US-12624960-B2 - Device and method for determining map data on the basis of observations

Abstract

A device for determining map data with respect to a route portion, configured to acquire a first set of first observations during a first trip along the route portion and a second set of second observations during a second trip along the route portion, determine values of a measure of similarity for different possible observation pairs each consisting of one first observation and one second observation, determine, on the basis of the values of the measure of similarity for different possible observation pairs, an overall association with one or more one-to-one associations between, in each case, one first observation from the first set and one second observation from the second set. The device is also configured to determine the map data with respect to the route portion on the basis of the overall association.

Inventors

• Wolfgang Hempel
• Martin Liebner
• Manuel Luitz
• David Pannen
• Fabian Zeller

Assignees

• BAYERISCHE MOTOREN WERKE AKTIENGESELLSCHAFT

Dates

Publication Date

20260512

Application Date

20211022

Priority Date

20201028

Claims (18)

1. 1 . A device for determining map data with respect to a route portion, wherein the device comprises: at least one processor configured to: acquire a first set of first observations during a first trip along the route portion, and a second set of second observations during a second trip along the route portion; determine values of a measure of similarity for a plurality of different potential observation pairs, each consisting of a first observation from the first set and a second observation from the second set, wherein the measure of similarity for each potential observation pair of the plurality of different potential observation pairs comprised of a specified first observation and a specified second observation indicates a similarity between an arrangement of one or more adjacent observations from the first set relative to the specified first observation, and an arrangement of one or more adjacent observations from the second set relative to the specified second observation; wherein to determine the value of the measure of similarity for the at least one potential observation pair, the at least one processor is configured to: determine a set of first translation vectors from the specified first observation to the one or more adjacent observations in the first set, wherein the specified first observation and the one or more adjacent observations in the first set are all from the first set of first observations acquired during the first trip; determine a set of second translation vectors from the specified second observation to the one or more adjacent observations in the second set, wherein the specified second observation and the one or more adjacent observations in the second set are all from the second set of second observations acquired during the second trip; and determine the value of the measure of similarity for the at least one potential observation pair based on a similarity of a magnitude and/or direction of the set of first translation vectors and a magnitude and/or direction of the set of second translation vectors; determine, on a basis of the values of the measure of similarity for the plurality of different potential observation pairs, an overall association between the first set and the second set comprising one or more one-to-one associations between, for each one-to-one association, one particular first observation from the first set and one particular second observation from the second set; and determine map data with respect to the route portion on a basis of the overall association; wherein a vehicle performs an automated driving function based at least in part on the map data, wherein the automated driving function comprises automated control of at least one of a speed, a braking, or a steering function of the vehicle.
2. 2 . The device according to claim 1 , wherein the at least one processor is configured to: determine a vector similarity based on a similarity of the magnitudes and/or directions of vector pairs from a first translation vector and a second translation vector respectively; and determine the value of the measure of similarity for the potential observation pair on the basis of the vector similarity of the one or more vector pairs.
3. 3 . The device according to claim 1 , wherein: the first set comprises Q first observations and the second set comprises R second observations; and wherein the at least one processor is configured to: determine, for each of the Q first observations, a respective set of Q−1 first translation vectors with regard to the respective Q−1 adjacent observations from the first set; determine, for each of the R second observations, a respective set of R−1 second translation vectors with respect to the respective R−1 observations from the second set; determine, for a specified potential observation pair comprising the specified first observation and the specified second observation, a vector similarity between the set of first translation vectors for the specified first observation and the set of second translation vectors for the specified second observation based on the similarity of the magnitudes and/or angles of the set of first translation vectors and the set of second translation vectors; and determine the value of the measure of similarity for the specified potential observation pair on a basis of the vector similarity determined.
4. 4 . The device according to claim 1 , wherein the at least one processor is configured to: for the determination of a vector similarity for a specified potential observation pair, for each potential combination of a first translation vector from the set of first translation vectors and a second translation vector from the set of second translation vectors, determine an individual value of the vector similarity between a respective first translation vector and a respective second translation vector in each case; determine, on a basis of individual values for the vector similarity, a set of one or more one-to-one vector associations established between a respective first translation vector and a respective second translation vector, for which an overall value of vector similarity is increased or maximized; and determine the value of the measure of similarity for the specified potential observation pair on a basis of the overall value of vector similarity.
5. 5 . The device according to claim 1 , wherein the at least one processor is configured to: determine, for the one or more one-to-one associations from the overall association, one or more respective transformation vectors, by which the second observation in a one-to-one association is transposed to the first observation of the one-to-one association; determine, on a basis of the one or more respective transformation vectors for the one or more one-to-one associations from the overall association, a uniform transformation vector for the second set of second observations; transform, respectively, the one or more second observations from the second set of second observations by application of the uniform transformation vector to determine a set of transformed observations; determine, on a basis of the set of transformed observations, an adapted overall association; and determine the map data with respect to the route portion on the basis of the adapted overall association.
6. 6 . The device according to claim 5 , wherein the at least one processor is configured to: determine, using a clustering algorithm, a set of clusters on a basis of the first set of first observations and on a basis of the set of transformed observations; and determine the adapted overall association on the basis of the set of clusters.
7. 7 . The device according to claim 1 , wherein: the first set of first observations and the second set of second observations respectively comprise observations of landmarks, each of which represents a landmark type from a plurality of different landmark types; the plurality of landmark types comprises at least one landmark type which is unsuitable for the determination of the measure of similarity; and wherein the at least one processor is configured to: determine the values for the measure of similarity only for potential observation pairs of observations which do not include an inappropriate landmark type.
8. 8 . The device according to claim 7 , wherein the at least one processor is configured to: determine an adapted overall association, at least in part, by considering one or more observations from the first set of first observations and/or from the second set of second observations which include the landmark type which is unsuitable for the determination of the measure of similarity.
9. 9 . The device according to claim 1 , wherein the at least one processor is configured to: determine the map data with respect to the route portion using a Simultaneous Localization and Mapping (SLAM) method, on a basis of the overall association.
10. 10 . The device according to claim 1 , wherein: the map data indicates one or more landmarks, including a position of one or more landmarks, on the route portion.
11. 11 . The device according to claim 1 , wherein: an observation indicates a position and/or type of a landmark; and/or an observation is made on a basis of sensor data from at least one sensor of the vehicle comprising an environment sensor and/or a position sensor during a trip along the route portion.
12. 12 . The device according to claim 1 , wherein the at least one processor is configured to: determine, for a plurality of potential one-to-one associations, respectively, clearance information with respect to a spatial clearance between a first observation and a second observation; and determine the overall association on a basis of the clearance information.
13. 13 . The device according to claim 1 , wherein the at least one processor is configured to: determine, for a plurality of potential one-to-one associations, respectively, type information with respect to a first landmark type in the first observation and a second landmark type in the second observation, wherein the type information indicates a similarity of the first landmark type and the second landmark type; and determine the overall association on a basis of the type information.
14. 14 . A method for determining map data with respect to a route portion, the method comprising: acquiring a first set of first observations during a first trip along the route portion, and a second set of second observations during a second trip along the route portion; determining values of a measure of similarity for a plurality of different potential observation pairs, each consisting of a first observation from the first set and a second observation from the second set, wherein the measure of similarity for each potential observation pair of the plurality of different potential observation pairs comprised of a specified first observation and a specified second observation indicates a similarity between an arrangement of one or more adjacent observations from the first set relative to the specified first observation, and an arrangement of one or more adjacent observations from the second set relative to the specified second observation; wherein determining the value of the measure of similarity for the at least one potential observation pair, further comprises: determining a set of first translation vectors from the specified first observation to the one or more adjacent observations in the first set, wherein the specified first observation and the one or more adjacent observations in the first set are all from the first set of first observations acquired during the first trip; determining a set of second translation vectors from the specified second observation to the one or more adjacent observations in the second set, wherein the specified second observation and the one or more adjacent observations in the second set are all from the second set of second observations acquired during the second trip; and determining the value of the measure of similarity for the at least one potential observation pair based on a similarity of a magnitude and/or direction of the set of first translation vectors and a magnitude and/or direction of the set of second translation vectors; determining, on a basis of the values for the measure of similarity for the plurality of different potential observation pairs, an overall association between the first set and the second set comprising one or more one-to-one associations between, for each one-to-one association, one particular first observation from the first set and one particular second observation from the second set; determining map data with respect to the route portion, on a basis of the overall association; and performing, by a vehicle, an automated driving function based at least in part on the map data, wherein the automated driving function comprises automated control of at least one of a speed, a braking, or a steering function of the vehicle.
15. 15 . The method according to claim 14 , comprising: determining a vector similarity based on a similarity of the magnitudes and/or directions of vector pairs from a first translation vector and a second translation vector respectively; and determining the value of the measure of similarity for the potential observation pair on the basis of the vector similarity of the one or more vector pairs.
16. 16 . The method according to claim 14 , comprising: for the determination of a vector similarity for a specified potential observation pair, for each potential combination of a first translation vector from the set of first translation vectors and a second translation vector from the set of second translation vectors, determining an individual value of the vector similarity between a respective first translation vector and a respective second translation vector in each case; determining, on a basis of individual values for the vector similarity, a set of one or more one-to-one vector associations established between a respective first translation vector and a respective second translation vector, for which an overall value of vector similarity is increased or maximized; and determining the value of the measure of similarity for the specified potential observation pair on a basis of the overall value of vector similarity.
17. 17 . The method according to claim 14 , comprising: determining, for the one or more one-to-one associations from the overall association, one or more respective transformation vectors, by which the second observation in a one-to-one association is transposed to the first observation of the one-to-one association; determining, on a basis of the one or more respective transformation vectors for the one or more one-to-one associations from the overall association, a uniform transformation vector for the second set of second observations; transforming, respectively, the one or more second observations from the second set of second observations by application of the uniform transformation vector to determine a set of transformed observations; determining, on a basis of the set of transformed observations, an adapted overall association; and determining the map data with respect to the route portion on the basis of the adapted overall association.
18. 18 . The method according to claim 14 , wherein: the first set of first observations and the second set of second observations respectively comprise observations of landmarks, each of which represents a landmark type from a plurality of different landmark types; and the plurality of landmark types comprises at least one landmark type which is unsuitable for the determination of the measure of similarity; and wherein the method further comprises: determining the values for the measure of similarity only for potential observation pairs of observations which do not include an inappropriate landmark type.

Description

FIELD The invention relates to a device and a corresponding method for determining map data with respect to a roadway, carriageway or route portion, particularly with respect to a node point. Map data can be determined on the basis of observations of landmarks. BACKGROUND AND SUMMARY A vehicle can comprise one or more driving functions which support the guidance of the vehicle by the driver, particularly longitudinal guidance and/or lateral guidance. An exemplary driving function for supporting the longitudinal guidance of a vehicle is an Adaptive Cruise Control (ACC) function, which can be employed for the longitudinal guidance of the vehicle at a stipulated set speed or target speed and/or with a stipulated target clearance from the vehicle which is driving ahead of the vehicle. The driving function can also be employed at a signaling unit (particularly at a traffic light) situated at a traffic node point (for example, at an intersection), in order to execute automated longitudinal guidance, for example an automatic delay, at the signaling unit. Consideration of a signaling unit at a node point (wherein the signaling unit comprises one or more signal generators) can be executed by reference to map data, wherein map data comprises one or more mapping attributes with respect to a landmark which is to be considered, for example a signaling unit. The quality of the driving function is thus typically dependent upon the quality of map data available, and particularly upon the quality of available mapping attributes. The present document particularly addresses the technical object of the improvement of map data with respect to a route portion, particularly with respect to a node point having at least one signaling unit, in order to improve the convenience and/or safety of a driving function, particularly a driving function for automated longitudinal guidance at a signaling unit or node point. This object is fulfilled in accordance with the present disclosure. Advantageous embodiments are also described in the present disclosure. It should be observed that additional features of a patent claim which is dependent upon an independent patent claim, in the absence of the features of the independent patent claim or in combination with only a proportion of the features of the independent patent claim, can form a separate invention which is independent of the combination of all the features of the independent patent claim, and which can be the subject matter of an independent claim, a divisional application or a subsequent application. The same applies correspondingly to the technical instruction described in the description, which can form an invention which is independent of the features of the independent patent claims. According to one aspect, a device is described for determining map data with respect to a route portion, particularly with respect to a node point. Map data can thus be determined which identify attributes with respect to one or more landmarks in the route portion. Exemplary landmarks include a signal generator of a light signal installation, a lane marking, a traffic sign, a reflector post or pole, etc. Exemplary attributes of a landmark include the position of the landmark, the landmark type of the landmark (e.g. signal generator, traffic sign, post, etc.), etc. The device can be designed to acquire a first set of first observations during a first trip along the route portion, and a second set of second observations during a second trip along the route portion. In particular, the device can be designed to acquire a plurality of sets, each comprising one or more observations, for a corresponding plurality of trips along the route portion. An observation of a landmark can indicate the (measured or estimated) position and/or the (measured or estimated) type of the landmark. An observation can have been acquired on the basis of sensor data from at least one sensor, particularly an environment sensor and/or a position sensor, of a vehicle during a trip along the route portion. Thus, during different trips, respectively detected sets of one or more observations of landmarks can be determined. Sets of one or more observations can be delivered by one or more different vehicles, e.g. by transmission via a communication link. Different sets of observations can differ from each other, particularly on the grounds of offset errors between the different sets of observations. Moreover, a set of observations can potentially include multiple observations of the same landmark (e.g. on the grounds of intermittent coverage of the landmark during a trip along the route portion). As described hereinafter, by means of the measures described in the present document, multi-trip observations of the same landmark can be detected in a reliable manner. The device can be designed to employ the internal geometry of the route portion as a means of assigning observations from different sets of observations to one another, or for th