Method for Determining the Position of a Vehicle

Abstract

A computer implemented method for determining the position of a vehicle, wherein the method comprises: determining at least one scan comprising a plurality of detection points, wherein each detection point is evaluated from a signal received at the at least one sensor and representing a location in the vehicle environment; determining, from a database, a predefined map, wherein the map comprises a plurality of elements in a map environment, each of the elements representing a respective one of a plurality of static landmarks in the vehicle environment, and the map environment representing the vehicle environment; matching the plurality of detection points and the plurality of elements of the map; determining the position of the vehicle based on the matching; wherein the predefined map further comprises a spatial assignment of a plurality of parts of the map environment to the plurality of elements, and wherein the spatial assignment is used for the matching.

Claims

1. A method for determining a position of a vehicle equipped with a sensor system, wherein the sensor system includes at least one sensor for receiving electromagnetic radiation emitted from at least one emitter of the sensor system and reflected in a vehicle environment towards the at least one sensor, wherein the method comprises: determining at least one scan comprising a plurality of detection points (P), wherein each detection point is evaluated from the electromagnetic radiation received by the at least one sensor and representing a location in the vehicle environment; determining, from a database, a predefined map, wherein the predefined map comprises a plurality of elements in a map environment and a spatial assignment of a plurality of parts of the map environment to the plurality of elements, wherein: each of the elements representing a respective one of a plurality of static landmarks in the vehicle environment; the parts of the map environment are defined by spatial positions arranged as a mesh grid which comprise positions for any potential detection point; and at least a portion of the map environment representing the vehicle environment; matching, using the spatial assignment, the plurality of detection points (P) to the plurality of elements in the map environment, wherein the matching comprises: identifying, for a respective one of the plurality of detection points (P), a respective one of the plurality of parts of the map environment, the respective one of the plurality of parts of the map environment representing a part of the vehicle environment including the location represented by the respective one of the plurality of detection points; identifying a respective one of the plurality of elements being assigned to the identified respective one of the plurality of parts of the map environment; and assigning the identified respective one of the plurality of elements to the respective one of the plurality of detection points (P); and determining the position of the vehicle based on the matching.

2. The method according to claim 1, wherein the predefined map comprises a definition of a plurality of spatial segments, each of the spatial segments representing a respective one of the plurality of parts of the map environment and being assigned to a respective one of the plurality of elements.

3. The method according to claim 1, wherein the matching is carried out on the basis of identified respective ones of the plurality of elements and respective ones of the plurality of detection points (P) to which the identified respective ones of the plurality of elements have been assigned.

4. The method according to claim 1, the method further comprising: determining at least one confidence value for a respective one of the plurality of detection points on the basis of the spatial assignment, wherein: the at least one confidence value represents a probability that the respective one of the plurality of detection points is assigned to a respective one of the plurality of elements; and the matching is carried out on the basis of the at least one confidence value.

5. The method according to claim 4, wherein the at least one confidence value is determined on the basis of a Monte-Carlo simulation for the respective one of the plurality of detection points.

6. The method according to claim 5, wherein the at least one confidence value is transformed to an exponential weighting factor and wherein the matching comprises weighting by the exponential weighting factor.

7. The method according to claim 1, wherein the predefined map further comprises a plurality of distance values representing distances between the plurality of parts of the map environment being assigned to the plurality of elements, respectively, wherein the distance values are used for the matching.

8. The method according to claim 1, wherein the spatial assignment identifies, for a respective one of the plurality of parts of the map environment, a respective one of the plurality of elements having a minimum distance to the respective one of the plurality of parts of the map environment.

9. The method according to claim 1, wherein the matching further comprises determining a rigid transformation function by minimizing distances between transformed detection points and respective ones of the plurality of elements being assigned to the plurality of detection points by means of the spatial assignment, wherein the transformed detection points are respective ones of the plurality of detections points transformed by means of the rigid transformation function.

10. The method according to claim 1, wherein at least one element of the map comprises an extended geometrical object, in particular a line or a surface, preferably a straight line or a plane.

11. The method according to claim 1, wherein the matching further comprises particle filtering.

12. A system comprising: at least one emitter; and at least one sensor for receiving electromagnetic radiation emitted by the at least one emitter and reflected in a vehicle environment towards the at least one sensor; and at least one processor configured to determine a position of a vehicle equipped with the system by: determining at least one scan comprising a plurality of detection points (P), wherein each detection point is evaluated from the electromagnetic radiation received by the at least one sensor and representing a location in the vehicle environment and a spatial assignment of a plurality of parts of the map environment to the plurality of elements; determining, from a database, a predefined map, wherein the predefined map comprises a plurality of elements in a map environment, wherein: each of the elements represent a respective one of a plurality of static landmarks in the vehicle environment; the parts of the map environment are defined by spatial positions arranged as a mesh grid which comprise the positions for any potential detection point; and at least a portion of the map environment represents the vehicle environment; matching, using the spatial assignment, the plurality of detection points (P) and the plurality of elements, the matching comprises: identifying, for a respective one of the plurality of detection points (P), a respective one of the plurality of parts of the map environment, the respective one of the plurality of parts of the map environment representing a part of the vehicle environment including the location represented by the respective one of the plurality of detection points; identifying a respective one of the plurality of elements being assigned to the identified respective one of the plurality of parts of the map environment; and assigning the identified respective one of the plurality of elements to the respective one of the plurality of detection points (P); and determining the position of the vehicle based on the matching.

13. The system of claim 12, further comprising the vehicle.

14. A non-transitory computer-readable medium comprising instructions that, when executed, configure at least one processor to determine, using at least one emitter and at least one sensor for receiving electromagnetic radiation emitted by the at least one emitter and reflected in a vehicle environment towards the at least one sensor, a position of a vehicle by: determining at least one scan comprising a plurality of detection points (P), wherein each detection point is evaluated from the electromagnetic radiation received by the at least one sensor and representing a location in the vehicle environment and a spatial assignment of a plurality of parts of the map environment to the plurality of elements; determining, from a database, a predefined map, wherein the predefined map comprises a plurality of elements in a map environment, wherein: each of the elements represent a respective one of a plurality of static landmarks in the vehicle environment; the parts of the map environment are defined by spatial positions arranged as a mesh grid which comprise the positions for any potential detection point; and at least a portion of the map environment represents the vehicle environment; matching, using the spatial assignment, the plurality of detection points (P) and the plurality of elements, the matching comprises: identifying, for a respective one of the plurality of detection points (P), a respective one of the plurality of parts of the map environment, the respective one of the plurality of parts of the map environment representing a part of the vehicle environment including the location represented by the respective one of the plurality of detection points; identifying a respective one of the plurality of elements being assigned to the identified respective one of the plurality of parts of the map environment; and assigning the identified respective one of the plurality of elements to the respective one of the plurality of detection points (P); and determining the position of the vehicle based on the matching.

15. The non-transitory computer-readable medium according to claim 14, wherein the predefined map comprises a definition of a plurality of spatial segments, each of the spatial segments representing a respective one of the plurality of parts of the map environment and being assigned to a respective one of the plurality of elements.

16. The non-transitory computer-readable medium according to claim 14, wherein the instructions, when executed, configure the at least one processor to carry out the matching based on identified respective ones of the plurality of elements and respective ones of the plurality of detection points (P) to which the identified respective ones of the plurality of elements have been assigned.

17. The non-transitory computer-readable medium according to claim 14, wherein the instructions, when executed, further configure the at least one processor to determine the position of the vehicle by: determining at least one confidence value for a respective one of the plurality of detection points on the basis of the spatial assignment, wherein the at least one confidence value represents a probability that the respective one of the plurality of detection points is assigned to a respective one of the plurality of elements, and wherein the matching is carried out on the basis of the at least one confidence value.

18. The non-transitory computer-readable medium according to claim 17, wherein the instructions, when executed, configure the at least one processor to determine the at least one confidence value based on a Monte-Carlo simulation for the respective one of the plurality of detection points.

19. The non-transitory computer-readable medium according to claim 18, wherein the instructions, when executed, further configure the at least one processor to determine the position of the vehicle by: transforming the at least one confidence value to an exponential weighting factor; and matching the plurality of detection points (P) and the plurality of elements by weighting the plurality of detection points (P) and the plurality of elements by the exponential weighting factor.

20. The non-transitory computer-readable medium according to claim 14, wherein the predefined map further comprises a plurality of distance values representing distances between the plurality of parts of the map environment being assigned to the plurality of elements, respectively, wherein the instructions, when executed, further configure the at least one processor to determine the position of the vehicle by matching using the plurality of distance values.

Description

BRIEF DESCRIPTION OF THE DRAWINGS

[0042] Exemplary embodiments and functions of the present disclosure are described herein in conjunction with the following drawings, showing schematically:

[0043] FIG. 1 a block diagram of method for determining the position of a vehicle;

[0044] FIG. 2 distances between exemplary detection points and a line being assigned to the detection points;

[0045] FIG. 3 mathematical expressions for determining the distances from FIG. 2 and for determining a rigid transformation function;

[0046] FIG. 4 exemplary spatial segments assigned to lines of a map;

[0047] FIG. 5 the spatial segments from FIG. 2 as a line drawing;

[0048] FIG. 6 a portion of the map from FIG. 4 comprising the lines of the map; a portion of the map with spatial segments assigned to the lines, and a portion of the map representing distance values between the map environment and the lines;

[0049] FIG. 7 portions of the map with spatial segments assigned to lines and with a plurality of samples for a given detection point.

DETAILED DESCRIPTION

[0050] FIG. 1 depicts a block diagram of a method for determining the position of a vehicle. The method can be carried out by a computer system of the vehicle (not shown). In block 10, a plurality of scans are determined, each of the scans comprising a plurality of detection points. In block 12, a predefined map is determined from a database, wherein the map comprises a plurality of elements, each of the elements representing a static landmark. The database is stored in a non-transitory storage medium of the computer system. In the following, the elements of the map are assumed to be formed by lines. However, other types of elements are also possible.

[0051] The database comprises map data for a large area, wherein the map is determined by taking a portion of the map data, which corresponds to a preliminary position of the vehicle. In other words, the map comprises lines in a map environment, which corresponds to a current vehicle environment. The map also comprises a spatial assignment of parts of the map environment to the elements of the map. Particularly, the map environment is represented by a plurality of positions (e.g. a grid), wherein each of the positions is assigned a line of the map. The positions are potential positions of detection points of the scans. The map further comprises a distance value for each of the positions. The distance values give the minimum distance between the positions and the assigned lines, respectively. This will be described further with respect to FIG. 2 and FIG. 3.

[0052] In block 14, matching is carried out between the detection points and the elements of the map. The spatial assignment of the predefined map is used for the matching and allows for carrying out the matching with high efficiency. For a given detection point the corresponding position in the map environment is identified, which can comprise one or more rounding operations of one or more coordinates of the detection point to coordinates of the nearest position in the map environment. Other ways of identifying the position are also possible. On the basis of the spatial assignment the element being assigned to the position is also assigned to the detection point. No exhaustive search through all elements of the map is necessary. Instead, the nearest element for a given detection point can readily be identified using the spatial assignment, which is part of the predetermined map.

[0053] In block 16, the current position of vehicle is determined on the basis of the matching. This can comprise updating the preliminary position with information from the matching, in particular a rigid transformation function.

[0054] FIG. 2 shows two exemplary detection points P and P′, which are assigned to an element 18 of the map. The element 18 is considered to be the nearest element for both detection points P and P′. The element 18 is a line segment S, which extends between a first endpoint L.sub.min (i.e. startpoint) and a second endpoint L.sub.max. The distance between the detection point P and the line S is the Euclidean distance d(P), which is the minimum distance between P and S (orthogonal projection). For the detection point P′, the minimum distance is d(P′), as shown in FIG. 2, which is the distance between P′ and the end or corner point L.sub.max. In general, the minimum distance calculation can be expressed by formula 20 shown in FIG. 3. The first case of formula 20 describes the orthogonal projection, as for detection point Pin FIG. 2. The second case of formula 20 describes the minimum distance to one of the endpoints, as for the detection point P′ in FIG. 2.

[0055] Formula 20 is used to compute a spatial assignment between parts of the map and the elements of the map, for example the element 18. The result is shown in FIG. 4, which shows an exemplary map 24 comprising a plurality of elements, for example element 18, in a map environment 26. The map environment 26 is divided into spatial segments 28, which are shown in different grey scales in FIG. 4 and which form parts of the map environment 26. The spatial segments are shown without grey scales in FIG. 5, see segments 28 and 28′ for example. The spatial segments are assigned to the lines in the map environment 26, respectively. For example, the spatial segment 28 is assigned to line 18, which is the line with the minimum distance to all points within the spatial segment 28. The spatial segments 28 can be determined offline (e.g. before carrying out the method described in FIG. 1) on the basis of a regular grid of positions covering the map environment 26. For each position formula 20 is used to determine the minimum distance to all of the lines (i.e. elements) in the map environment 26. The minimum from all distance values for all of the lines is than selected and the respective line is assigned to the respective position. In this way, a spatial assignment is computed for each point of the grid, i.e. each position is assigned to the nearest line. The spatial assignment is stored as a look-up table as an additional layer of the map 24. Additionally, the corresponding distance value is stored for each of the positions as another additional layer of the map. During run time, see blocks 10, 12, and 14 in FIG. 1, the map is evaluated for each of the detection points. For example, if a given detection point P is located inside the spatial segment 28, the line 18 is readily determined as the nearest line from all lines of the map by using the spatial assignment of the map. Additionally, the distance value d is known from the map. Matching can then be carried out between detection points and the assigned lines, wherein the overall matching process is very efficient due to the spatial assignment and the distance values, which forms precomputed data.

[0056] The matching 14 comprises determining a rigid transformation function by minimizing the expression 22 of FIG. 3. Expression 22 is a double sum over N lines of the map and M detection points di. This corresponds to a least squares estimator for rotation R and translation t, wherein the j-th line is parameterized by the normal vector n.sub.j and the orthogonal distance b.sub.j between the j-th line and the origin. The expression f(p.sub.ij) denotes an exponential weighting factor for the matching and depends on the probability p.sub.ij of detection point di being assigned to the j-th line. This is done by performing a Monte-Carlo simulation incorporating current pose uncertainties (position and heading or the preliminary position of the vehicle) and the detection point uncertainties, e.g. standard deviation in angle and range for a radar detection point. The propagated error ellipse (mean and standard deviation) as distribution (Gaussian) for K randomly generated samples (inputs) for each detection point is considered. First, the samples are assigned to one or more lines by using the spatial assignment. For each line, the number of assignments from the samples is counted. The probability p.sub.ij is then calculated as the ratio between the number of assignments to the j-th line and the total amount of samples (inputs) K for the i-th detection point. The weighting factor can be used instead of a hard 0-1 assignment from the spatial assignment. This leads to an improved accuracy of the matching and hence to an improved accuracy of the determined position. The generation of K samples is illustrated in FIG. 7 showing the map portion 32′, which corresponds to the map portion 32 from FIG. 6 discussed further below. A point cloud 38 is shown in the figure and represents K random samples drawn from the distribution. The distribution is represented by a propagated Gaussian error ellipse 40. An enlarged portion 36 is also shown in FIG. 7. It is understood that by using the spatial assignment the samples 38 will be assigned to different lines because the samples 38 are located in different spatial segments. In this way, the probabilities p.sub.ij can be determined as described.

[0057] FIG. 6 shows a portion 30 of the map 24 with lines of the map 24, for example line 18. The lines represent the walls of a parking garage level. The portion 30 can also be regarded as the map 24, wherein the spatial assignment is an additional layer or portion of this map. Another portion 32 of the map 24 is shown in FIG. 6, which corresponds to the map 24 of FIG. 4 with colored spatial segments but without the lines. FIG. 6 includes another portion 34 of the map 24 representing distance values between the map environment 26 and the lines as color intensity, i.e. for a given point in the map portion 34 the color intensity indicates the distance to the nearest line according to formula 20. The distance values form another additional layer of the map 24.