The invention relates to a method for localizing a vehicle on a road in a lateral direction with respect to the road, wherein sensor data are received from at least one environmental sensor of the vehicle, and wherein the sensor data relate to at least part of the road. Further, the sensor data are analyzed and at least one road feature is determined on the basis of the sensor data. Moreover, the vehicle is localized in the lateral direction with respect to the road in dependency of the determined at least one road feature. The invention further relates to a computer program and a driver assistance system.

Vehicle localization systems aim to find the position of the vehicle on the road. Usually, the position is divided into lateral and longitudinal components. Localization systems can be roughly divided into two categories, namely one for localization using exteroceptive sensors, which usually involves the usage of a perception algorithm, and one for localization using proprioceptive sensors, such as odometry. Common localization algorithms that use perception focus on different road features such as lane markings, traffic signs, and so on. Such road features are usually detected by means of a camera. However, during bad weather or bad lighting conditions or in case the lane markings are degraded, the perception from the camera has bad quality. Thus, localization is not possible or at least not with sufficient high quality.

US 8478 493 B2 and US 8 195 342 B2 describe that an operating environment around the vehicle is identified, and sensor data are selected from a set of sensors. A dynamic condition is identified using a plurality of different types of sensors on the vehicle. Further, a knowledge base can be used, which comprises an online knowledge base that dynamically provides information to a machine-controlled process which enables adjustment to sensor data processing and site-specific sensor accuracy calculations, an a priori knowledge base that contains static information about the operating environment of a vehicle, and a learned knowledge base that contains knowledge learned as the vehicle spends time in a specific work area. Further, environmental sensors can be selected according to a stored table and based on identified operating conditions. The table indicates the sensor quality in different operating conditions. Excellent quality can be achieved by a lidar sensor at a height of eight feet in every operating condition, even in early fall or winter. In contrast thereto, visible camera images of a curb or street edge or visible camera images of a street crown have good to poor quality or are even unusable in early fall or winter.

However, a lidar at a position of eight feet height can only be used on a harvester or a similar high vehicle but not on e.g. normal passenger cars.

It is an object of the present invention to provide a method, a computer program and a driver assistance system, which allow for a very accurate and reliable localization of a vehicle in a lateral direction with regard to a road.

This object is solved by a method, a computer program and a driver assistance system with the features according to the independent claims. Advantageous embodiments of the invention are presented in the dependent claims, the description and the drawings.

According to the invention, a method for localizing a vehicle on a road in a lateral direction with respect to the road is provided. According to this method, sensor data from at least one environmental sensor of the vehicle are received, wherein the sensor data relate to at least part of the road. Moreover, the sensor data are analyzed and at least one road feature is determined on the basis of the analysis of the sensor data. Moreover, the vehicle is localized in the lateral direction with respect to the road in dependency of the determined at least one road feature. By analyzing the sensor data, a lateral ground surface profile of a ground, which the road is part of, is determined and the at least one road feature is determined in dependency of the determined lateral ground surface profile.

The invention is based on the finding that a road typically comprises a characteristic, geometric lateral surface profile, which can be used to identify the parts of the lateral ground surface profile which belong to the road and further allow to identify the borders of the road and/or even the middle of the road and thereby allow for a localization of the vehicle in the lateral direction with respect to the road based on such detected road features, which may then serve as landmarks. In other words, geometrical characteristics of the road surface profile in the lateral direction with respect to the road are used as road features, which can define landmarks, instead of or additionally to the use of lane markings or traffic signs. The road features can define landmarks and the position of the vehicle, at least the lateral component thereof, can be determined relative to these one or more landmarks. Thereby these geometrical characteristics of a typical road are the following: typical roads are not horizontal in the lateral direction but are presenting either a road crown or an inclination regarding the horizontal axes, namely the axis in the lateral direction. Thus, in the first case, the road camber profile is such that the road height decreases from the center of the road to both lateral borders and in the second case, the camber profile is configured such that road height decreases from either the left border to the right border or from the right border to the left border. Thus, by means of the lateral ground surface profile, the middle of the road in the lateral direction and/or the borders of the road can be determined directly and very accurately in the first case, and at least one of the borders of the road or both borders can be determined very accurately in the second case, and thus, also a localization of the vehicle in lateral direction can be performed very accurately and with high reliability, either with respect to the middle of the road, which is later also called center of the road in the lateral direction, of with respect to at least one border of the road. Another great advantage of the invention is, that the visibility of lane markings or the like is not required. Thus, a lateral localization of the vehicle with respect to the road is even possible in case there are no lane markings at all. Further, this geometrical analysis of the lateral ground surface profile can also be used additionally to known methods for vehicle localization. This allows for an even more accurate and reliable localization in even more different situations. For example it is possible to compare the information obtained based on the lateral ground surface profile with the information obtained by other methods to better estimate the localization quality. This allows to improve the quality and the availability of the position estimation of the vehicle. Thus, combining this additional information with other sources, namely information derived from another method for localizing a vehicle, can be used to increase the robustness of the localization system, especially as errors from 3D-perception, which is preferably used for providing the sensor data, in the course of the invention, are likely to have low correlation with errors from 2D-perception by cameras, which are usually used for a conventional localization of the vehicle, e.g. based on the lane markings as landmarks. The invention also allows the localization system to operate even when the perception from cameras has bad quality acquisitions, such as during bad weather or lighting conditions, or when the landmarks, like lane markings, traffic signs and so on used by the other, conventional methods are degraded. Further, also the combination with other information sources are possible, like a map and/or GPS (Global Positioning Service). To conclude, the localization of the vehicle in the lateral direction with regard to the road can be performed much more accurate and reliable when considering the lateral gradient of the road or camber pattern of the road for the vehicle localization. The method according to the invention and its embodiments can be executed by a localization system, already mentioned above. This localization system is also called driver assistance system for localizing a vehicle in the course of the invention.

The at least one environmental sensor is a sensor of the vehicle is a sensor that is capable of perceiving at least part of the surrounding of the vehicle and especially at least part of the road, on which the vehicle is currently positioned and/or driving. It is preferred, that the at least one environmental sensor is a 3D-information sensor, which is configured to provide 3D-information about the surrounding of the vehicle. The at least one environmental sensor can be configured as a stereo camera pair, a lidar (light determining and ranging) and/or radar, and so on. This has the advantage that 3D- topographical information can be provided by such an environmental sensor, whereas conventionally only image information coming from cameras is used. Such 3D-information sensors allow to detect the geometry of the road surface or generally the ground surface and therefrom derive the lateral ground surface profile. Moreover, the at least one environmental sensor can be positioned such on the vehicle, that the field of view of the at least one environmental sensor covers at least part of the environment of the vehicle in front of the vehicle and/or behind the vehicle. The sensor data can also be provided from more than only one environmental sensor.

Generally, in the course of the invention, the lateral direction is defined as being perpendicular to a local, longitudinal direction of the road, and perpendicular to a vertical direction, which can be defined by the direction of gravity. Further, when the vehicle is traveling on the road, then the longitudinal axis of the vehicle can be assumed to be parallel to the longitudinal direction of the road, at least approximately. Moreover, the orientation of the sensor coordinate system of the at least one environmental sensor with respect to the vehicle coordinate system is known. Further, when it is assumed that the longitudinal axis of the vehicle is parallel to the longitudinal direction of the road, then also the orientation of the vehicle coordinate system and the sensor coordinate system is known. Otherwise, the orientation of the vehicle coordinate system or sensor coordinate system with respect to the longitudinal direction of the road can be determined on the basis of two or more measurements or determinations of e.g. the road crown position or position of road edge points. For example, the connection line between two determined crown positions at different distances can be assumed parallel to the longitudinal direction of the road. The same applies for edge points of the road at the roadside. Moreover, the lateral ground surface profile of the ground is a profile in a direction not parallel to the longitudinal direction of the road. The lateral ground surface profile can be determined in the lateral direction with respect to the road or in a direction comprising a certain angle with respect to the longitudinal direction of the road, which is e.g. the case when the sensor coordinate system comprises a certain angle with the longitudinal direction of the road. As the orientation of the sensor coordinate system with respect to the road can easily be determined, as described above, the ground surface profile can also be easily determined in the lateral direction of the road.

Localizing the vehicle in the lateral direction with respect to the road means that the position of the vehicle in the lateral direction with respect to the road is determined. This means that a component of the position of the vehicle is determined, which is oriented parallel to the lateral direction of the road. The lateral position of the vehicle can be determined as comprising e.g. a certain distance from the middle of the road and/or from a border of the road.

Furthermore, the lateral localization of the vehicle can be performed repeatedly. This allows for a continuous lateral localization of the vehicle during the vehicle is driving on the road. Furthermore, this also allows to detect changes of the lateral surface profile, and based on such detected changes, further information can be derived, like the presence of intersections, which is explained later.

According to an advantageous embodiment of the invention, by analyzing the sensor data a lateral slope profile of the ground is determined and in dependency of the determined lateral slope profile the at least one road feature is determined. As explained above, a typical road comprises a characteristic gradient profile in the lateral direction. Thus, it is very advantageous to analyze the lateral slope profile of the ground. Based on such a slope profile, above mentioned road feature and landmarks can be detected very easily.

Furthermore the slope profile can be determined based on the determined lateral ground surface profile, e.g. by taking the first derivative. Thus, in some advantageous embodiment of the invention, the ground surface profile is determined as the spatial progression of the ground surface in the lateral direction with respect to at least one certain cross-section, especially parallel to the lateral and vertical direction, and the slope profile is determined as the spatial change of the ground surface profile in the lateral direction. The spatial change of the ground surface profile thus is the first spatial derivative of the ground surface profile. The ground surface profile thus is the spatial progression of the height of the ground in the lateral direction with respect to some arbitrary defined horizontal plane, which defines a height of zero.

According to another advantageous embodiment of the invention, based on the lateral ground surface profile the sensor data relating to the road are identified and at least one specific point of the road is identified, which is at least one edge point of at least one lateral edge of the road and/or a lateral center point of the road, and a lateral position of the vehicle is determined relative to the identified specific road point. The lateral edge of the road can also be called road border. As explained above, it is very easy based on the known geometric characteristics of a road to determine the position of lateral edge points, or the middle of the road based on the lateral ground surface profile and the therefrom derived lateral surface road profile and the lateral slope profile of the road. Then, advantageously, the position of the vehicle can easily be determined relative to such identified specific road points. Thus, such a road point can also be seen as being a road feature or at least as being determined based on a determined road feature. The determination of the position of the vehicle with respect to such a specific road point, especially in the lateral direction allows for many advantageous applications, like autonomously guiding the vehicle or warning a driver when he is about to accidentally leave the lane or cross middle of the street.

According to another very advantageous embodiment of the invention, as the at least one road feature a road crown is determined and a lateral center of the road is determined to be at the position of the determined road crown. In case the road is configured such that it comprises a road crown, this road crown can easily be detected based on the lateral slope profile of the road. Thus, in case such a road crown is detected, also the center of the road in the lateral direction is known, because a road crown is always positioned at the center of the road in the lateral direction. The great advantage of detecting the road crown is that this works for nearly every road, as nearly every road comprises such a road crown, and further, the position of the vehicle with respect to the center of the road can be determined very accurately. This not even requires the determination or even the presence of lane markings.

According to another very advantageous embodiment of the invention, the localizing of the vehicle in lateral direction with respect to the road comprises determining the lateral position of the vehicle relative to the position of the lateral center of the road. As already explained, when the road crown of the road is detected, the vehicle position can be determined with respect to this detected road crown. Further, in case the road comprises two lanes, those lanes are typically also separated by the road crown. So, when the vehicle position can be determined relative to such detected road crown or center of the road in the lateral direction, also the position of the vehicle with respect to the current driving lane can be determined. Generally, the number of the lanes of the road can for example be determined based on the determined width of the road, which can also easily be determined based on the lateral ground surface profile, and by assuming some standard width range for a driving lane, or the number of lane of the current road can be derived from other information sources like a map, which can be stored in the localization system of the vehicle or be received from some source external of the vehicle.

According to another embodiment of the invention, it is determined based on the slope profile whether the slope changes its sign, especially on a defined macroscopic scale, and in case it is determined that the slope changes its sign, the lateral position at which the sign change takes place is defined as the position of the road crown. In case the road comprises a road crown, the height of the road from left to right first increases and then at the position of the road crown starts to increase. From the left border to the road crown, the road comprises a positive slope, and from the road crown to the right border of the road, the road or road surface comprises a negative slope. Thus, the position of the road crown can be detected by detecting this slope change. Of course, depending on the roughness of the road surface, also further slope changes may exist on a microscopic length scale or mesoscopic length scale, like on a scale of millimeters or centimeters. The slope change, which indicates the presence of the road crown, occurs at larger length scales, here called macroscopic length scale. For detecting the slope change indicative of the presence of the road crown, suitable filters, averaging methods, standard deviation, variance, etc. can be used. For example, the slope values of the slope profile can simply be averaged on a length scale of some centimeters.

According to another advantageous embodiment of the invention, based on the ground surface profile, especially based on the slope profile, regions of the profile are classified in dependency of their roughness, and especially regions of the profile are classified as belonging to the road or belonging not to the road, especially to a sidewalk or roadside banquet, in dependency of their roughness. The roughness can be measured as noise in the ground surface profile and/or slope profile. Typically, regions adjacent to a road comprise a surface which is rougher as the surface of the road. In case, there is sidewalk adjacent to the road, there is usually a curb between the road and the sidewalk, which also can easily be detected on the basis of the slope profile, or the road and the sidewalk are separated by some ground portion, which also typically is very rough. Thus, regions of the ground surface profiles belonging to the road and not belonging to the road can easily be determined based on the roughness of the respective regions.

According to another advantageous embodiment of the invention, the lateral slope profile of the ground is determined for several cross-sections, especially at different distances from the environmental sensor at the same time and/or subsequently. In case, a lidar is used to detect the lateral slope profile, the lidar can be capable of scanning the environment at different distances at the same time. More specifically, such a lidar can be configured to provide several layers of scanning lines at the same time. The layers differ from each other by a certain vertical angle. Nevertheless, to provide the lateral ground surface profile, a lidar providing just a single scanning line could be sufficient as well. Information pertaining to different distances can then be acquired timely subsequently, e.g. in subsequent time steps when the vehicle is moving forward. Determining the lateral slope profile at different distances allows e.g. to determine the longitudinal direction of the road and/or to detect changes of the slope profile of the road being indicative of e.g. intersections. Also, this allows for a continuous localization of the vehicle in lateral direction.

According to another advantageous embodiment of the invention, at least one driver assistance system, especially lane departure warning system, an automatic cross and/or guidance system and/or an autonomous driving system, is controlled in dependency of the determined lateral position of the vehicle. Therefore, the determination of the lateral position of the vehicle can be used for a variety of advantageous applications. Especially for autonomous driving, a very accurate and reliable determination of the lateral position of the vehicle on the road is very important. Thus, the invention and its embodiments provide great advantages especially with regard to such autonomous driving systems or other advanced driver assistance systems.

According to another advantageous embodiment of the invention, the sensor data are provided by a lidar and/or a radar and or a stereo camera. As already explained, these sensors allow the system to provide 3D-information. This 3D-information can again be used to detect the geometry of the road and to determine the lateral road profile.

According to another advantageous embodiment of the invention, a position of the vehicle in a map, which is a topographic map or a map that contains at least lateral profile data of roads including the road on which the vehicle is currently positioned, is determined based on the determined lateral ground surface profile by comparing at least the determined lateral ground surface profile to the stored map data, especially for a certain region around the current position of the vehicle, which, for example, can be determined by GPS. This advantageously allows for a relocation of the vehicle on a map. Typically, by means of GPS, only a rough estimation of the position of the vehicle can be provided. By using a map, which contains more specific information of the environment, like a topographic map, which comprises the 3D-position information of map elements like roads, it is possible to provide a much more accurate determination of the position of the vehicle in the map by comparing the current detected lateral road profile with those provided as part of the map data and relating to the part of the road, on which the vehicle is currently driving. This part of the road can be determined based on GPS information, which is accurate enough for rough position estimate of the vehicle. Further, the lateral road profile can be represented by very few data. For example, in case the road comprises a road crown, the height of the road crown and the height of the left and right edge point would be sufficient to represent the lateral road profile at a certain longitudinal positon of the road. In case the road does not comprise a road crown, the height of the left and right edge point would be sufficient to represent the lateral road profile at a certain longitudinal positon of the road.

Moreover, the determined lateral ground surface profile or road surface profile and especially also the slope profile and detected road features can be used for a variety of further advantageous applications, which are mentioned in the following: for example, based on the determined ground surface profile or slope profile a free space, especially around the vehicle or at least in part of the surrounding of the vehicle, is determined. The free space describes a space, in which no objects are present, with which the vehicle can collide. Knowing the free space is very important e.g. for path planning in the course of autonomous driving. This method can thus also be used to detect the free space around the vehicle by detecting the limit of the road, e.g. the sidewalk. This can ensure that all vehicle maneuvers are always safely performed in this free space.

By using a topographic map of roads, that matches with other localization systems, such as GPS, it is possible to accurately find the lateral position of the vehicle. In addition, by using displacement information, for example odometry, and variations in road profile due to intersections, it is also possible to find the longitudinal position of the vehicle, for example, with respect to specific points, like intersections. Odometric data can be provided by odometric sensors of the vehicle. Using odometry the vehicle can determine and track its own position starting from a known position and using current speed and acceleration information, steering information, and so on. Moreover, the ground surface profile and/or slope profile can be determined repeatedly and based on the repeatedly determined surface profiles and/or slope profiles a street intersection or road junction is detected, when a predefined significant change of the road slope profile is detected. So, advantageously, the lateral ground surface profile or slope profile can also be used to identify intersection areas. Indeed, the profile in these areas changes significantly, for example, there is often no pavement, no road crown and so on. These variations can also be detected on the basis of the determined ground surface profiles and thus be used to label these areas, e.g. as intersections in a map.

Moreover, based on the determined ground surface profile or slope profile a roughness of the road surface is determined and at least one component of the vehicle is controlled in dependency of the determined roughness. In case of strong roughness for example a warning can be output to a driver, or the vehicle can be controlled such that it is slowed down, i.e. the vehicle speed is decreased. Also, depending on the determined roughness, one or more dampers or shock absorbers of the vehicle can be automatically adjusted. Thus, it is also possible to adapt the behavior of the vehicle according to the type of profile automatically. So in the case of a noisy profile, the vehicle can, for example, slow down, adjust the dampers, warn the user, and so on.

The invention further relates to a computer program comprising instructions, which, when executed out by a computer system, cause the computer system to carry out a method according to the invention or its embodiments.

The invention also relates to a computer readable storage medium storing a computer program according to the invention or one of its embodiments.

The invention also relates to a driver assistance system for a vehicle for localizing a vehicle on a road in lateral direction with respect to the road, wherein the driver assistance system is configured to receive sensor data from at least one environmental sensor of the vehicle, wherein the sensor data relate to at least part of the road, to analyze the sensor data and to determine at least one road feature on the basis of the analysis of the sensor data, and to localize the vehicle in lateral direction with respect to the road in dependency of the determined at least one road feature. Furthermore, the driver assistance system is configured by analyzing the sensor data to determine a lateral ground surface profile of a ground, which the road is part of, and in dependency of the determined lateral ground surface profile to determine the at least one road feature. The advantages described with respect to the method according to the invention and its embodiments similarly apply for the driver assistance system according to the invention. Moreover, the embodiments mentioned in context of the method according to the invention allow further corresponding embodiments of the driver assistance system according to the invention.

The invention also relates to a vehicle comprising a driver assistance system according to the invention or one of its embodiments.

Further features of the invention are apparent from the claims, the figures and the description of figures. The features and feature combinations mentioned above in the description as well as the features and feature combinations mentioned below in the description of figures and/or shown in the figures alone are usable not only in the respectively specified combination, but also in other combinations without departing from the scope of the invention. Thus, implementations are also to be considered as encompassed and disclosed by the invention, which are not explicitly shown in the figures and explained, but arise from and can be generated by separated feature combinations from the explained implementations. Implementations and feature combinations are also to be considered as disclosed, which thus do not have all of the features of an originally formulated independent claim. Moreover, implementations and feature combinations are to be considered as disclosed, in particular by the implementations set out above, which extend beyond or deviate from the feature combinations set out in the relations of the claims.

In the following exemplary implementations of the invention are described. The figures show:

Fig. 1 a schematic illustration of the road camber of a road descending from left to right;

Fig. 2 a schematic illustration of a road camber of a road descending from a road mid-point to the borders of the road;

Fig. 3 a schematic illustration of the determination of the lateral ground surface profile according to an embodiment of the invention; and Fig. 4 a schematic illustration of the determined lateral ground surface profile and the lateral slope profile derived therefrom according to an embodiment of the invention.

Fig. 1 shows a schematic illustration of a vehicle 1 driving on a road 2a. The vehicle 1 may be configured as a vehicle 1 according to an embodiment of the invention and may comprise a driver assistance system 3 according to an embodiment of the invention. The driver assistance system 3 may comprise a control unit 4, especially an electronic control unit 4. The control unit 4 is configured to receive sensor data from at least one environmental sensor 5, which preferably is configured as a 3D-information sensor capable of capturing 3D-information from the environment 6 of the vehicle 1 , and especially at least part of the road 2a, in particular in front of the vehicle 1 . For example, the environmental sensor 5 can be configured as a lidar. The environmental sensor 5 can also be part of the driver assistance system 3. The sensor data provided by the environmental sensor 5 are transmitted to the control unit 4 and analyzed by the control unit 4. The control unit 4 is configured to perform a lateral localization of the vehicle 1 based on the analysis of the captured sensor data, as explained later.

In contrast to conventional systems, the driver assistance system 3 according to the invention or its embodiments considers the lateral gradient of the road 2a or camber pattern of the road 2a for the lateral vehicle localization. Lateral localization means that the lateral position component of the vehicle position is determined with respect to the road 2a. The lateral direction in this example is defined parallel to the x-direction of the illustrated coordinate system. The longitudinal direction of the road 2a in this case is oriented in or against the illustrated y-direction. The road 2a is illustrated in a cross- sectional view with regard to a cross-section perpendicular to the longitudinal direction y of the road 2a.

In road design, roads are not horizontal and are presenting either a road crown or an inclination regarding the horizontal axis, in this case the x-axis. Fig. 1 shows an example of a road camber 7, which comprises a camber profile that increases with regard to the road height from the left border 2b to the right border 2c of the road 2a. In particular, the road surface 2d decreases with regard to its height in the z-direction from a left border 2b of the road 2a to the right border 2c. Moreover, in this example, the road 2a comprises e.g. only one single lane. The camber 7, which defines the lateral road profile, can in this case be represented by the height of the left and right edge points of the lateral borders 2b, 2c of the road 2a.

Fig. 2 shows the vehicle 1 according to an embodiment of the invention driving on a road 2a comprising a different lateral camber profile. In this example, the camber 7 decreases from a midpoint which is also called road crown 8 with regard to its height in the z- direction to the lateral borders 2b, 2c of the road 2a. The camber 7, which defines the lateral road profile, can in this case be represented by the height of the road crown 8 and the height of the left and right edge points of the borders 2b, 2c.

The invention or its embodiments now allow to use the geometry of the road 2a, especially the camber 7, the road crown 8, sidewalks or borders 2b, 2c and so on, to laterally locate the vehicle 1 on the road 2a. This geometry can be detected by 3D- information sensors, like the environmental sensor 5, such as a stereo camera pair, a lidar sensor 5, a radar and so on. This 3D-information can then be processed, especially by the control unit 4, to extract salient features such as the road crown 8, the camber 7. The position of these features can then be analyzed to calculate the lateral position of the vehicle 1 on the road 2a.

Fig. 3 shows a schematic illustration of the determination of the lateral ground surface profile according to an embodiment of the invention. Fig. 3 again illustrates the vehicle 1 in a very schematic illustration. The vehicle 1 can be configured as described before. The vehicle 1 is driving on the road 2a, which in this case is illustrated in a perspective view. The environmental sensor 5 in this case is configured as a multi-layer scanner. The ground camber of the road 2a in this example is similar to that in the illustration of Fig. 2. The road 2a is limited in the lateral direction by the respective borders 2b, 2c, which are also called road edges 2b, 2c. The road edges 2b, 2c may comprise a curb, but need not. The ground of the vehicle 1 also extends in lateral direction beyond the road 2a on both sides of the road 2a. The ground is denoted by 9 in Fig. 3, wherein the road 2a, especially the road surface 2d, is also part of the ground 9.

In Fig. 3 also the impact patterns 10 of the lidar or radar scanner 5 are illustrated. In case the environmental sensor 5 is configured as a multi-layer lidar, the lidar emits a light pattern in different layers, which differ from each other by the angle of emittance in the vertical direction. The impact of this emitted light pattern on the road 2a provides the impact patterns 10. Each dotted line 10 represents light associated with a single layer of the multi-layer lidar. The light is then reflected on the road surface 2d and can be detected by the lidar 5. The received sensor data, which can be preprocessed by the environmental sensor 5, can be analyzed and further processed by the control unit 4 as described before. From the illustrated impact patterns 10, one can clearly see that the road 2a has two inclinations that can be determined by estimating a fitting plane for each side of the road 2a. The road crown 8 is the midpoint of the road 2a and can be estimated as the intersections of the left and right planes. The determination of the road crown 8 based on the received sensor data is now explained with regard to Fig. 4.

Fig. 4 shows a schematic illustration of the lateral ground surface profile 11 according to an embodiment of the invention. The ground surface profile 11 in the lateral direction x is the spatial progression of the ground surface 2b in the lateral direction x with respect to the cross-section perpendicular to the length direction y of the road 2a. To determine such a lateral ground surface profile 11 , one impact pattern 10 corresponding to one single layer of the scanner 5 would be sufficient. Based on such a single impact pattern 10, the height profile of the road 2a can be obtained, as illustrated in Fig. 4. Also illustrated in Fig. 4 is the lateral slope profile 12. The lateral slope profile 12 is the derivative signal of the lateral ground surface profile 11 . As one can see, the derivative signal, namely the lateral slope profile 12, is nearly constant in the regions 13, that are identified as road surface 2d. In particular, these regions 13 can be identified as road surface 2d based on the slope profile 12, especially as the regions 12a, 12b of the slope profile 12, which are nearly constant over large areas. Moreover, the slope profile 12 also changes its sign at a position X0. By identifying the sign change in the derivative signal between these two regions 12a and 12b with constant slope, the road crown 8 can be estimated, i.e. its position X0 can be determined. Therefore, one can identify the separation between the left and right lanes of the road 2a in the road crown 8, since it represents a change in the height signal derivative, namely the lateral slope profile 12. Additionally, the right border 2c of the road 2a is similarly detected due to the derivative noise 12c of the slope profile 12. Thus, an edge point Xr of the right border 2c of the road 2a can be determined, especially the position Xr in the lateral direction can be determined.

Such edge points Xr in the road crown 8 constitute road features which can be used to laterally locate the vehicle 1 . In other words, the position of the vehicle 1 can then be determined with respect to the position of the detected road features Xr, 8 with respect to the lateral direction. In the illustration of Fig. 4, the position of the vehicle 1 in the lateral direction is denoted by Xv and is located in the center of the illustrated coordinate system, which can also be the coordinate system associated with the environmental sensor 5. To conclude, the proposed solution, when implemented in a vehicle, allows a vehicle to be laterally located on the road. This can then be used to guide a vehicle or locate elements around the road. In order to perform this localization, the geometry of the road is used. Not considering the elevation of the road, the localization process could end up with errors in the localization result, which is the position of the vehicle, and the environment will be detected with an error. Due to the consideration of the lateral gradient of the road or camber pattern of the road for the vehicle localization, the vehicle localization in the lateral direction with respect to the road can be performed much more accurately and reliably.

The ideas as explained above are also applicable to roads having only one lane.