TECHNICAL FIELD

The invention relates to a method for determining lane boundaries, to a driving system that is configured to determine lane boundaries and to a vehicle with said driving system.

BACKGROUND

The determination of lane boundaries is frequently performed using environmental sensors such as cameras, radar and/or lidar. These environmental sensors detect landmarks such as road markings, curb stones, guide posts, guide rails and the like to determine the lane boundaries. However, in the absence of these landmarks, the determination of lane boundaries using the environmental sensors fails.

An alternative determination of lane boundaries is provided by high resolution maps that comprise said lane boundaries in combination with a highly accurate satellite navigation system. However, said high resolution maps are costly and processing said high resolution maps requires a high computational effort which is, again, costly.

SUMMARY

It is therefore an object of the present invention to provide a method for determining lane boundaries, a related driving system and a vehicle comprising such a driving system that overcome the above mentioned problems, in particular that allow the determination of lane boundaries when environmental sensors fail to detect landmarks, without the use of high resolution maps. The method may be a computer-implemented method. The object of the present invention is solved by the subject-matter of the independent claims, wherein further embodiments are incorporated in the dependent claims.

According to an aspect of the invention, a method for determining lane boundaries is provided. In this context, lane boundaries are boundaries of lanes of a road.

According to the method, environmental measurements are acquired from environmental sensors. Said acquisition of environmental measurements may be performed in regular intervals, e.g., once every second, once every 100 ms or once every 10 ms.

Based on said environmental measurements, a lane boundary detection is performed. In particular, the environmental measurements may comprise information on landmarks such as road markings, curb stones, guide posts, guide rails and the like, which indicate the lane boundaries.

It is then determined whether the lane boundary detection has failed. Such determination may be based on a confidence value of the lane boundary detection: if said confidence value is greater than a predetermined threshold, the lane boundary detection is deemed to have succeeded and if the confidence value is equal or less than the predetermined threshold, the lane boundary detection is deemed to have failed. If many landmarks are detected, a success of the lane boundary detection is likely. However, if only few landmarks or even no landmarks are detected, a failure of the lane boundary detection is likely.

In response to determining that the lane boundary detection has failed, a satellite navigation measurement from a satellite navigation system is acquired to obtain current position data. Said satellite navigation system may be any global navigation satellite system (GNSS), such as NAVSTAR GPS, GLONASS, Galileo or Beidou. Further, a standard navigation map is provided. Using the current position data obtained by the satellite navigation system, it is determined whether a current position corresponds to a specific situation in the standard navigation map.

Based on determining that the current position corresponds to a specific situation in the standard navigation map, road curvature data is obtained from the standard navigation map, i.e., the road curvatures are determined based on the data provided by the standard navigation map.

Further, vehicle location and orientation data is acquired. In this context, vehicle orientation refers to a yaw angle of the vehicle, i.e., to a rotation of the vehicle around the vertical axis.

Finally, the lane boundary is predicted based on the road curvature data, the vehicle location and orientation data and historical environmental measurements. In this context, the historical environmental measurements are environmental measurements in the past that lead to a successful lane boundary detection.

Hence, the lane boundary can be predicted even when the lane boundary detection based on the environmental measurements fails. Further, for this lane boundary prediction, a standard navigation map is sufficient and there is no need for a costly high precision map.

The prediction of the lane boundary based on the road curvature data, the vehicle location and orientation data and the historical environmental measurements is limited to the time while the current position corresponds to the specific situation: when the vehicle has left the specific situation, said prediction is no longer performed.

According to an embodiment which may be combined with any above-described embodiment or with any below described further embodiment, the environmental sensors comprise a camera, a radar and/or a lidar. In particular a combination of at least one camera and radar or lidar is very efficient at detecting landmarks and their distance to the vehicle. However, the distance to the vehicle may also be measured using stereo cameras or by tracking the detected landmarks across several frames. Some or all of the above mentioned sensors are already installed in current vehicles, hence no additional hardware is required.

According to an embodiment which may be combined with any above-described embodiment or with any below described further embodiment, the standard navigation map is defined by a navigation data standard, NDS, and/or an advanced driver assistance systems interface specification, ADASIS. Hence, the standard navigation map can be easily obtained and is relatively low cost.

The standard navigation map may be defined to be a navigation map with a resolution of between 1 m and 10 m, preferably between 2 m and 5 m. This resolution is sufficient for the method, yet it keeps the file size of the navigation map and the required computing resources low.

According to an embodiment which may be combined with any above-described embodiment or with any below described further embodiment, the specific situation is an intersection, a historic city area and/or a rural area. These specific situations are prone to having few landmarks that may be used to determine a lane boundary. The method allows to predict the lane boundaries even in these specific situations, which increases safety in these specific situations.

According to an embodiment which may be combined with any above-described embodiment or with any below described further embodiment, the vehicle location and orientation data comprises a yaw angle and/or a lateral position. In this context, the yaw angle is an angle of the vehicle around a vertical axis and the lateral position is the transverse deviation of the vehicle from the center of the lane or the center of the road. Both the yaw angle and the lateral position are easy to obtain and provide the most important information on vehicle location and orientation.

According to an embodiment which may be combined with any above-described embodiment or with any below described further embodiment, the vehicle location and orientation data is acquired from gyroscopic measurements from a gyroscopic sensor. In this context, the gyroscopic sensor may at least measure rotational accelerations around the vertical axis. Further, the gyroscopic sensor may measure linear accelerations, in particular in the horizontal plane. By integrating over these rotational and/or linear accelerations, the location and/or orientation of the vehicle compared to a previously known location and/or orientation may be determined. This yields a very accurate determination of the vehicle location and orientation.

Alternatively, or additionally, the vehicle location and orientation data is determined from an odometry accumulation. That is, the vehicle speed is integrated to provide a distance traveled by the vehicle, preferably in combination with a recorded steering of the vehicle. This determination of the vehicle location and orientation does not require an extra gyroscopic sensor and therefore reduces the cost of the vehicle.

Alternatively, or additionally, the vehicle location and orientation data is determined via the satellite navigation system. This determination of the vehicle location and orientation is reasonably accurate and provides absolute values of the vehicle location.

According to an embodiment which may be combined with any above-described embodiment or with any below described further embodiment, predicting the lane boundary comprises interpolating a map-based road curvature based on the road curvature data. That is, the rather coarse road curvature data is smoothened by interpolation, e.g., using polynomials or Bezier curves. This smoothened result is more accurate and true-to-life.

According to an embodiment which may be combined with any above-described embodiment or with any below described further embodiment, predicting the lane boundary comprises matching the map-based road curvature based on the road curvature data with a historical data-based road curvature based on the historical environmental measurements. Said matched road curvature may be called target curvature. A potential offset between the actual environmental measurements and the curvature based on the standard navigation map, which may be due to the lack of precision of the standard navigation map, is eliminated by said matching. The matching may depend on the quality of the map-based road curvature and the historical data-based road curvature, such that more weight is put on the road curvature with the higher quality. Finally, the lane boundaries with respect to the vehicle are then determined using the vehicle location and orientation data. Hence, a very accurate prediction of the lane boundaries is achieved, combining the accurate historical data-based road curvature with the map-based road curvature.

According to an embodiment which may be combined with any above-described embodiment or with any below described further embodiment, the method further comprises issuing an alert that the lane boundary cannot be determined, based on determining that the current position does not correspond to a specific situation in the standard navigation map. Hence, no attempt is made to predict the lane boundary when the environmental sensors fail to determine the lane boundaries and the vehicle is not in a specific situation, leading to enhanced safety.

According to an embodiment which may be combined with any above-described embodiment or with any below described further embodiment, the determined lane boundary is provided to a driver assistance system and/or to an autonomous driving system. In this context, a driver assistance system may be a lane keep assist system, an adaptive cruise control and the like. By providing the determined lane boundary to the driver assistance system and/or to the autonomous driving system, the operation of these systems may be enabled, i.e., the availability of these systems may be increased, and/or improved.

According to another aspect of the invention, a driving system is provided. The driving system comprises environmental sensors, a satellite navigation system and a computing unit and is configured to be operated according to the above description. The advantages and further embodiments correspond to those given in the above description.

According to an embodiment which may be combined with any above-described embodiment or with any below described further embodiment, the driving system is a driver assistance system and/or an autonomous driving system. In either case, the operation of the driving system is improved by the provided lane boundaries.

According to yet another aspect of the invention, a vehicle is provided. The vehicle comprises a driving system according to the above description. Hence, the advantages and further embodiments correspond to those given in the above description.

BRIEF DESCRIPTION OF THE DRAWINGS

These and other aspects of the invention will be apparent from and elucidated further with reference to the embodiments described by way of examples in the following description and with reference to the accompanying drawings, in which

Fig. 1 shows a flowchart of an embodiment of a method for determining lane boundaries;

Fig. 2 shows a flowchart of another embodiment of a method for determining lane boundaries; and

Figs. 3a and 3b show lane boundaries based on road curvature data and on historical environmental measurements.

In the figures, elements which correspond to elements already described may have the same reference numerals. Examples, embodiments or optional features, whether indicated as non-limiting or not, are not to be understood as limiting the invention as claimed.

DESCRIPTION OF EMBODIMENTS

Figure 1 shows a flowchart of an embodiment of a method 1 for determining lane boundaries.

According to the method 1 , environmental measurements 3 are acquired from environmental sensors 2. Said environmental sensors 2 may comprise a camera, a radar and/or a lidar. The environmental measurements 3 may comprise information on landmarks such as road markings, curb stones, guide posts, guide rails and the like. Based on the environmental measurements 3, a lane boundary detection 4 is performed. It is then determined 5 whether the lane boundary detection 4 has failed. In case that the lane boundary detection 4 has succeeded (not shown here), the lane boundaries are determined.

In response to determining 5 that the lane boundary detection 4 has failed, a satellite navigation measurement is acquired from a satellite navigation system 6 to obtain current position data 7. Further, a standard navigation map 8 is provided. Said standard navigation map 8 may be defined by a navigation data standard, NDS, and/or an advanced driver assistance systems interface specification, ADASIS.

It is then determined 9 whether the current position 7 corresponds to a specific situation, such as an intersection, a historic city area and/or a rural area, in the standard navigation map 8. Based on determining 9 that the current position 7 does not correspond to a specific situation in the standard navigation map 8 (not shown here), an alert may be issued that the lane boundary cannot be determined.

Based on determining 9 that the current position 7 corresponds to a specific situation in the standard navigation map 8, road curvature data 10 is obtained from the standard navigation map 8. Further, vehicle location and orientation data 12 is acquired, e.g., from motion sensors 11 of the vehicle. Said motion sensors 11 may be gyroscopic sensors and/or odometry sensors such as speedometers.

Then, based on the road curvature data 10, the vehicle location and orientation data 12 and historical environmental measurements 3, the lane boundary 13 is predicted. Said prediction of the lane boundary 13 may comprise matching a map-based road curvature based on the road curvature data 10 with a historical data-based road curvature based on the historical environmental measurements 3.

Hence, the lane boundary 13 can be predicted even when the lane boundary detection 4 based on the environmental measurements 3 fails. Further, for this lane boundary prediction, a standard navigation map 8 is sufficient and there is no need for a costly high precision map.

Figure 2 shows a flowchart of another embodiment of a method 1 for determining lane boundaries 13. Compared to the method 1 of Figure 1 , the determined lane boundaries 13 are provided to a driving system 14 such as a driver assistance system and/or an autonomous driving system. Hence, the driving system 14 may operate even when the standard lane boundary detection 4 fails and/or the operation of the driving system 14 may be improved in this case.

Figure 3a schematically shows historical data-based road curvature 15 (solid lines) based on the historical environmental measurements 3 as well as map-based road curvature 16 (dashed lines) for an intersection. Due to the low precision of the standard navigation map 8, there is an offset between the historical data-based road curvature 15 and the map-based road curvature 16.

By matching the historical data-based road curvature 15 and the map-based road curvature 16, as shown in Figure 3b, the offset disappears and the result are predicted lane boundaries 13.

Other variations to the disclosed embodiments can be understood and effected by those skilled in the art in practicing the claimed invention, from the study of the drawings, the disclosure, and the appended claims. In the claims, the word “comprising” does not exclude other elements or steps and the indefinite article “a” or “an” does not exclude a plurality. The mere fact that certain measures are recited in mutually different dependent claims does not indicate that a combination of these measures cannot be used to advantage. Any reference signs in the claims should not be construed as limiting the scope of the claims. List of reference signs

1 method

2 environmental sensor

3 environmental measurement

4 lane boundary detection

5 determination

6 satellite navigation system

7 current position data

8 standard navigation map

9 determination

10 road curvature data

11 motion sensor

12 vehicle location and orientation data

13 lane boundary

14 driving system

15 historical data-based road curvature

16 map-based road curvature