Methods for operating an autonomous vehicle in an environment and autonomous vehicle

Abstract

The invention relates to a method for operating an autonomous vehicle (1) in an environment, wherein the autonomous vehicle (1) has at least one laser scanner (2) for acquiring laser scans for detecting objects, and wherein the autonomous vehicle (1) has a map of the environment, and wherein the environment comprises at least one object to which an object (4) is associated, which is recorded in the map, wherein, during a movement of the autonomous vehicle (1) along a trajectory in the environment, a laser scan of a part of the environment, which contains at least one object, is acquired by the at least one laser scanner (2) of the autonomous vehicle (1); the laser scan is compared with at least one object (4) recorded in the map by calculating a correlation (K) of objects detected in the laser scan with objects (4) recorded in the map for several orientations (A) of the laser scan;A pose of the autonomous vehicle (1) on the map is determined based on the orientation (A) of the laser scan for which the calculated correlation (K) of the objects detected in the laser scan with the objects (4) recorded on the map has the highest value. Alternatively, for several orientations (A) of the laser scan, a deviation of the objects detected in the laser scan with the objects (4) recorded on the map is calculated; and a pose of the autonomous vehicle (1) on the map is determined based on the orientation (A) of the laser scan for which the calculated deviation of the objects detected in the laser scan with the objects (4) recorded on the map has the lowest value. The invention also relates to an autonomous vehicle (1) that can be operated using one of the methods according to the invention.

Description

[0001] The invention relates to methods for operating an autonomous vehicle in an environment, wherein the autonomous vehicle has at least one laser scanner for recording laser scans to detect objects, and wherein the autonomous vehicle has a map of the environment, and wherein the environment comprises at least one object to which an object is assigned that is recorded on the map. The invention also relates to an autonomous vehicle that can be operated using the methods according to the invention.

[0002] The environment is, in particular, a technical facility, such as a production plant, an industrial hall, or a logistics center. Autonomous vehicles are used, for example, to transport materials within the technical facility. The technical facility also contains other objects, such as walls, columns, production machines, pallets, boxes, containers, or transport trolleys, as well as people and other autonomous vehicles. The autonomous vehicles also have sensors, particularly laser scanners for recording laser scans, for detecting such objects.

[0003] Objects in the environment, such as walls, columns, production machinery, pallets, boxes, containers, or transport trolleys, are assigned to objects that are recorded on the map. If an object in the environment is detected by a laser scanner of an autonomous vehicle, the location of the autonomous vehicle on the map can be determined by comparing the detected object with an object recorded on the map.

[0004] DE 10 2021 000 349 A1 discloses a method for operating a technical installation that has at least one mobile system. A map of the technical installation is generated, which contains information about at least one accessible area and at least one restricted area.

[0005] It is problematic if a comparison of a detected object with an object shown on the map does not produce a clear result. Such a comparison can lead to an incorrect determination of the autonomous vehicle's location on the map.

[0006] A generic method for scan matching is known from the document "KONECNY, J., et al.: Scan Matching by Cross-Correlation and Differential Evolution, Electronics, 1 August 2019. Vol. 8, no. 8, 856. DOI: 10.3390 / electronics8080856".

[0007] The invention is based on the object of developing methods for operating an autonomous vehicle in an environment and an autonomous vehicle.

[0008] The object is achieved by a method for operating an autonomous vehicle in an environment having the features specified in claim 1. The object is also achieved by a method for operating an autonomous vehicle in an environment having the features specified in claim 2. Advantageous embodiments and further developments are the subject of the subclaims. The object is also achieved by an autonomous vehicle having the features specified in claim 14.

[0009] A first method for operating an autonomous vehicle in an environment is proposed. The autonomous vehicle has at least one laser scanner for taking laser scans to detect objects, and the autonomous vehicle has a map of the environment. The environment includes at least one object to which an object is assigned, which is listed on the map.

[0010] While the autonomous vehicle is moving along a trajectory in the environment, the at least one laser scanner of the autonomous vehicle acquires a laser scan of a part of the environment containing at least one object. The laser scan is compared with at least one object recorded on the map by calculating a correlation between objects detected in the laser scan and objects recorded on the map for several orientations of the laser scan. A pose of the autonomous vehicle in the map is determined based on the orientation of the laser scan for which the calculated correlation between the objects detected in the laser scan and the objects recorded on the map has the highest value.

[0011] Typically, there is no complete match between the objects detected in the laser scan and those recorded on the map. The laser scan orientation with the most matches has the highest calculated correlation compared to the other laser scan orientations. By determining the pose of the autonomous vehicle in the map based on this laser scan orientation, the probability of correctly determining the pose increases. The pose of the autonomous vehicle includes the location and orientation of the autonomous vehicle.

[0012] A second method for operating an autonomous vehicle in an environment is also proposed. The autonomous vehicle has at least one laser scanner for taking laser scans to detect objects, and the autonomous vehicle has a map of the environment. The environment includes at least one object, which is associated with an object listed on the map.

[0013] While the autonomous vehicle is moving along a trajectory in the environment, the at least one laser scanner of the autonomous vehicle acquires a laser scan of a part of the environment containing at least one object. The laser scan is compared with at least one object recorded on the map by calculating, for several orientations of the laser scan, a deviation between objects detected in the laser scan and objects recorded on the map. A pose of the autonomous vehicle in the map is determined based on the orientation of the laser scan for which the calculated deviation between the objects detected in the laser scan and the objects recorded on the map has the lowest value.

[0014] Typically, there is no complete match between the objects detected in the laser scan and those recorded on the map. The laser scan orientation with the most matches has the lowest calculated deviation compared to the other laser scan orientations. By determining the pose of the autonomous vehicle in the map based on this laser scan orientation, the probability of correctly determining the pose increases. The pose of the autonomous vehicle includes the location and orientation of the autonomous vehicle.

[0015] According to an advantageous embodiment of the invention, the laser scan orientations for which a correlation or deviation between objects detected in the laser scan and objects recorded on the map is calculated are each offset from one another by one angular step. The correlation or deviation is thus calculated for discrete laser scan orientations. This saves computing power and time.

[0016] If the objects captured by the laser scanner in the laser scan do not have any distinct features, especially sharp edges, it is difficult to find the laser scan orientation for which the calculated correlation has the highest value. Likewise, it is difficult to find the laser scan orientation for which the calculated deviation has the lowest value. The larger the angular steps between the individual laser scan orientations, the more difficult it is to find the laser scan orientation for which the calculated correlation has the highest value or for which the calculated deviation has the lowest value.The smaller the angular steps between the individual orientations of the laser scan, the higher the probability of finding the orientation of the laser scan for which the calculated correlation has the highest value or for which the calculated deviation has the lowest value.

[0017] According to an advantageous embodiment of the invention, the size of the angular step is specified for regions along the trajectory in the surrounding area. Typically, it is known in advance in which regions of the trajectory the objects detected in the laser scan do not exhibit any distinct features. If the vehicle is located in such an area, the angular step between two adjacent orientations of the laser scan is selected to be smaller. This increases the probability of finding the orientation of the laser scan for which the calculated correlation has the highest value or for which the calculated deviation has the lowest value.

[0018] According to an advantageous embodiment of the invention, the size of the angular step depends on a previously determined pose of the autonomous vehicle in the map. Typically, it is known in advance which poses of the autonomous vehicle do not reveal any distinct features on the objects detected in the laser scan. If the vehicle assumes such a pose, the angular step between two adjacent orientations of the laser scan is selected to be smaller. This increases the probability of finding the orientation of the laser scan for which the calculated correlation has the highest value or for which the calculated deviation has the lowest value.

[0019] According to an advantageous embodiment of the invention, the size of the angular step depends on a gradient of the previously calculated correlation. This increases the probability of finding the laser scan orientation for which the calculated correlation has the highest value. Alternatively, the size of the angular step depends on a gradient of the previously calculated deviation. This increases the probability of finding the laser scan orientation for which the calculated deviation has the lowest value.

[0020] Preferably, the angular step is selected to be smaller as the magnitude of the correlation gradient increases. This increases the probability of finding the laser scan orientation for which the calculated correlation has the highest value. Alternatively, the angular step is selected to be smaller as the magnitude of the deviation gradient increases. This increases the probability of finding the laser scan orientation for which the calculated deviation has the lowest value.

[0021] According to an advantageous development of the invention, the map comprises a plurality of cells. Cells occupied by an object are marked as occupied, and cells free of objects are marked as empty. For multiple orientations of the laser scan, a first partial correlation is calculated as the correspondence between objects detected in the laser scan and cells occupied by objects, and a second partial correlation is calculated as the correspondence between empty areas detected in the laser scan and empty cells. The correlation between the objects detected in the laser scan and the objects recorded in the map is calculated as the sum of the first partial correlation multiplied by a first weighting factor and the second partial correlation multiplied by a second weighting factor.

[0022] By appropriately selecting the weighting factors, it is possible, for example, to weight objects differently—in particular, more heavily—than empty cells when calculating the correlation. This can influence the probability of correctly determining the pose when calculating the correlation.

[0023] For example, the first weighting factor is larger than the second weighting factor, in particular between two and ten times as large. However, it is also conceivable that the first weighting factor is equal to the second weighting factor.

[0024] According to a preferred embodiment of the invention, a positive cell value is assigned to an occupied cell, a positive detection value is assigned to a detected object, a negative cell value is assigned to an empty cell, and a negative detection value is assigned to a detected empty area. The first partial correlation is calculated as the sum of the products of the detection values ​​of the detected objects with the cell values ​​of the assigned occupied cells, and the second partial correlation is calculated as the sum of the products of the detection values ​​of the detected empty areas with the cell values ​​of the assigned empty cells. This further increases the probability of finding the orientation of the laser scan for which the calculated correlation has the highest value.

[0025] According to an alternative advantageous development of the invention, for multiple orientations of the laser scan, multiple partial deviations are calculated as the difference between the objects detected in the laser scan and an object recorded on the map. The deviation between the objects detected in the laser scan and the objects recorded on the map is then calculated from the partial deviations. The partial deviations correspond to the distances of the detected objects to the objects recorded on the map.

[0026] According to an advantageous embodiment of the invention, the deviation of the objects detected in the laser scan with the objects recorded in the map is calculated as the sum of the amounts of the previously calculated partial deviations.

[0027] According to an alternative advantageous embodiment of the invention, the deviation between the objects detected in the laser scan and the objects recorded on the map is calculated as the sum of the squares of the previously calculated partial deviations. The partial deviations correspond to the distances of the detected objects to the objects recorded on the map. The deviation between the objects detected in the laser scan and the objects recorded on the map thus corresponds to the sum of the squares of the said distances.

[0028] An autonomous vehicle according to the invention comprises at least one laser scanner for recording laser scans to detect objects, wherein the autonomous vehicle has a map of its surroundings. The autonomous vehicle can be operated using one of the methods according to the invention.

[0029] The invention is not limited to the combination of features in the claims. Further possible combinations of claims and / or individual claim features and / or features of the description and / or the figures will become apparent to those skilled in the art, particularly from the problem and / or the problem posed by comparison with the prior art.

[0030] The invention will now be explained in more detail with reference to the accompanying drawings. The invention is not limited to the exemplary embodiments shown in the drawings. The drawings only represent the subject matter of the invention schematically. They show: Fig. 1: a schematic representation of a map of an environment, Fig. 2: a graphical representation of a calculated correlation depending on an orientation of a laser scan according to a first example and Fig. 3: a graphical representation of a calculated correlation depending on an orientation of a laser scan according to a second example.

[0031] Fig. Figure 1 shows a schematic representation of a map of an environment. The environment is, for example, a technical facility, such as a production plant, an industrial hall, or a logistics center. It is also conceivable, for example, that the environment represents part of a town or a residential area. The environment includes several autonomous vehicles 1. Only one such autonomous vehicle 1 is shown in the present illustration. The autonomous vehicles 1 are used, in particular, to transport materials within the environment.

[0032] The autonomous vehicle 1 has a drive device, an electrical energy storage device for supplying the drive device, and a control unit for controlling the drive device. Furthermore, the autonomous vehicle 1 has a communication device for wireless communication with other autonomous vehicles 1 and with a higher-level server. The communication device of the autonomous vehicle 1 is designed, for example, for data transmission via WLAN, Bluetooth, or light.

[0033] The autonomous vehicle 1 has two laser scanners 2. The laser scanners 2 are used to capture laser scans to detect objects in the surrounding area. When an object is detected, the laser scanners 2 record the distance to the object and the direction in which the object is located. The laser scanners are mounted at opposite corners of the autonomous vehicle 1 and each detect objects within an angular range of approximately 270°.

[0034] The environment includes several objects. Examples of such objects are walls, columns, and production machines. Objects generally remain in a fixed location in the environment and are therefore not moved. Each object is assigned an object 4. The objects 4 are listed on the map.

[0035] The map comprises a plurality of cells. Cells occupied by an object 4 are marked as occupied. Cells that are free of objects 4 are marked as empty. The cells of the map are square and arranged two-dimensionally, directly adjacent to one another. In this case, one side length of a cell corresponds to a distance of approximately 2 cm to 10 cm in the surrounding area.

[0036] Autonomous vehicle 1 has a pose in the environment. The pose of autonomous vehicle 1 includes the location and orientation of autonomous vehicle 1 in the environment. The pose of autonomous vehicle 1 is determined using laser scans recorded by laser scanners 2.

[0037] The autonomous vehicle 1 moves along a trajectory in the environment. While the autonomous vehicle 1 moves along the trajectory in the environment, the laser scanners 2 of the autonomous vehicle 1 each acquire a laser scan of a portion of the environment. Such a laser scan extends over an angular range of approximately 270°.

[0038] The recorded part of the environment contains at least one object. The objects located in that part of the environment are recorded in the laser scan. The laser scan is compared with the objects 4 recorded on the map.

[0039] For multiple orientations A of the laser scan, a correlation K is calculated between the objects detected in the laser scan and the objects 4 recorded on the map. An orientation A of the laser scan corresponds to an angle relative to a defined reference axis. The reference axis is, for example, a longitudinal axis of the autonomous vehicle 1. The orientations A extend within the angular range of the laser scan.

[0040] The individual orientations A of the laser scan, for each of which a correlation K is calculated between the objects captured in the laser scan and the objects 4 recorded on the map, are each offset from one another by one angular step. The size of this angular step is variable. A relatively large angular step enables a relatively fast calculation of the correlation K with relatively low required computing power. A relatively small angular step enables a relatively precise calculation of the correlation K with relatively high required computing power.

[0041] The map contains a number of cells. Cells occupied by an object 4 are marked as occupied. Cells free of objects 4 are marked as empty.

[0042] For each of the multiple orientations A of the laser scan, a first partial correlation is calculated. The first partial correlation is calculated as the correspondence between the objects detected in the laser scan and the cells occupied by objects 4. The first partial correlation is multiplied by a first weighting factor to produce a weighted first partial correlation.

[0043] For each orientation A of the laser scan for which a first partial correlation is calculated, a second partial correlation is also calculated. The second partial correlation is calculated as the correspondence between empty areas captured in the laser scan and empty cells. The second partial correlation is multiplied by a second weighting factor to produce a weighted second partial correlation.

[0044] For the orientations A of the laser scan, for which a first partial correlation and a second partial correlation are calculated, the first weighted partial correlation and the second weighted partial correlation are added to a correlation K. The correlation K of the objects detected in the laser scan with the objects 4 recorded in the respective part of the map is calculated as the sum of the first weighted partial correlation and the second weighted partial correlation.

[0045] The correlation K of the objects detected in the laser scan with the objects 4 recorded in the map is thus calculated as the sum of the first partial correlation multiplied by the first weighting factor and the second partial correlation multiplied by the second weighting factor.

[0046] The first weighting factor is generally larger than the second weighting factor, in particular between two and ten times as large. When calculating the correlation K between the objects detected in the laser scan and the objects 4 recorded on the map, the correspondence between the objects detected in the laser scan and the cells occupied by objects 4 is given greater consideration than the correspondence between empty areas detected in the laser scan and empty cells.

[0047] It is also conceivable that the first weighting factor is equal to the second weighting factor. When calculating the correlation K between the objects detected in the laser scan and the objects 4 recorded on the map, the correspondence between the objects detected in the laser scan and the cells occupied by objects 4 is then given equal weight as the correspondence between empty areas detected in the laser scan and empty cells.

[0048] It is also conceivable that the first weighting factor is greater than zero and the second weighting factor is equal to zero. When calculating the correlation K between the objects detected in the laser scan and the objects 4 recorded on the map, only the correspondence between the objects detected in the laser scan and the cells occupied by objects 4 is taken into account, and the correspondence between the empty areas detected in the laser scan and empty cells is not taken into account.

[0049] Furthermore, it is also conceivable that a positive cell value is assigned to an occupied cell, and that a positive detection value is assigned to a detected object, and that a negative cell value is assigned to an empty cell, and that a negative detection value is assigned to a detected empty area.

[0050] The first partial correlation is calculated as the sum of the products of the detection values ​​of the detected objects with the cell values ​​of the assigned occupied cells, and the second partial correlation is calculated as the sum of the products of the detection values ​​of the detected empty areas with the cell values ​​of the assigned empty cells.

[0051] A match between a detected object and an occupied cell provides a positive contribution to the first partial correlation. However, a match between a detected object and an empty cell provides a negative contribution to the first partial correlation.

[0052] A match between a detected empty area and an empty cell provides a positive contribution to the second partial correlation. However, a match between a detected empty area and an occupied cell provides a negative contribution to the second partial correlation.

[0053] As already mentioned, said correlation is calculated for several orientations A of the laser scan, with said orientations A of the laser scan each offset by one angular step. The pose of the autonomous vehicle 1 in the map is determined based on the orientation A of the laser scan for which the calculated correlation K of the objects detected in the laser scan with the objects 4 recorded in the map has the highest value.

[0054] Fig. Figure 2 shows a graphical representation of a calculated correlation K as a function of an orientation A of a laser scan according to a first example. As already mentioned, the correlation K is calculated for discrete orientations A of the laser scan. However, here the correlation K is shown as a continuous curve.

[0055] The correlation K curve in the first example exhibits several local maxima, with one of the local maxima simultaneously representing a global maximum. The pose of autonomous vehicle 1 in the map is determined based on the orientation A of the laser scan to which the global maximum of the correlation K is assigned.

[0056] The orientations A of the laser scan for which the correlation K is calculated are initially offset from each other by a constant, relatively large angular step. Around the local maxima of the correlation K, the magnitude of the gradient of the correlation K increases. The gradient of the correlation K corresponds to a gradient or a derivative of the course of the correlation K with respect to the orientation A of the laser scan and is calculated, for example, using gradient triangles.

[0057] If it is detected that the magnitude of the gradient of the correlation K is increasing, the angular step between two adjacent orientations A of the laser scan is chosen to be smaller. The size of the angular step thus depends on the gradient of the previously calculated correlation K.

[0058] Fig. Figure 3 shows a graphical representation of a calculated correlation K as a function of an orientation A of a laser scan according to a second example. As already mentioned, the correlation K is calculated for discrete orientations A of the laser scan. However, here the correlation K is shown as a continuous curve.

[0059] The correlation K curve in the second example exhibits several local maxima, with one of the local maxima simultaneously representing a global maximum. The pose of autonomous vehicle 1 in the map is determined based on the orientation A of the laser scan to which the global maximum of the correlation K is assigned.

[0060] Compared to the Fig. In the first example shown in Figure 2, several local maxima are located relatively close to one another in the second example shown here, with one of the local maxima simultaneously representing the global maximum. Such a situation occurs, for example, when the objects detected by the laser scanners 2 in the laser scans do not have any distinct features, in particular no sharp edges.

[0061] Typically, it is known in advance in which areas of the trajectory in the environment along which the autonomous vehicle 1 is moving such situations occur. If the vehicle is located in such an area, the angular step between two adjacent orientations A of the laser scan is selected to be smaller. The size of the angular step is predetermined for areas along the trajectory in the environment.

[0062] Typically, the poses of autonomous vehicle 1 in the map at which such situations occur are also known in advance. When the vehicle assumes such a pose, the angular step between two adjacent orientations A of the laser scan is chosen to be smaller. The size of the angular step thus depends on a previously determined pose of autonomous vehicle 1 in the map.

[0063] Similar to the one in Fig.In the first example shown in Figure 2, an amount of the gradient of the correlation K increases around the local maxima of the correlation K. The gradient of the correlation K corresponds to a gradient or a derivative of the course of the correlation K according to the orientation A of the laser scan and is calculated, for example, using gradient triangles.

[0064] If it is detected that the magnitude of the gradient of the correlation K is increasing, the angular step between two adjacent orientations A of the laser scan is chosen to be smaller. The size of the angular step thus depends on the gradient of the previously calculated correlation K. List of reference symbols 1 autonomous vehicle 2 laser scanners 4 A Object Alignment K Correlation QUOTES CONTAINED IN THE DESCRIPTION

[0000] This list of documents submitted by the applicant was generated automatically and is included solely for the convenience of the reader. This list is not part of the German patent or utility model application. The DPMA assumes no liability for any errors or omissions. Cited patent literature

[0000] DE 10 2021 000 349 A1

[0004] Cited non-patent literature

[0000] KONECNY, J., et al.: Scan Matching by Cross-Correlation and Differential Evolution, Electronics, August 1, 2019. Vol. 8, no. 8, 856. DOI: 10.3390 / electronics8080856

[0006]

Claims

[1] Method for operating an autonomous vehicle (1) in an environment, wherein the autonomous vehicle (1) has at least one laser scanner (2) for recording laser scans for detecting objects, and wherein the autonomous vehicle (1) has a map of the environment, and wherein the environment comprises at least one object, to which an object (4) is assigned, which is shown on the map, characterized by , that during a movement of the autonomous vehicle (1) along a trajectory in the environment a laser scan of a part of the environment which has at least one object is recorded by the at least one laser scanner (2) of the autonomous vehicle (1); the laser scan is compared with at least one object (4) recorded on the map by for multiple orientations (A) of the laser scan a correlation (K) of objects detected in the laser scan with objects recorded in the map (4) is calculated; a pose of the autonomous vehicle (1) in the map is determined based on the orientation (A) of the laser scan, for which the calculated correlation (K) of the objects detected in the laser scan with the objects recorded in the map (4) has the highest value. [2] Method for operating an autonomous vehicle (1) in an environment, wherein the autonomous vehicle (1) has at least one laser scanner (2) for recording laser scans for detecting objects, and wherein the autonomous vehicle (1) has a map of the environment, and wherein the environment comprises at least one object, to which an object (4) is assigned, which is shown on the map, characterized by , that during a movement of the autonomous vehicle (1) along a trajectory in the environment a laser scan of a part of the environment which has at least one object is recorded by the at least one laser scanner (2) of the autonomous vehicle (1); the laser scan is compared with at least one object (4) recorded on the map by for multiple orientations (A) of the laser scan a deviation of objects detected in the laser scan with objects recorded in the map (4) is calculated; a pose of the autonomous vehicle (1) in the map is determined based on the deviation of the laser scan, for which the calculated deviation of the objects detected in the laser scan with the objects recorded in the map (4) has the lowest value. [3] Method according to one of claims 1 or 2, characterized by that the alignments (A) of the laser scan, for each of which a correlation (K) or a deviation of objects detected in the laser scan with objects recorded on the map (4), are offset from each other by one angular step. [4] Method according to claim 3, characterized by that a size of the angular step is specified for areas along the trajectory in the environment. [5] Method according to one of claims 3 to 4, characterized by that a size of the angular step depends on a previously determined pose of the autonomous vehicle (1) in the map. [6] Method according to one of the preceding claims, characterized by that a size of the angular step from a gradient of the previously calculated correlation (K) or depends on a gradient of the previously calculated deviation. [7] Method according to claim 6, characterized by that the angle step is chosen smaller, when an amount of the gradient of the correlation (K) increases, or when an amount of the gradient of the deviation becomes larger. [8] Method according to one of the preceding claims, characterized by that the map comprises a plurality of cells, and that Cells occupied by an object (4) are marked as occupied, and cells free of objects (4) are marked as empty, and for multiple orientations (A) of the laser scan in each case, a first partial correlation is calculated as the correspondence of objects detected in the laser scan with cells occupied by objects (4); and in each case, a second partial correlation is calculated as the correspondence of empty areas detected in the laser scan with empty cells; and in each case the correlation (K) of the objects recorded in the laser scan with the objects recorded in the map (4) as a sum the first partial correlation multiplied by a first weighting factor and the second partial correlation multiplied by a second weighting factor. [9] Method according to claim 8, characterized by , that the first weighting factor is greater than the second weighting factor, especially between twice and ten times as large. [10] Method according to one of claims 8 to 9, characterized by that a positive cell value is assigned to an occupied cell, and that a positive detection value is assigned to a detected object, and that a negative cell value is assigned to an empty cell, and that a negative detection value is assigned to a detected empty area, and that the first partial correlation is calculated as the sum of the products of the detection values ​​of the detected objects with the cell values ​​of the associated occupied cells; and that the second partial correlation is calculated as the sum of the products of the detection values ​​of the detected empty areas with the cell values ​​of the associated empty cells. [11] Method according to one of the preceding claims, characterized by that for several orientations (A) of the laser scan several partial deviations are calculated as the difference between the objects recorded in the laser scan and an object (4) recorded in the map; and that the deviation of the objects recorded in the laser scan with the objects recorded in the map (4) is calculated from the partial deviations. [12] Method according to claim 11, characterized by , that the deviation of the objects recorded in the laser scan with the objects recorded in the map (4) is calculated as the sum of the amounts of the partial deviations. [13] Method according to claim 11, characterized by , that the deviation of the objects recorded in the laser scan with the objects recorded in the map (4) is calculated as the sum of the squares of the partial deviations. [14] Autonomous vehicle (1), comprising at least one laser scanner (2) for recording laser scans for detecting objects, wherein the autonomous vehicle (1) has a map of an environment, and wherein the autonomous vehicle (1) can be operated in the environment using one of the methods according to one of the preceding claims.

Images