) of the vehicle (401), and wherein the calculation (105) of the availability value comprises: Determining (109) a sensor type and / or a directional characteristic (410) of at least one environmental sensor (405) of the vehicle (401); and determining (111) the detectability of the features (413) of the roadway (403) to be traveled on by the vehicle (401) for the sensor type and / or the directional characteristic (410) of the at least one environmental sensor (405). Method (100) according to claim 1 or 2, wherein the availability criterion comprises detectability of the features by the environmental sensors of the vehicle taking into account traffic volume on the roadway to be traveled on, and wherein calculating (105) the availability value comprises: Determine (113) the volume of traffic on the carriageway to be used by the vehicle; and Determining (111) the detectability of the features of the roadway to be traveled on by the vehicle by calculating the coverage of the roadway features by the traffic volume on the roadway. Method (100) according to claim 3, wherein determining (113) the traffic volume (409) comprises: Determining (115) an average number and / or an average speed of vehicles per unit of time on the roadway to be traveled by the vehicle; and wherein determining (111) includes the detectability: Calculating (117) an average coverage (414) of the feature (413) of the roadway based on the average number and / or speed of vehicles per unit of time. Method (100) according to any one of claims 2 to 4, wherein the calculation (105) of the availability value comprises: Determining (119) at least one characteristic of the features (413), wherein the at least one characteristic includes: the type of feature, the number of features, the number of features of a particular type or category; the recognizability of the features, the extent and / or size of the features, the reflectivity of the features (413); and wherein the determination (111) of the detectability is carried out with respect to at least one specific characteristic of the features and a detectability-corrected characteristic is calculated. Method (100) according to claim 5, wherein the calculation (105) of the availability value comprises: Comparing (121) the at least one characteristic of the features (413) of the feature map (411) with the corresponding detectability-corrected characteristic. Method (100) according to any one of claims 2 to 6, wherein the determination (111) of the detectability is carried out for at least one previously known position (P) of the vehicle (401) on the roadway (403) or for different previously known positions of the vehicle (401) on different lanes (404) of the roadway (403).Method (100) according to any one of claims 2 to 7, wherein the determination (111) of detectability is performed by a suitably trained artificial intelligence (419), wherein the artificial intelligence (419) is trained on environmental sensor data (406) from environmental sensors (403) of at least one vehicle (401), and wherein the environmental sensor data (406) were recorded during a plurality of journeys of the vehicle (401) along the at least one roadway (403) during a plurality of different traffic conditions (409). Method (100) according to any one of the preceding claims, wherein the features (413) are characteristic objects arranged at an edge of the roadway (403) and include: buildings, traffic signs, road markings. Method (200) for route planning for an autonomously controlled vehicle (401) comprising: Execution (201) of the procedure (100) for the availability prediction of a feature-based localization of a vehicle (401) according to one of the The preceding claims 1 to 9 relate to a plurality of possible routes of the vehicle (401) between a predetermined starting point and a predetermined endpoint; and Determining (203) a route (421) of the vehicle (401) with the highest predicted availability of a feature-based localization of the vehicle (401) based on the executed availability prediction. Method (300) for controlling a vehicle, comprising: executing (301) the route planning method (200) for an autonomously controllable vehicle according to claim 10; and Steering (303) of the vehicle (401) based on the selected route. Method (300) according to claim 11, wherein steering the vehicle comprises executing a maneuver planning, the maneuver planning comprising a driving maneuver of the vehicle that enables optimal availability of the features of the feature map. Computing unit (415, 417) configured to execute the method (100) for predicting the availability of a feature-based localization of a vehicle according to any one of claims 1 to 9 and / or the method (200) for route planning for an autonomously steerable vehicle according to claim 10 and / or the method (300) for steering a vehicle according to claim 11 or 12.Computer program product (500) comprising instructions which, when the program is executed by a data processing unit, cause the data processing unit to execute the method (100) for predicting the availability of a feature-based localization of a vehicle according to any one of the preceding claims 1 to 9 and / or the method (200) for route planning for an autonomously controllable vehicle according to claim 10 and / or the method (300) for controlling a vehicle according to claim 11 or 12.

Images