Method for determining a map of a surrounding area

Abstract

The invention relates to a method for determining a map of a surrounding area in which a mobile device moves or is intended to move, having the steps of: providing (200) sensor information (202) which has been captured by means of an inertial sensor of the mobile device during a movement of the mobile device on an underlying surface in the surrounding area; determining (210), on the basis of the sensor information (202), one or more points of detectable changes in the underlying surface; determining (220), on the basis of the one or at least one of the plurality of points of detectable changes, one or more extended transition regions in the surrounding area, each extended transition region indicating a boundary between one type of underlying surface and another type of underlying surface; creating or expanding (230) a map of the surrounding area, information relating to the one or at least one of the plurality of transition regions in the surrounding area being entered into the map; and providing (240) the map of the surrounding area, in particular in order to navigate the mobile device.

Description

[0001] Description

[0002] title

[0003] Method for determining a map of an environment

[0004] The present invention relates to a method for determining a map of an environment in which a mobile device such as a vacuum or mop robot or another household robot moves or is intended to move, a system for data processing and a computer program for carrying out the method, and a mobile device.

[0005] Background of the invention

[0006] Mobile devices such as vacuum or floor-mopping robots or other household robots typically move across a surface in an environment to be serviced, such as a home. Sensors can be used to detect the environment, especially obstacles, and navigate based on this.

[0007] Disclosure of the invention

[0008] According to the invention, a method for determining a map of an environment, a data processing system and a computer program for implementing the method, as well as a mobile device having the features of the independent patent claims are proposed. Advantageous embodiments are the subject of the dependent claims and the following description.

[0009] The invention relates to mobile devices that move or are intended to move on a surface in an environment, e.g., along a specific movement path. A typical example of such a mobile device is a household robot, such as a vacuum and / or floor mopping robot. Although the invention is described below primarily with reference to household robots, it is equally applicable to other robots or mobile devices that move or are intended to move on a surface in an environment, e.g., lawnmowers, floor or street cleaning devices such as street sweeping robots (or automated street sweepers) and the like, but also other so-called service robots or other at least partially automated vehicles, such as passenger transport vehicles or goods transport vehicles (including so-called industrial trucks, e.g., in warehouses).

[0010] For mobile devices such as household robots, movements within the environment (e.g., within the household or apartment) must be carried out in a systematic and planned manner. For this purpose, household robots are equipped with sensors such as lidar sensors or laser scanners (rangefinders) or cameras to perceive their surroundings and, based on relevant information such as positions, distances, distances, and dimensions of (insurmountable) obstacles, build a map of their environment, within which they can ultimately plan their movements, e.g., the systematic vacuuming of an entire apartment (coverage path planning or "coverage problem"). These robots typically perceive obstacles such as walls or furniture in their environment and enter them into the map. In this way, for example, an area of ​​the ground that should be covered by the movement of the mobile device or robot can be determined.

[0011] As it turns out, there are also uneven surfaces which the mobile device can actually overcome or drive over, but which - if overcome, driven over or even driven over - can lead to damage, for example, either to the mobile device or to the surface. The mobile device can also get stuck, tangled up or have other problems. Such uneven surfaces include, in particular, steps, elevations and / or depressions in the surface. In the case of a household robot, examples are the edges of carpets (which lie on the floor and can be driven over) or different heights between the floors of two adjoining rooms or with different floor coverings. In the case of a street sweeper, these could be, for example, (low) curbs.

[0012] Against this background, a possibility is proposed to circumvent or at least reduce this problem by detecting such unevenness and plotting it on the map. The map of the surroundings can then be used, in particular, for navigation of the mobile device. This means, for example, that based on the map, a path of movement can be determined that the mobile device should follow. Suitable navigation or control information can then also be determined, which can be used to operate the mobile device accordingly.

[0013] For this purpose, sensor information is provided that has been recorded by an inertial sensor on the mobile device while the mobile device is moving across the surrounding ground. Based on the sensor information, one or more locations of detectable changes in the ground are then determined. These locations of detectable changes are uneven surfaces that the mobile device can overcome or drive over and that can also be detected using the sensor information or sensor data from an inertial sensor, thus generating sufficiently large changes in the recorded data.

[0014] Based on one or at least one of the several locations of detectable changes, one or more extended transition zones are then determined in the environment. A transition zone represents a boundary between one type of surface and another. This can be, for example, a boundary between a smooth floor and a carpet, or even a door threshold, where the different types of surface typically only differ in height.

[0015] A map of the surroundings is then created, with information about one or at least one of the several transition areas in the surroundings being entered into the map. Instead of creating a (possibly new) map, an existing map of the surroundings can also be expanded accordingly. The map of the surroundings is then made available for further use, such as navigation. It is also conceivable, for example, to make the map available for viewing by a user, e.g., on a display, so that the user can see the detected transition areas.

[0016] Regardless of its further use, such an extended transition area can, for example, have a linear shape. Several such linear transition areas can then also form a geometric shape, such as a rectangle, which then indicates an area with a different type of surface, such as a carpet.

[0017] Instead of individual detections or locations of changes, for example, at the edge of a carpet, the geometry of the carpet can be reconstructed as a rectangle or polygon. This applies similarly to other types of substrates instead of a carpet.

[0018] Various methods can be used for this. For example, neighboring points (the detected locations) can be connected and the resulting lines, polylines, or polygons can be simplified, e.g., using the so-called Douglas-Peucker algorithm.

[0019] In one embodiment, the one or at least one of the multiple extended transition areas is further determined based on one or more expected geometries. These can be, for example, rectangles or circles in the case of carpets. In general, such an extended transition area, in contrast to individual locations, allows for more precise mapping and delineation of larger areas with a different type of substrate. The detection of expected geometries can be performed, for example, using RANSAC. Subsequently, a fitting can also be performed based on identified inlier points, for example, by minimizing the squared distance.

[0020] Such points of detectable changes, unevenness, or ground transitions can be virtually invisible to other sensors such as cameras or laser scanners, e.g., because they correspond to only a very small but steep elevation difference, or because they are located in complete darkness. However, with the help of inertial sensors—these can be, for example, gyroscopes and / or linear acceleration sensors—such transitions or points can be detected by detecting brief vibrations while driving over them.

[0021] This particularly takes advantage of the fact that mobile devices such as household robots often already have one or more inertial sensors – also known as inertial measurement units (IMUs). In one embodiment, an inertial sensor can, for example, combine a 3D rotation rate sensor (also called a gyroscope) with a 3D linear acceleration sensor (also called an accelerometer).

[0022] In each measurement step, the angular rate sensor provides a vector with the angular rates around all three axes of the inertial sensor, x I ,y I ,z I . A simple measurement model for the angular rate G) m is e.g.:

[0023] This corresponds to the actual rotation rates, the gyroscope bias, a systematic error that can be estimated because it changes only slowly (e.g. due to temperature fluctuations), and n^a normally distributed noise with mean 0.

[0024] Assuming that the inertial sensors are installed so that the x-axis points forward, y-axis points left, and z-axis points upward, the following can be expected when the robot is moving on a flat surface (assuming a 2D movement): The linear acceleration sensor then provides a vector a in each measurement step mfor the acceleration, which combines the linear accelerations in all three axes of the inertial sensor with the gravitational acceleration:

[0025] Here, a corresponds to the linear acceleration of the sensor in sensor coordinates, g to the gravitational constant, R IW the rotation matrix, which converts vectors in world coordinates into inertial sensor coordinates, ö a the accelerometer bias, a systematic error that can be estimated, and n a a normally distributed noise with mean 0.

[0026] When the robot travels or moves at a constant speed (unaccelerated) on a plane (assuming a movement in 2D), the following measurements can be expected:

[0027] A particularly simple variant for detecting floor transitions is that a deviation of the measured values ​​of a) m and / or a mfrom expected values. If the deviations exceed previously defined thresholds, a transition is detected, ie, a point with a detectable change.

[0028] Other variants for detecting such locations or floor transitions include, for example, taking into account several consecutive measurements in a time window of a predetermined length, transforming the measurements using, for example, a discrete Fourier transformation in order to detect certain frequency components, decomposing the signal (i.e. the sensor information) using what is known as “singular spectrum analysis”, or even more complex detectors, right up to the use of machine learning methods. A combination of two or more of these variants is also conceivable. The sensor information (or sensor data) can, for example, be passed on to a computing unit or a data processing system (control unit) in the mobile device, where it is then made available, where the determinations are made as mentioned, and ultimately the map is generated or expanded. It is also conceivable, however, that this environmental information, if necessary.with further data, to a server or data center (e.g. via wireless communication); there, the determinations can then be made as mentioned and the map or navigation information or the like based on it can be sent back to the mobile device.

[0029] When driving over uneven surfaces, there is an increased risk that the mobile device or robot will get stuck (e.g. because the wheels spin, so-called "wheel slip") or that it will only be able to localize itself with less accuracy (e.g. vibrations cause so-called "motion blur" in camera images or lead to uneven speed in the lidar sensor or laser scanner). The transitions or transition areas mapped in this way can also be combined with other semantic information about the environment to support semantic mapping. Transitions or transition areas detected in this way can, for example, indicate door thresholds or transitions to other rooms. They can therefore provide useful clues for room segmentation, which has so far been created primarily based on so-called "occupancy grid maps", i.e. maps of comprehensively detected obstacles. Transitions or transition areas detected in this wayTransition areas can also indicate boundaries between carpet and other flooring and can therefore also be used to refine or verify visually detected carpet boundaries.

[0030] In one embodiment, position and / or orientation information is also provided, indicating a position and / or orientation (in combination also referred to as pose) of the mobile device at one or at least one of the multiple locations of detectable changes. This information can be determined, for example, using odometry and / or cameras and / or depth sensors such as laser scanners, in particular also within the framework of so-called SLAM. In this case, one or at least one of the multiple transition areas in the environment can then be determined based on the position and / or orientation information. Due to the simultaneously known pose of the mobile device or robot relative to its map, these specific extended transition areas can be entered (mapped) into this map with particular precision.

[0031] In one embodiment, the information relating to the one or at least one of the multiple transition areas (which are entered into the map) includes a description of the transition area and / or a description of the type of subsurface on at least one side of the transition area. Such a description may, for example, include semantic information, conceivably, for example, a subsurface class. This description of the transition area and / or the type of subsurface may, for example, be determined based on data from other sensors, as already mentioned.

[0032] In one embodiment, it is further provided that, based on the sensor information, a type of subsurface is determined at least in a region of the one or at least one of the multiple locations of detectable changes in the subsurface. The information relating to the one or at least one of the multiple transition regions then includes information about the type of subsurface on at least one side of the transition region.

[0033] In other words, in addition to the floor transitions or unevenness, a floor signature or floor classification is determined using the inertial sensors. A type of subsurface or a floor signature can be determined, for example, using a Fourier transformation of the sensor data or sensor information from the inertial sensor in the area of ​​the locations, particularly within a specific time window and / or distance window, and subsequent binning of the spectral components, particularly both in frequency and amplitude. When driving over or driving on a carpet, for example, a pattern in the vibrations of the robot results, depending on the surface structure of the carpet, which can be recorded and evaluated by the inertial sensor.

[0034] This ground signature can then be entered into the map, for example, based on the position and / or orientation of the mobile device. This ground signature can also determine the type of subsurface based on extended transition areas, such as multiple lines. For example, a detected rectangle can be classified as a carpet and entered into the map accordingly. By determining the spatial gradient of this ground signature map and incorporating the transition detections, areas can be identified with greater certainty.

[0035] In one embodiment, the mobile device has multiple ground contact points, wherein, based on multiple locations of detectable changes in the ground that correspond to the multiple ground contact points, a location of detectable changes in the ground is determined for further use. In other words, multiple locations of detectable changes that are known to belong together based on the design of the mobile device are combined into one location.

[0036] If the mobile device has several ground contact points, it is possible, depending on the angle at which a ground transition is crossed, for a transition detection to occur for each of the ground contact points, i.e. a point of detectable changes to be identified. In this case, detections that occur spatially in close succession at a short distance can be combined into a single detection. In the case of only two ground contact points, the two detections can be replaced, for example, by a detection in their middle. The known geometry of the robot can be used to determine more precisely, based on the wheel positions, at which point the ground transition, such as a door threshold, must be located. A system according to the invention for data processing or a computing unit, e.g. a control device or a control unit of a mobile device, or a server or other computer, is set up, in particular in terms of programming, to carry out a method according to the invention, e.g.in one of the described embodiments.

[0037] The invention also relates to a mobile device such as a household robot, in particular a vacuum and / or floor-mopping robot, a lawnmower, or a floor or street cleaning device. In addition to the aforementioned computing unit or data processing system, the mobile device also has at least one inertial sensor and, if appropriate, additional sensor units such as a camera and / or a laser scanner to capture environmental information.

[0038] The implementation of a method according to the invention in the form of a computer program or computer program product with program code for carrying out all method steps is also advantageous, since this entails particularly low costs, in particular if an executing control unit is also used for other tasks and is therefore already present. Finally, a machine-readable storage medium is provided with a computer program stored thereon, as described above. Suitable storage media or data carriers for providing the computer program are, in particular, magnetic, optical, and electrical memories, such as hard disks, flash memories, EEPROMs, DVDs, and others. Downloading a program via computer networks (Internet, intranet, etc.) is also possible. Such a download can be wired or cable-based or wireless (e.g., via a WLAN network, a 3G, 4G, 5G, or 6G connection, etc.).

[0039] Further advantages and embodiments of the invention will become apparent from the description and the accompanying drawings.

[0040] The invention is illustrated schematically in the drawing using an exemplary embodiment and is described below with reference to the drawing. Brief description of the drawings

[0041] Figures 1a, 1b schematically show a mobile device in one embodiment in an environment.

[0042] Figure 2 shows schematically a sequence of a method in one embodiment.

[0043] Figure 3 schematically shows an aspect of the invention in one embodiment.

[0044] Embodiment(s) of the invention

[0045] Figure 1a schematically illustrates a mobile device 100 in one embodiment in an environment 120, e.g., a room, in a top view. Figure 1b illustrates the situation from Figure 1a in a side view. Figures 1a and 1b will be described comprehensively below.

[0046] The mobile device 100 is, for example, a vacuum cleaner robot with a control or regulating unit 102 and a drive unit 104 (with wheels) for moving the vacuum cleaner robot 100, e.g., along a movement path 130. Furthermore, the vacuum cleaner robot 100 has, for example, a sensor 106 designed as a lidar sensor with a detection range.

[0047] For better illustration, the detection range is chosen to be relatively small here; in practice, however, the detection range can also be up to 360° (e.g., but at least 180° or at least 270°).

[0048] The environment 120 can be detected by the sensor 106, i.e., environmental information can be generated. In a scan using the lidar sensor, for example, a set of points, a so-called point cloud, can be generated, with each point indicating the distance of an object from which the laser beam is reflected, from the sensor. Furthermore, the robot vacuum cleaner 100 has a computing unit or a system 108 for data processing, e.g., a control unit, by means of which data can be exchanged with a higher-level system 110, e.g., via an indicated radio connection. The control or regulating unit 102 and the system 108 can also be combined.

[0049] In the system 110, for example, a map and movement paths (or general navigation information) can be determined, which are then transmitted to the system 108 in the robot vacuum cleaner 100, which the robot vacuum cleaner is then to follow. However, it can also be provided that a map and / or a movement path (or general navigation information) is determined in the system 108 itself or is received there in some other way. Instead of a movement path or the navigation information, the system 108 can also receive, for example, control information that has been determined based on a movement path or the navigation information, and according to which the control or regulating unit 102 can move the robot vacuum cleaner 100 via the drive unit 104 to follow a movement path. The movement path 130 is indicated here only as an example.

[0050] Furthermore, the robot vacuum cleaner 100 has an inertial sensor; multiple inertial sensors are also conceivable. In one embodiment, an inertial sensor or inertial measuring unit can, for example, combine a 3D rotation rate sensor (gyroscope) with a 3D linear acceleration sensor (accelerometer).

[0051] The robot vacuum cleaner 100 is located on a surface 122, a floor, of a room 120, which is bordered on the left side by, for example, a wall as an obstacle. To the right of the robot vacuum cleaner 100, a carpet 124 lies on the floor, which is also part of the surface and, like the floor, is to be vacuumed by the robot vacuum cleaner. The edges 126 of the carpet 124 each form an unevenness in the surface 122, which the robot vacuum cleaner 100 can overcome.

[0052] If the robot vacuum cleaner 100 follows the movement path shown as an example in Figure 1a, it will pass over the edge 126, i.e., the unevenness, at a point 128. Such a location of detectable changes can be determined based on the sensor information acquired by the inertial sensor.

[0053] Figure 2 schematically illustrates the flow of a method in one embodiment. The method is generally used to determine a map of an environment, and in particular also for the navigation of a mobile device such as the robot vacuum cleaner shown in Figures 1a, 1b.

[0054] In a step 200, sensor information 202 is provided, which has been acquired by an inertial sensor of the mobile device or the robot vacuum cleaner, even during a movement of the robot vacuum cleaner on a surface in the environment. For example, data can be acquired repeatedly or continuously with the inertial sensor during a movement of the robot vacuum cleaner, e.g., along the movement path shown in Figure 1a.

[0055] In a step 210, one or more locations of detectable changes in the subsurface are then determined based on the sensor information 202. Such a location is, as mentioned, designated 128 in Figure 1a, for example. In this way, further such locations can be determined, e.g., wherever the robot vacuum cleaner moves from the floor to the carpet or from the carpet to the floor. Likewise, there can be other types of such transitions with corresponding locations, as already mentioned.

[0056] In a step 212, position and / or orientation information 214 may also be provided, which indicates a position and / or orientation of the robot vacuum cleaner (in the environment) at the plurality of detectable changes.

[0057] In Figure 3, a plurality of such locations in an environment 320 are shown as an example, the locations being numbered 328a.1 etc.

[0058] In a step 220, one or more extended transition areas are then determined based on the multiple locations of detectable changes, as well as, if applicable, the position and / or orientation information. An extended transition area indicates a boundary between one type of subsurface and another type of subsurface. As mentioned, an extended transition area can be linear, for example.

[0059] This is illustrated by way of example in Figure 3. Based on the locations 328a.1 to 328a.5, for example, the linear transition region 326a can be determined. For example, it can be expected that the locations 328a.1 to 328a.5 lie at least approximately on a line, so that the line 326a can then be found using suitable methods. Similarly, based on the locations 328b.1 to 328b.5, for example, the linear transition region 326b can be determined, based on the locations 328c.1 to 328c.3, the linear transition region 326c, and based on the locations 328d.1 to 328d.4, the linear transition region 326d. Each of the transition regions separates, for example, a smooth floor from a (slightly raised) carpet as different types of subfloor.

[0060] However, one might also expect a rectangle formed by the individual transition areas or points. This can also be taken into account when determining the individual transition areas.

[0061] The vacuum cleaner robot 100 shown in Figures 1a and 1b has, for example, two ground contact points, namely where each of the two wheels touches the ground or subsurface.

[0062] Figure 4a shows the robot vacuum cleaner 100 again, with the ground contact points of the wheels 104 indicated by 404a for the left wheel and 404b for the right wheel. The x and y axes of the robot vacuum cleaner are also shown, with a center or center of gravity 405.

[0063] Depending on how the robot vacuum cleaner travels over an uneven floor, such as the edge 126 shown in Figure 1a, it is possible that the two wheels 104 travel over the edge at different times and at different locations.

[0064] In Figure 4b, such an edge or unevenness, e.g., the edge of a carpet, is designated 426. On the left, a situation is shown in which the robot vacuum cleaner is just passing over edge 426, i.e., the x-axis is perpendicular to edge 426. Accordingly, both wheels pass over edge 426 simultaneously, so that the vibration detected by the inertial sensor occurs simultaneously for both ground contact points. Accordingly, for example, only one point is detected, the position of which can be determined as the position of the center of gravity 405.

[0065] On the right, however, a situation is shown in which the robot vacuum cleaner crosses edge 426 at an angle, i.e., the x-axis is at an angle of, for example, between 0° and 90° to edge 426. Accordingly, both wheels cross edge 426 at different times, so that the vibration detectable by the inertial sensor occurs at different times for both ground contact points. Accordingly, two locations are detected. In a simple example, the position of the center of gravity 405 could be assigned to each of the two locations.

[0066] Preferably, however, both locations are combined into one location, to which the position of the center of gravity 405 is then assigned as the position. The fact that these two locations can be assigned to a crossing of the edge by the robot vacuum cleaner can be determined based on the time interval between the detection of the two locations; this time interval is usually very short.

[0067] It is also conceivable, however, that each of the two locations is assigned the position of the specific ground contact point as a position, provided this is known with sufficient accuracy. However, since, depending on the type of evaluation of the sensor data, it is not possible to distinguish, for example, which wheel drove over the edge first, combining them into one location with one position is particularly useful. In a step 230, a map of the surroundings can then be expanded, with information on the transition areas being entered into the map. The map can then be made available for further use in step 240, e.g., for navigation of the robot vacuum cleaner.

[0068] For this purpose, Figure 5 shows, as an example, a map 500 of an environment (which here does not correspond to the environment according to Figures 1a, 1b). Map 500 shows, as an example, environmental information in the form of lidar points 502; such points 502 thus indicate, for example, the boundaries of obstacles such as walls or larger objects. A trajectory 504 is also shown, along which the mobile device is traveling or has traveled, for example, while the environmental information was being acquired.

[0069] Furthermore, various locations of detectable changes are shown here as examples, which were determined based on the sensor information of the inertial sensor during the movement of the mobile device along the trajectory 504. For example, the locations labeled 506 are door thresholds, the locations labeled 508 are a small other threshold or elevation, and the locations labeled 510 are the edges of a carpet.

[0070] At this point it should be mentioned that for the sake of clarity and explanation only the locations themselves are entered in the map, but as described above, for example with reference to Figure 3, extensive transition areas are determined based on such locations, which are then entered in the map 500.

[0071] In an optional step 250, it can also be provided that, based on the sensor information, a type of subsurface is determined at least in a region of the one or at least one of the multiple locations of detectable changes in the subsurface. For this purpose, for example, a Fourier transformation of the sensor data can be determined in at least the region of the one or at least one of the multiple locations. The information relating to the one or at least one of the multiple transition areas in the map then includes information relating to the type of subsurface on at least one side of the transition area.

[0072] For this purpose, Figure 6a shows an example of a trajectory 630 along which the mobile device moves, for example. The trajectory is in two parts here, e.g. because a connecting part is not shown or because both parts were traveled on different journeys. In addition, locations 628 of detectable changes are shown, based on which extensive transition areas 626 can be determined. In this case, it can be difficult, for example, to recognize a larger, contiguous area, e.g. a carpet, based solely on these transition areas 626, as indicated here by the dashed line. Although the mobile device moves on the carpet, it only moves over an edge of the carpet in a few places.

[0073] Figure 6b, however, now indicates that a type of subsurface can also be determined based on sensor information from other areas, particularly along the entire trajectory. Where the carpet is located, designated here as 624, a different ground signature results than in the rest of the area, with the ground signature being recognizable, for example, through a Fourier transformation of the sensor information. This allows the area in which the carpet is located to be determined much more precisely.

[0074] This can also be entered into the map, making it even more accurate.

Claims

Claims 1 . A method for determining a map of an environment in which a mobile device (100), in particular a robot, preferably a household or vacuum cleaning robot, moves or is to move, comprising: Providing (200) sensor information (202) that has been detected by an inertial sensor (107) of the mobile device during a movement of the mobile device on a surface (122) in the environment (120); Determining (210), based on the sensor information (202), one or more locations (128) of detectable changes in the subsurface; Determining (220), based on the one or at least one of the plurality of locations (128) of detectable changes, one or more extended transition regions (326a, 326b, 326c, 326d) in the environment, each extended transition region indicating a boundary between one type of subsurface and another type of subsurface; Creating or expanding (230) a map of the environment, wherein information about the one or at least one of the several transition areas in the environment is entered into the map; and Providing (240) the map of the surroundings, in particular for navigation of the mobile device.

2. The method of claim 1, further comprising: Providing (212) position and / or orientation information (214) indicating a position and / or orientation of the mobile device at the one or at least one of the plurality of detectable changes, wherein the one or at least one of the plurality of transition areas in the environment is further determined based on the position and / or orientation information.

3. The method according to claim 1 or 2, wherein the one or more extended transition regions each have a line shape.

4. The method of any preceding claim, wherein the one or at least one of the plurality of extended transition regions is further determined based on one or more expected geometries.

5. Method according to one of the preceding claims, wherein the information relating to the one or at least one of the plurality of transition regions comprises a description of the transition region and / or a description of the type of subsurface on at least one side of the transition region.

6. Method according to one of the preceding claims, further comprising: Determining (250), based on the sensor information, a type of subsurface at least in a region of the one or at least one of the plurality of locations of detectable changes in the subsurface, wherein the information about the one or at least one of the plurality of transition regions comprises information about the type of subsurface on at least one side of the transition region.

7. The method according to claim 6, wherein determining the type of the subsurface comprises: determining a Fourier transform of the sensor data in at least the area of ​​the one or at least one of the plurality of locations, and determining the type of the subsurface based on the Fourier transform 8. The method according to any one of the preceding claims, wherein the mobile device has a plurality of ground contact points (404a, 404b), and wherein a location of detectable changes in the ground is determined for further use based on a plurality of locations of detectable changes in the ground corresponding to the plurality of ground contact points.

9. Method according to one of the preceding claims, further comprising: Providing environmental information that has been captured at least partially using a camera and / or a depth sensor, wherein the map of the environment is generated based on the environmental information.

10. The method according to any one of the preceding claims, wherein the mobile device is navigated based on the map. 11 . A data processing system comprising means for carrying out the method according to any one of the preceding claims 12. A mobile device (100) comprising at least one inertial sensor for acquiring sensor data and a system according to claim 11.

13. Mobile device (100) according to claim 12, which is designed as a household robot, in particular a vacuum and / or wiping robot, or as a floor or street cleaning device or as a lawnmower.

14. A computer program comprising instructions which, when executed by a computer, cause the program to carry out the method steps of a method according to any one of claims 1 to 10 when executed on the computer.

15. A computer-readable storage medium on which the computer program according to claim 14 is stored.

Images