Method for identifying and determining geometric properties of an object and driver assistance system

Abstract

The invention proposes a method for recognizing and determining geometric properties of objects (2) using ultrasound, wherein ultrasound pulses (20) are emitted and ultrasound echoes (31, 43) reflected at the objects (2) are received. An n-dimensional grid map is created for each object type. Each cell represents a parametric object with geometric parameters attributed to the cell and has a counter. Upon reception of a direct echo (31) the counter of the cell representing the parametric object whose surface is tangent to a circular arc (51) or a spherical surface defined by the distance attributed to the direct echo is incremented. Upon reception of a cross echo (43) the counter of the cell representing the parametric object whose surface is tangent to an elliptic arc (63) or an ellipsoidal surface defined by the distance attributed to the cross echo (43) and the relative position of the ultrasound sensor (11, 12, 13, 14) involved is incremented. For each object type candidate objects are determined. The invention also relates to a computer program and to a driver assistance system.

Description

Technical Field

[0001] This invention relates to a method for identifying and determining the geometric characteristics of an object using ultrasound, wherein ultrasonic pulses are emitted and the ultrasonic echoes reflected at the object are received again by at least two ultrasonic sensors arranged spaced apart from each other. Other aspects of the invention relate to a computer program and a driver assistance system configured to implement the method. Background Technology

[0002] Different driver assistance systems are known that assist a vehicle driver in performing various driving maneuvers, provide advance warnings of dangers, and / or at least temporarily guide the vehicle automatically. To achieve these functions, these driver assistance systems rely on accurate data about objects in the surrounding environment (such as other road users and obstacles like trees or pillars). Accurately determining the position of these objects relative to the vehicle is particularly important.

[0003] The least squares method can be used to determine the location of an object. In the least squares method, the distance between each sensor and the object is determined using multiple sensors that are spatially spaced apart. For example, ultrasonic distance sensors can be used for location determination based on the least squares method.

[0004] When multiple objects are present in the surrounding environment, the challenge lies in assigning the measured distances to the correct objects using the least squares method, ensuring that the correct circles or ellipses are tangent to each other. This primarily occurs in scenes with a large number of different objects, due to incorrect tangency points leading to incorrect locations or objects being inferred from places where no objects are present. The more sensors used, the more complex the selection of circles and ellipses to be tangent to each other becomes.

[0005] DE 10 2013 223 803 A1 describes a method for segmenting an occupancy grid for an environment model. Here, in addition to occupancy level, each grid cell is associated with at least one additional piece of information about that grid cell, such as velocity or height. The method includes reading in an occupancy grid having multiple grid cells, each of which is associated with grid cell information, and, using that grid cell information, assigning at least one object to the multiple grid cells.

[0006] A system and control device for an autonomous vehicle are known from EP 3 552 908 A2. The system includes a sensor module, which may in particular include ultrasonic sensors. To assess the vehicle's surrounding environment, a grid map and a list of dynamic objects can be used.

[0007] DE 10 2015 122 825 A1 describes a method for grouping target elements for object fusion. Here, an xy space is represented by a grid, in which sensor measurement points are shown. The grid is used to group sensor measurements, for example, by assigning them to adjacent cells.

[0008] The problem with the known method is that it does not classify objects or determine object size. Summary of the Invention

[0009] A method is proposed for identifying objects and determining their geometric properties using ultrasound, wherein ultrasonic pulses are emitted and the ultrasonic echoes reflected at the object are received by at least two ultrasonic sensors arranged spaced apart from each other. In this method, object types are defined with separately defined sets of geometric parameters, including the object's location, and an n-dimensional grid map is created for each object type, where n is the number of parameters associated with the object type. Here, each cell represents a parametric object having geometric parameters associated with the corresponding cell, and each cell in the grid map has a counter associated with that cell.

[0010] In the case of a direct echo, the counter of a cell representing a parameter object in each grid map is incremented if the surface of that parameter object is tangent to a circular or spherical surface, defined by the distance associated with the direct echo. In the case of a cross echo, the counter of a cell representing a parameter object is incremented if the surface of that parameter object is tangent to an elliptical or ellipsoidal surface, defined by the distance associated with the cross echo and the relative position of the participating ultrasonic sensors. When determining whether tangency exists, tolerances are considered for measurement error, scattering, and / or the parameter range associated with the cell. Specifically, the tolerances are chosen such that the scattering behavior of the reflecting surface is taken into account. Such a "scattering" surface does not have tangential reflection behavior applicable to "an incident angle equal to a reflection angle" and resulting in tangential contact when checking for tangency. Alternatively, scattering within a certain angular range, which is also identified as tangential through appropriate tolerance selection, is also considered tangential.

[0011] In addition, candidate objects are determined for each object type, and for each candidate object, the corresponding raster map cell has a counter higher than a pre-given boundary value.

[0012] The object can be, for example, a traffic participant such as a vehicle or pedestrian, or an obstacle such as a tree or pillar.

[0013] To identify objects and determine their geometric properties, at least one of the ultrasonic sensors can, for example, emit an ultrasonic pulse, wherein two or more ultrasonic sensors can receive ultrasonic echoes reflected by objects in the surrounding environment. Here, the ultrasonic echo received by the ultrasonic sensor that originally emitted the ultrasonic pulse is called a direct echo, while the ultrasonic echo received by one or more other ultrasonic sensors is called a cross echo. The distance to the object can be determined by the propagation time from the emission of the ultrasonic pulse to the reception of the ultrasonic echo and by the known speed of sound. Depending on the situation, especially when multiple objects are present in the surrounding environment, multiple ultrasonic echoes may be received by a single ultrasonic sensor for the emitted ultrasonic pulse.

[0014] Each ultrasonic sensor can emit ultrasonic pulses sequentially. Alternatively, two or more ultrasonic sensors can be configured to emit ultrasonic pulses simultaneously, wherein the ultrasonic pulses are preferably encoded for differentiation. For encoding, the ultrasonic pulses can be modulated and / or different frequencies can be used.

[0015] A set of geometric parameters is predefined for each object type to be identified, which describes each valid object of that type. These parameters at least describe the object's position and may additionally include other specifications, such as geometric dimensions like diameter, length, height, and / or width. The position-describing parameters do not necessarily have to be specified in Cartesian coordinates; they can also be specified in other coordinate systems, such as polar coordinates. Depending on the object type, the position-describing parameters can be one-dimensional, two-dimensional, or three-dimensional, i.e., including one to three parameters to explicitly describe the position.

[0016] For example, a wall in three-dimensional space can be represented as a plane (with an infinitely extending scale), where the plane can be explicitly represented by points in a three-dimensional parameter space (n=3). If only vertical walls are considered, even two parameters and a corresponding two-dimensional parameter space (n=2) are sufficient for description. Possible choices for the two parameters or dimensions of this parameter space include the distance from the plane to the origin and the angle between the projection of the plane's normal vector onto the xy-plane and the x-axis. A vertical column, such as a pillar or support, can be represented by its x-coordinate and its y-coordinate, and therefore can also be represented by a two-dimensional parameter space (n=2). If, for example, the radius of the column should be additionally considered, the parameter space is correspondingly extended to three parameters (n=3).

[0017] Accordingly, in this method, the defined object type is preferably selected from planes (especially vertical planes), cylinders (especially vertical cylinders), spheres, and lines (especially horizontal lines). It is preferable to define at least two different object types and create at least two raster maps accordingly.

[0018] For example, a sphere object can be used with a four-dimensional parameter description (center point position and radius). Alternatively, a point object can be derived from a three-dimensional parameter description without a radius.

[0019] Line objects, in particular, can represent one-dimensional boundaries. A horizontal line object, for example, can describe the position of a curbstone, where arbitrary three-dimensional and four-dimensional parametric descriptions can be implemented.

[0020] When defining the object type, the surface scattering behavior of the reflective surface can already be considered. For this purpose, object types such as planes and cylinders can be used as a basis, where the parameter descriptions themselves are preferably not changed. For example, objects with uneven or irregular surfaces (such as shrubs or bicycles) can be modeled as planes with diffuse reflection. However, when selecting the raster cells to be filled, the conditions for tangential reflection behavior (checking tangency) are softened in such a way that a larger tolerance is allowed for the deviation between the angle of incidence and the angle of reflection. This results in additional cells being filled by incrementing a counter, for example, if this involves a raster map for shrubs. Correspondingly, it is preferable to pre-define the tolerances used for checking tangency differently for different object types.

[0021] It is preferable to use different raster maps for object types with tangential reflection behavior and for object types with diffuse reflection behavior (e.g., walls relative to shrubs), even if these object types are described using the same parameters.

[0022] Each cell in a raster map represents a parametric object with parameters assigned to that corresponding cell. For the example of thin vertical cylinders, where the radius is not considered as a parameter, each cell corresponds to a thin vertical cylinder at its assigned X / Y position in the surrounding environment. In this example, the raster map could correspondingly represent possible vertical cylinders within an environmental area having, for example, dimensions of 2.5m × 2.5m (length × width), using, for example, a 5cm grid. For this example, a raster map with 50 × 50 = 2500 cells is obtained.

[0023] The size and division of the raster map can be fixedly pre-defined or dynamically matched.

[0024] It is preferable to select the size of the raster map in such a way that the parameter objects can be described in the raster map in length and width (road plane) ranging from 1m to 10m, preferably from 2m to 5m, and in height ranging from 0m to 5m, preferably from 1m to 2m.

[0025] When receiving ultrasonic echoes, it is determined which of the objects represented by the grid map the ultrasonic echoes may originate from. Here, the elliptic of revolution E corresponding to the distance values ​​'a' of the obtained direct and cross echoes is determined. a And for each object type T, it is determined whether the object represented by the unit parameter is related to the elliptic body E of the ultrasonic echo. a They share a tangential hyperplane (in addition to measurement tolerances).

[0026] The projection of an ellipsoid of revolution in Cartesian space is a circle for direct echoes and an ellipse for cross echoes in two dimensions. In three-dimensional viewing, the ellipsoid of revolution is a sphere for direct echoes and an ellipsoid with at least two identical axes for cross echoes. However, depending on the object type, the spatial dimensions of a raster map can be defined with deviations from Cartesian coordinates, for example, in the case of vertical walls, as the distance to the origin and the angle between the normal vector and the X-axis in the XY plane. In this case, transformations can be performed to determine the point of tangency or a common hyperplane.

[0027] This transformation can be implemented similarly to the Hough transform. The Hough transform is particularly known to those skilled in the art of recognizing complex parameterizable geometries in images. For example, the representation of an elliptic is transformed into the corresponding parameter space of the corresponding raster map. The counters of raster map cells that are tangent to the transformed elliptic are then incremented. In Cartesian space, this corresponds to incrementing the counters of cells representing objects whose surfaces are tangent to the elliptic within a pre-defined tolerance.

[0028] Preferably, before incrementing the counter of a cell, it is checked whether the parameter object represented by the cell is located within a region defined by the field of view of the ultrasonic sensor that has received the ultrasonic echo. Here, for example, the field of view can be defined by a cone, where it can be checked, for example, whether the object represented by the corresponding cell is tangent to the volume of the cone.

[0029] If the cell counters increment by 1, then in the case of a 2D raster map (n=2), the pre-defined boundary value is preferably ≥1, while in the case of a 3D raster map (n=3), it is preferably ≥2. In higher dimensions, the boundary value is preferably increased accordingly to the number of dimensions. A higher boundary value results in a lower probability of recognizing an object, even if no real object exists at that location. Conversely, increasing the boundary value decreases sensitivity, thus increasing the probability of not recognizing a real object as an object.

[0030] This boundary value can be fixedly pre-defined or dynamically matched. For example, if a particularly large number of echoes are measured within a measurement cycle, the boundary value for that measurement cycle can be increased by, for example, by 1.

[0031] Preferably, the unit is assigned a corresponding received ultrasonic echo as the counter is incremented. This can be done, for example, by storing a reference to a set of data representing the received ultrasonic echoes assigned to the corresponding unit. Preferably, it is stored which of the ultrasonic sensors have received the corresponding ultrasonic echoes, and preferably other attributes, such as distance determined by the propagation time of the ultrasonic echo and / or the amplitude of the ultrasonic echo, are stored.

[0032] Preferably, the order of candidate objects is determined based on classification parameters, wherein the ranking parameters are selected from a combination of at least two of the following: object type, distance from the candidate object to the ultrasonic sensor, number of ultrasonic echoes associated with the candidate object, quality of the associated ultrasonic echoes, and ranking parameters. As indicators of ultrasonic echo quality, amplitude and / or signal-to-noise ratio can be used, for example. Preferably, the number of echoes associated with the candidate object is used as the most important parameter for classification and weighted accordingly.

[0033] Preferably, candidate objects are examined according to the determined order in the following aspects: whether the amount of ultrasound echoes associated with the candidate object is significant. Candidate objects with a significance value higher than a pre-given threshold are identified as target objects, wherein the number of target objects to which the corresponding ultrasound echoes associated with the candidate object have been assigned is checked as a parameter used to determine significance.

[0034] For example, an evaluation is performed for each ultrasonic echo assigned to a candidate object, where the more target objects an ultrasonic echo has been assigned to, the lower its evaluation score. The significance value of a candidate object specifically incorporates the sum of the evaluations of each assigned ultrasonic echo. Especially when the probability of the presence of a real object has been established for the echoes using the ultrasonic sensor itself, this probability of presence is preferably considered, for example, as a weighting factor.

[0035] To test significance, other criteria may be considered when determining the significance value, in particular from the object type and the number of ultrasound echoes belonging to the candidate object.

[0036] Once the target object is identified, it is explicitly characterized by parameters assigned to the corresponding cell. Furthermore, the target object is also classified as an object of that type using a raster map assigned to it.

[0037] Ideally, the target object, along with its defined parameters and classification, should be used together to provide driver assistance functions. To this end, the target object can be input into an environment map that is consistent across all object types.

[0038] The identified target object is then studied in more detail, thereby enabling a more accurate determination of its parameters. For this purpose, it is preferable to reduce the step size of the corresponding raster map within the target object's region to improve the accuracy of determining the target object's geometric characteristics.

[0039] Alternatively or additionally, to more accurately verify the target object, the description of the target object's location derived from the parameters of the corresponding cell is used as a pre-selection for subsequent determination of the target object's location when using the least squares method. This is particularly true if such a cell is associated with a received ultrasonic echo, allowing the selection of the ultrasonic echo or the distance associated with it for performing the least squares method. The location determination accuracy achieved by this least squares method can be superior to that of a grid map.

[0040] Preferably, the ultrasonic echoes are removed from the grid map after a predetermined minimum time (e.g., 0.5 s) and / or after traveling a minimum distance (e.g., 0.1 m), wherein the counter of the corresponding cell is decremented. For this purpose, it is preferable to assign a reference to the corresponding ultrasonic echo to one of these cells while incrementing the cell's counter.

[0041] Alternatively, after a measurement cycle or a predetermined number of measurement cycles, the counters of all cells in all raster maps can be reset, specifically by setting all cell counters to a neutral value of "0".

[0042] Preferably, a measurement cycle is considered complete after the ultrasonic sensor has emitted an ultrasonic pulse and received the corresponding ultrasonic echo, or after a maximum predetermined waiting time for receiving ultrasonic echoes has elapsed. For example, one ultrasonic sensor may emit a signal in one measurement cycle, while up to 12 ultrasonic sensors may receive the signal. Since each ultrasonic sensor is capable of receiving multiple ultrasonic echoes, for example up to 20 ultrasonic echoes, it is possible to receive up to 240 echoes in a single measurement cycle.

[0043] The proposed method is preferably repeated continuously, so that the object position is obtained or updated continuously. The obtained target object and its determined position are preferably used to provide driver assistance functions, or to be used by other driver assistance systems that provide driver assistance functions. The driver assistance function may be, for example, parking assistance, which automatically guides the vehicle to a parking space.

[0044] According to the invention, a computer program is also provided, according to which the methods described herein are implemented when executed on a programmable computer device. The computer program may, for example, be a module for implementing a driver assistance system or its subsystems in a vehicle. The computer program may be stored on a machine-readable storage medium, such as a permanent or rewritable storage medium, or attached to a computer device, or stored on a removable CD-ROM, DVD, Blu-ray disc, or USB strip. Alternatively or additionally, the computer program may be provided on the computer device (e.g., on a server) for download, for example via a data network (such as the Internet) or via a communication connection (e.g., a telephone line or wireless connection).

[0045] Furthermore, according to the present invention, a driver assistance system for determining the position of objects in the environment surrounding a vehicle is provided, wherein the driver assistance system includes a plurality of ultrasonic sensors. Here, the driver assistance system is configured to implement the method described herein.

[0046] This invention particularly enables efficient identification of objects and determination of their geometric characteristics when using ultrasonic sensors that receive echoes from multiple objects. Direct object classification is achieved by using a grid map for each type of object to be identified. Furthermore, by using a grid map, the computational overhead required to determine object characteristics increases only slightly with the increase in the number of sensors and / or objects in the surrounding environment, thus allowing the method to be implemented resource-efficiently. This is especially true for implementation in typical control devices used to operate driver assistance systems.

[0047] Advantageously, the determined object characteristics can include not only its location, but also, for example, its size. Therefore, the method is capable not only of object identification, but also of classification and measurement.

[0048] Unlike current ultrasonic sensor-based localization methods, the method according to the present invention can locate not only the nearest objects individually, but also multiple objects in the surrounding environment. This method enables localization by dividing the ultrasonic echoes of a measurement cycle into more groups than by grouping only the ultrasonic echoes matched to the nearest object. Attached Figure Description

[0049] The embodiments of the present invention will now be described in more detail with reference to the accompanying drawings.

[0050] The attached diagram shows:

[0051] Figure 1 : A schematic diagram of the transmission and reception of ultrasonic waves used to identify objects in the environment surrounding a vehicle;

[0052] Figure 2 A schematic diagram of the first example of a grid map of the vehicle's surroundings;

[0053] Figure 3 A schematic diagram of a second example of a grid map in two-dimensional parametric space on a vertical plane extending 3m in front of the vehicle, and

[0054] Figure 4 : A schematic diagram of the field of view of the ultrasonic sensor in a vehicle. Detailed Implementation

[0055] In the following description of embodiments of the invention, the same reference numerals are used to denote the same or similar elements, and in some cases, repeated descriptions of said elements are omitted. The drawings are for illustrative purposes only, showing the subject matter of the invention.

[0056] Figure 1 The front of a vehicle 1 with a driver assistance system 300 is shown, which is used to identify and determine the geometric characteristics of an object 2 in the environment surrounding the vehicle 1.

[0057] exist Figure 1 In the illustrated embodiment, the driver assistance system 300 includes four ultrasonic sensors 11, 12, 13, and 14, all of which are arranged at the front of the vehicle 1. For example, the ultrasonic sensors 11, 12, 13, and 14 can be arranged in the bumper of the vehicle 1. The driver assistance system 300 further includes a control device 100 connected to the ultrasonic sensors 11, 12, 13, and 14. The control device 100 is configured to manipulate the connected ultrasonic sensors 11, 12, 13, and 14 to emit ultrasonic pulses 20 and to process the received ultrasonic echoes 31 and 43.

[0058] exist Figure 1In the scenario shown, object 2 is located in front of vehicle 1. Obviously, there could be more than one object 2 in the surrounding environment, but for clarity, the process of this method is described with respect to a single object 2. Furthermore, the following exemplary description of this method illustrates the process for identifying and determining the characteristics of the thin vertical column that is object 2. Object 2 is a vertical columnar support.

[0059] To identify object 2 in the surrounding environment and determine its location, ultrasonic sensors 11, 12, 13, and 14 emit ultrasonic pulses 20 and receive ultrasonic echoes 31 and 43 reflected by object 2. For clarity, in Figure 1 Only the emission of ultrasonic pulse 20 via the first ultrasonic sensor 11 is shown. Clearly, the other ultrasonic sensors 12, 13, and 14 are also capable of emitting ultrasonic pulse 20.

[0060] Ultrasonic pulse 20 is reflected by object 2. Here, the ultrasonic echo received by the first ultrasonic sensor 11 is referred to as direct echo 31, because the original ultrasonic pulse 20 was emitted by the first ultrasonic sensor 11. The ultrasonic echoes received by the other sensors 12, 13, and 14 are referred to as cross echo 43, wherein, for simplicity... Figure 1 The indication only indicates that the cross echo 43 is received by the third ultrasonic sensor 13.

[0061] Based on the propagation time from the emission of the ultrasonic pulse 20 to the reception of the direct echo 31, and based on the known speed of sound, the distance from the object 2 to the first ultrasonic sensor 11 can be determined. A distance can also be assigned to the cross echo 43 based on the propagation time.

[0062] The proposed method is configured to create a raster map 200 for each object type to be identified, see [link to relevant documentation]. Figure 2 and Figure 3 When a thin vertical column is used as the object type to be identified, two parameters are sufficient to clearly describe object 2: its X and Y coordinates. Accordingly, a raster map 200 is created as follows: this raster map is two-dimensional and has an X coordinate as the first dimension and a Y coordinate as the second dimension.

[0063] In the case of direct echo 31, the radius of arc 51 is determined based on the distance obtained from the propagation time. This arc describes the possible X / Y values ​​of the coordinates of object 2. Here, the first ultrasonic sensor 11 is located at the center of arc 51. In the grid map 200, a counter associated with cell 210 is incremented accordingly, the counter describing the vertical column at the location intersected by arc 51.

[0064] In the case of cross echo 43, the possible X / Y values ​​of the coordinates of object 2 are described by elliptical arc 63, where the first ultrasonic sensor 11, which transmits, is located at one focus, while the third ultrasonic sensor 13, which receives, is located at another focus. This can be similar to the gardener's method, for example. An elliptical arc 63 is constructed, wherein the sum of the distances between points on the elliptical arc 63 is given by the distance determined from the signal propagation time of the cross echo 43.

[0065] against Figure 1 The situation shown, Figure 2 A raster map 200 is shown with a thin vertical column as the object type. In the example shown, the raster map 200 is a two-dimensional raster map with 10×10 elements or cells 210, which represent the parameter space of the X and Y values ​​of the coordinates of the thin vertical column.

[0066] In order to determine the target Figure 1 The conditions shown represent candidate objects 2 in the surrounding environment, and each of the units 210 is associated with a counter. If the unit 210 shares a common hyperplane with the rotating ellipsoids associated with the ultrasonic echoes 31 and 43, the counter is incremented by a value of 1. In other words, in Figure 2 In the example, if the arc 51 associated with the direct echo 31 is tangent to the corresponding cell 210, then the counter of cell 210 is incremented. Figure 2 In the illustration, the first unit 211 is marked with a first shaded line. Similarly, this is applied to the unit associated with the cross echo 43 (see...). Figure 1 The counters of those units 210 cut by the elliptical arc 63 increment by a value of 1. Figure 2 In the illustration, the second unit 213 is marked with a second shading line.

[0067] exist Figure 2 In the example shown, there are two overlapping units 215, which are marked not only by a first shading line but also by a second shading line. The counter of the overlapping unit 215 is incremented twice and therefore has a value of 2. If a boundary value of 1 is given in advance, only the counter of the overlapping unit 215 exceeds this boundary value, and the parameter object represented by the unit 210 is considered as a candidate object.

[0068] Figure 3 A grid map 200 is shown, where vertical walls are treated as object types. Therefore, contrary to treating pillars as object 2, according to... Figure 3 For example, the vertical wall is located 1.6m in front of vehicle 1 at an angle of α = 0° with the X-axis. See [reference needed]. Figure 1The raster map 200 has an angle α as the first dimension and a distance D as the second dimension. The cell 210 of the raster map 200 has a side length of 4 cm in the distance dimension and a side length of 2° in the angle dimension.

[0069] Three measurements were entered into the raster map 200. The three overlapping cells 215 are approximately located at... Figure 3 At the center of the diagram, the counters of these overlapping units increment three times. These overlapping units 215 represent candidate objects.

[0070] Figure 4 For a vehicle 1 equipped with a driver assistance system 300, the fields of view 81, 83 of the first ultrasonic sensor 11 and the third ultrasonic sensor 13 are schematically shown. The second ultrasonic sensor 12 and the fourth ultrasonic sensor 14 also obviously have corresponding fields of view; for clarity, these fields of view are not shown in the diagram. Figure 4 As shown in the diagram, fields of view 81 and 83 represent the following areas of the surrounding environment of vehicle 1: Corresponding ultrasonic sensors 11, 12, 13, and 14 are capable of detecting object 2 within these areas by receiving ultrasonic echoes 31 and 43. This can be used, for example, when creating a grid map 200 (see...). Figure 2 and 3 The counter is incremented only for the following units 210: the units are located within the fields of view 81, 83 of the ultrasonic sensors 11, 12, 13, 14 that receive the corresponding ultrasonic echoes 31, 43.

[0071] This invention is not limited to the embodiments described herein and the aspects emphasized therein. Rather, various modifications that are within the scope of the claims and are of skill to those skilled in the art can be implemented.

Claims

1. A method for identifying and determining the geometric properties of an object (2) using ultrasound, wherein, The ultrasonic pulse (20) is emitted, and the ultrasonic echo reflected at the object (2) is received by at least two ultrasonic sensors (11, 12, 13, 14) arranged spaced apart from each other, characterized in that, The object type is defined by a set of separately defined geometric parameters, wherein the geometric parameters include the position of the object (2). An n-dimensional raster map (200) is created for each object type, where n is the number of parameters associated with the object type, and each cell (210) of the raster map (200) represents a parameter object with geometric parameters associated with the corresponding cell (210), and each cell (210) of the raster map (200) has a counter. In the case of receiving a direct echo (31), in each grid map (200), the counter is incremented for those cells (210) representing the following parameter objects: the surface of the parameter object is tangent to an arc (51) or a sphere surface, the arc or sphere surface being defined by the distance associated with the direct echo (31); and in the case of receiving a cross echo (43), the counter is incremented for those cells (210) representing the following parameter objects: the surface of the parameter object is tangent to an elliptical arc (63) or an ellipsoidal surface, the elliptical arc or ellipsoidal surface being defined by the distance associated with the cross echo (43) and the relative position of the participating ultrasonic sensors (11, 12, 13, 14), and Candidate objects are determined for each object type. For the candidate objects, the cells (210) of the corresponding raster map (200) have counters higher than a pre-given boundary value. When determining whether tangency exists, tolerances are considered for measurement error, scattering, and / or parameter intervals belonging to the cells. The tolerances are pre-given in a differentiated manner for different object types.

2. The method according to claim 1, characterized in that, When the counter is incremented, the unit (210) is assigned a corresponding received ultrasonic echo.

3. The method according to claim 1 or 2, characterized in that, The order of the candidate objects is determined based on classification parameters, wherein the sorting parameters are selected from the following parameters: object type, distance of the candidate object to the ultrasonic sensor (11, 12, 13, 14), number of ultrasonic echoes associated with the candidate object, quality of the associated ultrasonic echoes, and a combination of at least two of the sorting parameters.

4. The method according to claim 3, characterized in that, According to the determined order, the candidate objects are examined in the following ways: whether the amount of ultrasound echoes assigned to the candidate objects is significant, wherein candidate objects with a significance value higher than a pre-given threshold are identified as target objects, and wherein the number of target objects to which ultrasound echoes have been assigned is checked as a parameter for determining significance.

5. The method according to claim 4, characterized in that, Reduce the step size of the grid map (200) within the target object area to improve the accuracy of determining the geometric characteristics of the target object.

6. The method according to any one of claims 1 to 5, characterized in that, The defined object type can be selected from plane, cylinder, sphere, and line.

7. The method according to any one of claims 1 to 6, characterized in that, After a predetermined minimum time and / or after traveling a minimum distance, the ultrasonic echo is removed from the grid map (200), wherein the counter of the corresponding unit (210) is decremented.

8. The method according to any one of claims 1 to 7, characterized in that, Before incrementing the counter of unit (210), it is checked whether the object (2) represented by unit (210) is located within the area defined by the field of view (81, 83) of the ultrasonic sensors (11, 12, 13, 14) that have received ultrasonic echoes.

9. A computer program product comprising instructions that, when executed by a computer, cause the computer to perform the steps of the method according to any one of claims 1 to 8.

10. A driver assistance system (300) for identifying and determining the geometric characteristics of an object (2) in the surrounding environment of a vehicle (1), wherein, The vehicle (1) includes a plurality of ultrasonic sensors (11, 12, 13, 14), characterized in that the driver assistance system (300) is configured to implement the method according to any one of claims 1 to 8.

Images