Methods, systems, and computer program products for determining the pose of a mobile unit.

Abstract

The present invention relates to a method for determining the pose of a mobile unit (1) using at least one sensor device and a map (MAP), the sensor device being configured to acquire an environmental image (PIC) of the mobile unit (1), in which at least one drivable area (2) is marked. In this method, at least one environmental image (PIC) is acquired by means of the at least one sensor device, at least one collision-free area (3, 3.1, 3.2) is calculated based on the acquired at least one environmental image (PIC), and the pose of the mobile unit (1) is determined by comparing the calculated at least one collision-free area (3, 3.1, 3.2) with the at least one drivable area (2) marked in the map (MAP). Furthermore, a system for performing the method is also described.

Description

Technical Field

[0001] This invention relates to methods, systems, and computer program products for determining the pose of a mobile unit. Background Technology

[0002] Localization methods, for example, those used in robotics or autonomous driving, can be used to accurately determine the pose or position of a vehicle. Here, landmarks, such as traffic signs or signal lights, are detected in images captured by a camera, these landmarks are classified, and then assigned to known landmarks with predetermined location information. The vehicle's pose or position can then be determined, for example, from this assignment.

[0003] The accuracy of this method is highly sensitive to the number of landmarks detected and the uniqueness of the landmarks assigned to known landmarks. In particular, this method is not feasible or its accuracy is insufficient in areas where only a few landmarks can be detected, landmarks are completely undetectable, or landmarks cannot be explicitly assigned. Summary of the Invention

[0004] Therefore, the object of the present invention is to avoid the disadvantages mentioned above and to provide a reliable and accurate method, system and computer program product for determining the pose of a mobile unit.

[0005] According to the invention, this objective is achieved by the features mentioned in claims 1, 13, and 15. Advantageous variations are derived from the features mentioned in the dependent claims.

[0006] This invention relates to a method for determining the pose of a mobile unit using at least one sensor device and a map. The at least one sensor device is configured to acquire an environmental image of the mobile unit, and at least one drivable area is marked on the map. In this method, at least one environmental image is acquired by means of the at least one sensor device, and at least one collision-free area is calculated based on the at least one acquired environmental image. The pose of the mobile unit is determined by comparing the at least one collision-free area with at least one drivable area marked on the map.

[0007] The proposed method allows for the reliable and flexible determination of the pose of mobile units. Particularly advantageously, this method can be used in environments where few or no landmarks are detectable, or where few or no landmarks are known, or where few or no landmarks are marked on the map, such as on small side roads or in remote areas with sparse traffic. This method is particularly useful when the allocation of detected landmarks to known landmarks cannot be definitively or reliably determined using landmark-based localization methods. Even in such challenging situations, collision-free areas can always be reliably and robustly calculated based on at least one acquired environmental image and compared with at least one drivable area marked on the map.

[0008] For example, the mobile unit can be a land vehicle (e.g., a motor vehicle or robot) or a vessel (e.g., a small boat or ship). The pose of the mobile unit can at least include position information. The position information can be formed, for example, using the values ​​of two or three coordinates. The pose can also include azimuth information. For example, azimuth information can be formed using the values ​​of two or three azimuth angles. The pose of the mobile unit preferably includes position information and azimuth information. Here, a pose can be represented by a vector with six components. Here, three components can correspond to coordinates, such as geographic coordinates or GPS coordinates (GPS = Global Positioning System). The other three components can correspond to azimuth angles.

[0009] At least one marked drivable area may at least partially correspond to a sub-region of road surface or water surface. Preferably, the map has map location information for at least one marked drivable area. Here, the map location information can be formed using two or three coordinates in a map coordinate system. At least one marked drivable area can be determined, for example, by a two-dimensional or three-dimensional set of points or pixels. Preferably, at least one marked drivable area corresponds to a two-dimensional surface or set of points in a two-dimensional or three-dimensional map coordinate system. The map coordinate system may be a global reference system, which may be formed, for example, using geographic coordinates or GPS coordinates.

[0010] The map may be a semantic map. Preferably, the map includes at least one known landmark with map location information. The at least one landmark may be, for example, a traffic sign, a traffic light, or a fire hydrant. The map location information of the landmark may include the values ​​of two or three coordinates of the corresponding landmark in the map coordinate system.

[0011] At least one sensor device may have a camera, preferably a monocular camera. Particularly preferably, calibration is performed via at least one sensor device. Calibration allows at least one coordinate to be assigned to a point or pixel of at least one environmental image, for example, its height relative to the local ground plane of the mobile unit.

[0012] The local ground plane of the mobile unit may correspond to at least one sub-region of the mobile unit's drivable surface. The local ground plane may be determined, for example, by the support points of the mobile unit's wheels on the mobile unit's drivable surface.

[0013] Calibration can be performed based on a camera coordinate system, wherein the optical axis of the camera, which is at least one sensor device, can be equivalent to the coordinate axis of the camera coordinate system. Furthermore, the height of at least one sensor device, motion unit, or optical axis relative to the local ground plane of the motion unit can be predetermined and / or taken into account during calibration.

[0014] At least one collision-free region may be a set of points or pixels in two or three dimensions. At least one collision-free region may, for example, correspond to a set of points in a two- or three-dimensional coordinate system. At least one collision-free region may be calculated as a sub-region of a surface (e.g., a road surface or water surface) recorded by at least one environmental image and accessible by the mobile unit. At least one collision-free region may particularly form an area of ​​the mobile unit's drivable surface accessible from the mobile unit along the direction of travel. At least one collision-free region may particularly be free of obstacles, such as other mobile units, sidewalks, or fire hydrants.

[0015] At least one collision-free region can be extracted from at least one environment image as a mask, wherein a bit value can be assigned to each pixel of the at least one environment image. Pixels with a bit value of 1 can be associated with the collision-free region, for example, while pixels with a bit value of 0 can be associated with the complementary portion of the collision-free region. The mask defined in this way can be represented by a binary matrix, wherein the rows and columns of the binary matrix can correspond to the row and column arrangements of the pixels of the at least one environment image.

[0016] Preferably, at least one collision-free region is calculated based on at least one environmental image using semantic segmentation. Particularly preferably, at least one collision-free region is calculated using machine learning methods, particularly using a trained neural network.

[0017] The method may include an initialization step. The position or pose of the mobile unit may be initially estimated, for example, by means of at least one other sensor device. At least one other sensor device may include, for example, a satellite positioning system for acquiring GPS data, at least one speed sensor for acquiring mileage data, and / or a LIDAR system (LIDAR = light detection and ranging). For example, the initially estimated pose or position of the mobile unit can be used to retrieve or obtain a map via a mobile communication network through an external server and / or from the mobile unit's electronic storage unit.

[0018] The method may also include geometric projection. For example, at least one calculated collision-free area may be preferably projected onto the local ground plane of the mobile unit based on calibration. Preferably, at least one drivable area marked on the map may also be projected onto the local ground plane of the mobile unit, for example, based on the initial estimated pose or position of the mobile unit or based on an assumption about the pose of the mobile unit. This geometric projection allows for a particularly fast and efficient comparison between at least one calculated collision-free area and at least one drivable area marked on the map.

[0019] Particularly advantageously, the proposed method can be combined with known positioning methods, especially landmark-based positioning methods. For example, at least one assumption about the pose of the moving unit can be determined or calculated first by means of known positioning methods, especially landmark-based positioning methods.

[0020] A plausibility check can be performed when comparing at least one calculated collision-free area with at least one drivable area marked on the map. The plausibility check can be used to check or determine the plausibility of at least one hypothesis, or to select one hypothesis from multiple hypotheses, preferably the plausible one, and take it into account when determining the pose of the mobile unit. Alternatively or additionally, the plausibility check can also be used to determine the reliability of the initially estimated pose of the mobile unit.

[0021] Preferably, the localization method is a landmark-based localization method. A landmark-based localization method may include detecting and / or classifying at least one landmark in at least one environmental image. Detection may be performed, for example, based on a trained neural network.

[0022] Alternatively or additionally, classification can be performed using a trained neural network.

[0023] At least one assumption about the pose of the mobile unit can be determined based on the association between at least one landmark detected in at least one acquired environmental image and at least one known landmark marked on a map. For this purpose, the localization method may include, for example, a particle filter or a Kalman filter, preferably a multi-hypothesis Kalman filter.

[0024] Landmark-based localization methods may not be able to definitively assign multiple detected landmarks to multiple known landmarks. In such cases, multiple hypotheses can be determined. Each hypothesis can correspond to a specific assignment of the detected landmarks to the known landmarks. For the corresponding assignment or hypothesis, a hypothesis about the pose of the mobile unit can then be determined using the landmark-based localization method. In particular, multiple hypotheses about the pose of the mobile unit can be determined in this way.

[0025] If multiple hypotheses are identified using a positioning method, it is advantageous to perform a reasonableness check on all of these hypotheses. If multiple reasonable hypotheses are subsequently identified, these multiple reasonable hypotheses can be weighted during comparison or reasonableness checks.

[0026] Reasonableness checks may include determining at least one collision-free area that can be explicitly / unambiguously assigned to at least one drivable area marked on a map, based on at least one assumption. For comparison or reasonableness checks, at least one collision-free area and at least one drivable area may be transformed to a common coordinate system based on at least one assumption and corresponding coordinate transformations.

[0027] Preferably, based on at least one assumption and / or calibration, at least one collision-free area and at least one drivable area are projected onto a surface in a common coordinate system. This surface may, for example, correspond to the local ground plane of the mobile unit. It may also correspond to a surface defined by a marked drivable area in a map. Furthermore, it may correspond to a surface or hypersurface defined by at least one marked drivable area in a map coordinate system or camera coordinate system.

[0028] An assumption is considered reasonable if the ratio of the size of a sub-region of a collision-free area that can be explicitly assigned to a marked drivable area based on at least one assumption to the size of its complementary portion exceeds a critical value. The weight of a reasonable assumption can also be determined based on this determined ratio of the size of the identified sub-region to the size of its complementary portion.

[0029] For example, the amount of intersection between at least one collision-free area projected onto a common surface and at least one marked drivable area can be determined. The number of valid pixels can correspond to the number of pixels contained in the intersection (pixels that can be explicitly assigned to at least one marked drivable area). The number of invalid pixels can correspond to the number of pixels contained in at least one collision-free area but not belonging to the intersection (pixels that cannot be explicitly assigned to at least one marked drivable area).

[0030] Therefore, the reasonableness check may include comparing the ratio of valid pixels to invalid pixels with a predetermined threshold. If the specified threshold is exceeded, at least one hypothesis can be classified as reasonable. The weights of multiple reasonable hypotheses can then be determined based on the ratio of valid pixels to invalid pixels, where the highest weight corresponds to the numerically largest ratio of valid pixels to invalid pixels.

[0031] During the comparison, the pose of the mobile unit, or a new estimate of the pose of the mobile unit, can be determined based on a reasonableness check. For example, the pose of the mobile unit determined by comparison may correspond to a reasonable assumption. The pose of the mobile unit determined by comparison may also correspond to the initially estimated pose of the mobile unit. For example, if no reasonable assumption is determined by the reasonableness check, the pose of the mobile unit determined by comparison may correspond to the initially estimated pose of the mobile unit. If multiple reasonable assumptions are determined by the credibility check, and the comparison includes a weighted average of the multiple reasonable assumptions, the pose of the mobile unit determined by comparison may also correspond to the reasonable assumption with the highest weight.

[0032] As an addition to or alternative to the localization method, multiple assumptions about the pose of the mobile unit can also be determined as initial values ​​or initial vectors in the form of regular or random grids. The grid can be determined, for example, based on the initially estimated position or pose of the mobile unit, or include it as an initial value or initial vector. The hypothetical position and / or azimuth information determined in this way can correspond to a point cloud in a coordinate system (e.g., a two-dimensional or three-dimensional Bravais grid, preferably a square or cubic grid). Alternatively, the multiple assumptions as initial vectors can also be randomly generated using a random generator.

[0033] As an addition to or alternative to rationality checks, the pose or position of the moving unit can be updated.

[0034] To update the pose or position of a mobile unit, a cost function can be minimized when comparing at least one calculated collision-free region with at least one drivable region marked on the map. For at least one assumption, the cost function can be given by characteristic distances between at least one calculated collision-free region and at least one drivable region marked on the map in a common coordinate system. Here, the minimum value of the cost function can be determined by means of optimization methods, such as least squares, Monte Carlo simulation, linearization in the correction step of the Kalman filter, or simulated annealing. The determined mobile unit pose corresponds to the minimum value of the cost function by comparison. At least one characteristic distance can be determined based on the shortest distance between the corresponding 3D point that can be assigned to at least one collision-free region and the surface defined by at least one marked drivable region on the map. For example, this at least one characteristic distance can correspond to the average or sum of all shortest distances that can be determined in this way.

[0035] Preferably, minimizing the cost function is used to determine the update of the position or pose of the mobile unit.

[0036] In a preferred embodiment, the cost function is minimized for one or more assumptions determined by the positioning method. Therefore, the method for minimizing or optimizing the cost function can also be combined with a landmark-based positioning method.

[0037] Equally feasible is to perform cost function minimization here for several assumptions specifically determined as initial values ​​or initial vectors. The cost function minimization or optimization method can then be advantageously used for mobile units in environments where landmarks are undetectable, or where no known landmarks are marked on the map for that environment, thus landmark-based localization methods are not feasible. It can also be specified that the initial estimated pose of the mobile unit is determined by means of other sensor devices, at least one assumption is determined by means of a localization method, and / or at least one other assumption is determined by minimizing the cost function. A reasonableness check can also be performed on the initial estimated pose, at least one assumption, and / or at least one other assumption. The pose of the mobile unit can be determined by comparing at least one calculated collision-free area with at least one drivable area marked on the map, i.e., it corresponds to the initial estimated pose, at least one assumption, or at least one other assumption.

[0038] The present invention also relates to a system comprising a mobile unit having at least one sensor device, an electronic storage unit, and an electronic analysis and control unit. The at least one sensor device is configured to acquire at least one environmental image of the mobile unit. A map having at least one drivable area marked on a map is stored in the electronic storage unit. The electronic analysis and control unit is configured to calculate at least one collision-free area based on at least one environmental image acquired by means of the at least one sensor device. The electronic analysis and control unit is further configured to determine the pose of the mobile unit by comparing the at least one collision-free area with at least one drivable area marked on the map.

[0039] Preferably, at least one sensor device has a calibrated monocular camera. Particularly preferably, the electronic analysis and control unit is configured to calculate at least one collision-free region based on a sequence of environmental images acquired by means of the calibrated monocular camera.

[0040] Alternatively, at least one sensor device may have a 3D sensor, such as a LiDAR sensor.

[0041] At least one sensor device may also have a combination of one or more 3D sensors and / or one or more 2D sensors (e.g., multiple cameras).

[0042] The electronic storage unit may be part of the electronic analysis and control unit or electronically connected to an external server. The electronic storage unit may be a read-only memory or a working memory.

[0043] The electronic analysis and control unit may have a CPU (central processing unit), a GPU (graphics processing unit), and / or a computing unit. The system, mobile unit, and / or electronic analysis and control unit may also have at least one other sensor device and / or mobile communication system.

[0044] At least one additional sensor device may be configured to initially estimate the pose or position of the mobile unit. This additional sensor device may, for example, have a satellite positioning system, preferably a GPS system (Global Positioning System), or be configured as part of a satellite positioning system. This additional sensor device may also have at least one speed sensor for estimating mileage data.

[0045] The mobile communication system may be a communication system that supports WLAN (wireless local area network) and / or mobile radio. For example, the mobile communication system may be configured to query or retrieve maps from an electronic storage unit via an external server.

[0046] The present invention also provides a computer program product. The computer program product has a computer program (or instruction sequence). In this case, the computer program has software media / software components for performing the methods described above or manipulating the systems described above when the computer program is run in a computing unit.

[0047] Preferably, the computer program product may be directly loaded into the internal electronic memory or memory unit of the computing unit, or may already be stored there and typically includes partial program code for performing the described methods when the computer program product is run or executed in the computing unit. The program code or a portion thereof may be developed using a scripting language or a compiled language such as C, C++, or Python. The computer program product may be stored on a machine-readable medium, preferably a digital storage medium. Attached Figure Description

[0048] Embodiments of the invention are illustrated in the accompanying drawings, and will be further explained below by means of... Figures 1 to 2 It is explained in detail, including:

[0049] Figure 1 A schematic illustration of the acquired environmental image showing the calculated collision-free area is shown, and

[0050] Figure 2 A schematic illustration of a map showing marked drivable areas is provided. Detailed Implementation

[0051] In the following text, Figure 1 and Figure 2 The features that appear repeatedly in the figure are each given the same reference numerals.

[0052] Figure 1 A schematic illustration shows an environmental image PIC of a vehicle (not shown) as a moving unit 1, acquired using a monocular camera. Based on the environmental image PIC, a collision-free region 3 is calculated using machine learning methods and marked in the environmental image PIC. Here, the collision-free region 3 corresponds to a sub-region of the road surface 5 that is traversable in the driving direction. Figure 1 In this context, the collision-free zone 3 is adjacent to vehicle 1.1 traveling in front of the mobile unit 1 and vehicle 1.2 approaching the mobile unit 1. Furthermore, traffic signs serving as landmarks 4.1 are detected using a trained neural network based on the environmental image PIC.

[0053] Figure 2 A schematic illustration of a semantic map (MAP) showing the marked drivable area 2 and known landmark 4.2 is provided. The map MAP also includes map location information for known landmark 4.2.

[0054] The pose of mobile unit 1 is determined by comparing at least one collision-free area 3 with at least one drivable area 2 marked in a map MAP. To this end, the position of mobile unit 1 is first initially estimated using a GPS system. Based on the initial estimated position of mobile unit 1, landmarks 4.1 detected in the environmental image PIC are assigned to landmarks 4.2 marked and known in the map MAP using a known landmark-based localization method. Based on the assignment of known landmarks 4.2 and map location information, assumptions about the pose of mobile unit 1 are determined using a known landmark-based localization method.

[0055] Based on the pose assumption of mobile unit 1 and the pre-calibrated monocular camera, Figure 1 The collision-free zone 3, marked and calculated in the environmental image PIC, is projected onto the drivable zone 2, marked in the map MAP. Figure 2 In the image PIC, the projection of the collision-free area 3, which is marked and calculated in the environment image PIC, onto the drivable area 2 marked in the map MAP corresponds to sub-regions 3.1 and 3.2.

[0056] Then, a reasonableness check is performed on the pose assumptions of mobile unit 1 during comparison. In the reasonableness check, sub-region 3.1, which can be explicitly / unambiguously assigned to the drivable area 2 marked in the map MAP, is determined from the calculated collision-free areas 3, 3.1, and 3.2. At this point, the number of valid pixels corresponds to the number of pixels contained in sub-region 3.1. The number of invalid pixels corresponds to the number of pixels contained in sub-region 3.2. Sub-region 3.2 is a supplement to sub-region 3.1. The higher the number of valid pixels, the more reasonable the pose assumptions of mobile unit 1 are. In particular, the ratio of the number of valid pixels to the number of invalid pixels is compared with a predetermined threshold. If the ratio determined in this way is greater than the specified threshold, reasonableness exists. Figure 2 In the example shown, the specified threshold value is 4. The ratio of valid pixels to invalid pixels is 5. Therefore, the assumptions determined by the landmark-based positioning method are reasonable. Thus, the pose of mobile unit 1 determined by comparing the calculated collision-free area 3 and the drivable area 2 marked in the map MAP corresponds to a reasonable assumption about the pose of mobile unit 1.

[0057] In another embodiment, multiple landmarks 4.1 are detected in the environmental image PIC. The map MAP also includes multiple known landmarks 4.2, each with map location information. Based on the initial estimated position of the mobile unit 1, multiple hypotheses are determined using a landmark-based localization method, which is itself known. Here, each hypothesis correspondingly assigns a known landmark 4.2 to the detected landmark 4.1. For each associated hypothesis, a hypothesis of the pose of the mobile unit 1 is then determined based on the map location information of the corresponding associated known landmark 4.2.

[0058] For each hypothesis determined in this way, the calculated collision-free region 3 is projected onto the drivable region 2 marked in the map MAP, and the ratio of the number of valid pixels to the number of invalid pixels is determined accordingly. The weight of the corresponding hypothesis is also determined based on the ratio determined in this way, wherein the hypothesis with the largest ratio determined in this way is assigned the largest weight. At this point, the determined pose of the mobile unit 1 is compared to the hypothesis for the pose of the mobile unit 1 with the largest weight.

[0059] In another embodiment, a landmark-based positioning method is not performed. Instead, a random number generator is used to randomly determine 10,000 assumptions about the pose of the mobile unit 1.

[0060] Then, a cost function is minimized in the comparison, whereby the cost function specifies the characteristic distance between the calculated collision-free region 3 and the drivable region 2 marked in the map MAP in a common coordinate system, for each randomly determined hypothesis. To determine the characteristic distance, based on the corresponding hypothesis, a coordinate transformation of the calculated collision-free region 3 from the camera coordinate system to the map coordinate system is performed, and the shortest distance between the corresponding 3D point in at least one collision-free region 3 and the surface defined by the marked drivable region 2 in the map MAP is determined. The sum of the shortest distances determined in this way then corresponds to the characteristic distance.

[0061] The cost function is minimized by linearization in the correction step of the Kalman filter. Therefore, the minimum value of the cost function corresponding to the determined pose of mobile unit 1 is used to determine the update of the pose of mobile unit 1.

[0062] The method described in the embodiments is executed using a system comprising a mobile unit 1 having a monocular camera as a sensor device, an electronic storage unit, and an electronic analysis and control unit. The system also includes a GPS system. The monocular camera is configured to acquire environmental images of the mobile unit 1. A map MAP showing marked drivable areas 2 is stored in the electronic storage unit. The electronic analysis and control unit is configured to calculate collision-free areas 3 using a series of environmental images acquired by the monocular camera. The electronic analysis and control unit is further configured to determine the pose of the mobile unit 1 by comparing the collision-free areas 3 with the drivable areas 2 marked in the map MAP.

[0063] Features disclosed only in the embodiments of various implementations may be combined with each other and individually claimed.

Claims

1. A method for determining the pose of a mobile unit (1), which is a land vehicle or vessel, using at least one sensor device and a map (MAP), wherein the at least one sensor device is configured to acquire an environmental image (PIC) of the mobile unit (1), marking at least one drivable area (2) in the map, the map (MAP) being stored in an electronic storage unit, the map (MAP) being a semantic map and / or including at least one known landmark (4.2) with map location information, in which, At least one environmental image (PIC) is acquired using the at least one sensor device, and at least one collision-free region (3) is calculated based on the acquired environmental image (PIC). The at least one collision-free region (3) corresponds to a surface from the moving unit (1) that allows for collision-free travel along the driving direction. The pose of the mobile unit (1) is determined by comparing the calculated at least one collision-free area (3) with the at least one drivable area (2) marked in the map (MAP).

2. The method according to claim 1, characterized in that, The pose of the mobile unit (1) includes position information and / or azimuth information, and / or the marked at least one drivable area (2) corresponds to a two-dimensional point set or a three-dimensional point set in the map coordinate system.

3. The method according to claim 1 or 2, characterized in that, The at least one collision-free region (3) corresponds to a two-dimensional point set or a three-dimensional point set in the camera coordinate system. And / or the at least one collision-free region (3) is calculated based on at least two sequentially acquired environmental images (PICs). And / or the at least one collision-free region (3) is calculated based on a monocular environment image sequence. And / or at least one collision-free region (3) is calculated by means of semantic segmentation, And / or the at least one collision-free region (3) is calculated by means of a machine learning method, which is a trained neural network.

4. The method according to claim 1 or 2, characterized in that, The calculated at least one collision-free area (3) and / or the at least one drivable area (2) marked in the map (MAP) are projected onto the local ground plane of the mobile unit (1).

5. The method according to claim 1 or 2, characterized in that, At least one hypothesis of the pose of the moving unit (1) is calculated by a positioning method, and the rationality of the calculated at least one hypothesis is checked in the comparison.

6. The method according to claim 5, characterized in that, The positioning method described is a landmark-based positioning method. Based on the allocation relationship between at least one landmark (4.1) detected in the acquired environmental image (PIC) and at least one known landmark (4.2) marked in the map (MAP), at least one assumption is made to determine the pose of the mobile unit (1). Multiple assumptions about the pose of the mobile unit (1) are determined by using a landmark-based positioning method.

7. The method according to claim 5, characterized in that, In the rationality check, the following sub-region (3.1) of the at least one collision-free zone (3) is determined: the sub-region can be explicitly assigned to at least one drivable zone (2) marked on the map (MAP) based on the at least one assumption. If the size of the sub-region of the at least one collision-free area (3) that can be assigned to the marked at least one drivable area (2) according to the at least one assumption is greater than the size of its complementary area, then the at least one assumption is reasonable.

8. The method according to claim 5, characterized in that, The pose determined by comparison of the moving unit (1) corresponds to a reasonable assumption. An initial pose estimation of the moving unit (1) is performed using another sensor device. The pose of the moving unit (1) determined by comparison corresponds to the initial estimated pose or reasonable assumption of the moving unit (1), and / or Multiple reasonable assumptions are determined by means of a reasonableness check, the comparison includes a weighted average of the multiple reasonable assumptions, and the pose of the moving unit determined by the comparison corresponds to the reasonable assumption with the largest weight.

9. The method according to claim 1 or 2, characterized in that, Determine at least one assumption about the pose of the mobile unit (1) and minimize the cost function during comparison, wherein the cost function of the at least one assumption gives at least one characteristic distance in a common coordinate system between at least one calculated collision-free area (3) and a drivable area (2) marked in the map (MAP), and the pose of the mobile unit (1) corresponds to the minimum value of the cost function.

10. The method according to claim 9, characterized in that, The at least one characteristic distance is determined based on the shortest distance between each three-dimensional point in the at least one collision-free area (3) and the surface defined by the at least one drivable area (2) marked in the map (MAP).

11. The method according to claim 9, characterized in that, The pose of the moving unit (1) is updated by minimizing the cost function.

12. A system comprising a mobile unit (1) having at least one sensor device, an electronic storage unit, and an electronic analysis and control unit, the mobile unit being a land vehicle or vessel, the sensor device being configured to acquire at least one environmental image (PIC) of the mobile unit (1), the electronic storage unit storing a map (MAP) having at least one marked drivable area (2), the map (MAP) being a semantic map and / or including at least one known landmark (4.2) with map location information; the electronic analysis and control unit being configured to calculate at least one collision-free area (3) based on the at least one environmental image (PIC) acquired by means of the at least one sensor device, the at least one collision-free area (3) corresponding to a surface from the mobile unit (1) that can be driven without collision in a driving direction, and to determine the pose of the mobile unit (1) by comparing the at least one collision-free area (3) with the at least one drivable area (2) marked in the map (MAP).

13. The system according to claim 12, characterized in that, The system and / or the mobile unit (1) and / or the electronic analysis and control unit have at least one other sensor device, wherein the at least one other sensor device includes a satellite positioning system and / or a mobile communication system.

14. A computer program product having a computer program having a software medium for performing the method according to any one of claims 1 to 11 or for manipulating the system according to any one of claims 12 to 13 when the computer program is run in a computing unit.

Images