Method for determining the passability of a driving path, driver assistance system, and vehicle

Abstract

The invention relates to a method for determining the passability of a driving path (6), wherein the method comprises emitting ultrasonic pulses using a plurality of ultrasonic sensors (2), receiving echoes from reflecting objects in the surroundings of a vehicle (1), and determining a passability probability P of a stenosis. It is provided that the passability probability P of the stenosis E is determined by taking into account at least two criteria tp, which are selected from the group consisting of an amplitude gradient, a gradient in the number of received echoes, a position of a focal point, a variation curve of a determined distance, and a detected characteristic object. Another aspect of the invention relates to a driver assistance system (100) which is configured to implement the method, and a vehicle (1) which comprises such a driver assistance system (100).

Description

Technical Field

[0001] The present invention relates to a method for determining the passability of a driving path, wherein the method comprises emitting ultrasonic pulses using a plurality of ultrasonic sensors, receiving echoes from reflecting objects in the vehicle's surroundings, and determining a passability probability P of a constriction E, wherein the constriction E is bounded on both sides by objects. Further aspects relate to a driver assistance system comprising a plurality of ultrasonic sensors and a control unit, wherein the driver assistance system is configured to carry out the method, and to a vehicle comprising such a driver assistance system. Background Art

[0002] Ultrasonic-based environmental sensing systems are used in modern vehicles. Their applications range from simple assistance systems, which warn the driver of obstacles in the surroundings, to systems that perform fully autonomous maneuvers, such as parking assistance. These systems create a map of the surroundings from the echoes of ultrasonic pulses reflected by objects in the vehicle's surroundings. Reliably identifying all objects that could cause a collision is crucial. Preventing false warnings or even unnecessary interventions, such as automatically initiated braking, is also crucial in situations that do not present a safety risk. Estimating the height of the objects is crucial, ensuring that low objects, such as door sills or falling curb edges, are not warned but rather appear passable, while reliably warning against tall objects, such as pillars or walls.

[0003] However, ultrasound-based systems that achieve a high degree of classification of detected objects often reach their limits when numerous objects are located in the vehicle's surroundings and their echoes partially overlap. This situation arises particularly when driving into a garage or parking structure, in tight parking spaces, or on narrow roads with oncoming traffic. In particular, when low curbs or thresholds must be driven over in such situations, these low objects can be misclassified, leading to false warnings or unnecessary brake interventions.

[0004] DE 10 2016 103 251 A1 discloses a method for operating a sensor for sensing a vehicle's surroundings. Here, a scenario describing the vehicle's surroundings is identified, such as a vehicle in a highway tunnel. Object detection algorithms are then controlled so that, for example, traffic signs that cannot be present in the identified scenario are not searched for. Furthermore, pedestrian detection can be adapted, for example, depending on the identified scenario, as overlooked pedestrians in urban areas are extremely important.

[0005] Document DE 10 2017 127 972 A1 describes a multi-sensor probabilistic object detection system. In this system, data about the surrounding environment is sensed and evaluated using multiple sensors, which may also include different sensor types. For example, if an object is being searched for behind a vehicle, the probability of the object being behind the vehicle is updated when the sensor detects the object. This probability can be calculated using a Bayesian statistical model, and the probability of the object being sensed can be updated using a Bayesian probability updating technique.

[0006] A method for classifying objects using distance data is known from DE 10 2007 061 235 A1. Here, for high-level classification, it is utilized that for ultrasonic sensor measurement signals, large objects have a greater dispersion in the distance data between two measurements than smaller objects. This is caused by multiple reflections, which can occur with large objects. With smaller objects, multiple reflections occur only with a lower probability.

[0007] A disadvantage of known methods is that they cannot reliably determine whether a driving path is passable or not in situations with partially overlapping multiple ultrasonic echoes. However, such situations often occur during daylight hours, for example when driving through a garage entrance, into a parking structure, or on narrow roads. Therefore, there is a need for a method that can also determine the passability of a driving path in such situations. Summary of the Invention

[0008] A method for determining the passability of a driving path is proposed. The method comprises emitting ultrasonic pulses using a plurality of ultrasonic sensors, receiving echoes from reflecting objects in the vehicle's surroundings, and determining a probability P of the passability of the constriction. The constriction E is bounded on both sides by objects. The method further provides for determining the probability P of the passability of the constriction E by considering at least two criteria tp, selected from the group consisting of:

[0009] the amplitude gradient of the received echo starting from the center of the driving corridor to the edge of the driving corridor,

[0010] the number gradient of the received echoes starting from the center of the driving corridor to the edge of the driving corridor,

[0011] - Intersection cluster points assigned to the spacing of the received echoes Relative to the driving channel,

[0012] - the course of the distance ascertained by the echo received by the ultrasonic sensor when the vehicle approaches a reflecting object, and

[0013] - Identify objects that represent stenosis.

[0014] In this setting, for each of the considered criteria tp, the probability of stenosis permeability P(E|tp) is determined by the following relation,

[0015] P(E|tp)= P(tp|E)·P(E) / P(tp) (1)

[0016] Where P(tp|E) represents the probability that the standard tp appears in relation to the narrow part that can be passed through, P(E) represents the probability that the narrow part can be passed through by the vehicle, and

[0017] P(tp) represents the probability that the standard tp is noticed.

[0018] Preferably, during vehicle operation, a continuous check is performed to determine whether the activation condition for the method, "the presence of a constriction," is met. To this end, the vehicle environment is monitored using ultrasonic sensors and checked to determine whether a constriction exists along the vehicle's travel path, i.e., an area bounded on both sides of the travel path by at least one object. For this purpose, sensors positioned on the vehicle's sides can be used, or ultrasonic echoes received by the side sensors can be given a higher weighting. Furthermore, the detection of an object approaching the constricted travel path can be used as an additional condition. Approach is detected by a decrease in the distance determined by the ultrasonic sensors during vehicle movement.

[0019] Here, a driving corridor is defined as the area of the vehicle's surroundings that is currently occupied by the vehicle and will be driven into in the future. The area that will be driven into in the future can be estimated, for example, by extrapolating the vehicle's motion. This can be done, for example, using the current steering angle and speed or based on a planned vehicle trajectory. The vehicle trajectory is the curve of the track along which the vehicle moves. The driving corridor is delimited laterally by a first boundary and a second boundary. When driving in a straight line, the distance between the first and second boundaries is essentially predetermined by the vehicle width.

[0020] If there are no objects in the driving path that cannot be driven over, then the driving path is considered to be passable. However, low objects, such as door sills or lowered curbs, are objects that can be driven over and can also be located in the driving path without impairing the passability of the driving path.

[0021] The narrow area through which the travel path passes has two objects, each located on opposite sides of the travel path and thus delimiting the maximum width of the travel path. Such delimiting objects can be, for example, pillars, walls, columns, or walls. Precise position determination is particularly difficult with conventional methods for extended objects such as walls, as the reflection point from which the ultrasonic pulse is reflected is not clearly defined. In situations where multiple objects are located in the vehicle's surroundings, height classification—that is, distinguishing between low objects that can be driven over and tall objects that cannot be driven over—is also difficult with conventional methods.

[0022] Therefore, the present invention provides for using a statistical method to determine the probability of a driving path being passable or impassable in the area of a narrow section. Two or more criteria are combined, each weighted according to its quality. In some cases, a single criterion is insufficient for a clear classification. For each of these criteria, a certain probability P(tp|E) of the criterion occurring within the narrow section is given, as well as a probability P(tp) that the criterion is fully present, i.e., also present in aspects other than the narrow section. According to Equation 1, the probability of a narrow section being passable is determined together with the probability P(E) that the vehicle can pass through the narrow section.

[0023] Preferably, a threshold value is predefined for the probability. If the probability determined using at least two criteria tp is above this limit value, the stenosis is considered to be passable.

[0024] Furthermore, if no other impassable objects are detected in the driving path, the driving path is also considered passable as a whole. Furthermore, if neither constrictions nor impassable objects are detected in or around the driving path, the driving path is of course considered passable.

[0025] Preferably, in this method, all of the aforementioned criteria are taken into account, namely the amplitude gradient, the gradient of the number of received echoes, the position of the focal point, the ascertained change in the spacing and the identified characteristic objects, in order to determine the probability P of the stenosis being permeable.

[0026] For example, the probabilities P(E|tp) determined for the considered criteria tp can be summarized by weighted summation to give the probability P of the permeability of the stenosis.

[0027] Preferably, for the standard tp, the probability P(tp|E) of the standard tp occurring in relation to the stenosis and the attention probability P(tp) are determined empirically and provided via a database.

[0028] Furthermore, it can be provided that weighting parameters for the weighted summation and / or the probability P(tp|E) and the attention probability P(tp) are predefined.

[0029] In one embodiment of the method, a threshold value and quality are predefined for each standard TP. The quality of a standard TP that exceeds the assigned threshold value is added to the overall quality, wherein the overall quality must exceed a predefined quality threshold in order for the stenosis to be considered passable. Depending on the quality or value of the standard TP, typically two or three criteria are required for classification.

[0030] In a further embodiment of the method, a decision tree is predefined in order to enable a classification of the stenosis as passable or non-passable, taking into account the criterion tp.

[0031] The thresholds, qualities, weighting factors, and / or decision trees used for classification can be predefined, for example, using a database of multiple known scenarios with stenosis. To this end, the database preferably includes received echoes for multiple different scenarios, each with information about the stenosis and information about the permeability of the corresponding stenosis. These information can be used, for example, using statistical methods to analyze how often a specific criterion tp associated with a permeable stenosis is identified. Accordingly, when the criterion tp exists, P(E|tp) represents the probability of a permeable stenosis. Given this correlation, P(E) represents the probability of the permeability of the stenosis for all observed scenarios with a stenosis.

[0032] For example, a decision tree for classifying stenosis using standard TP can be generated using an offline training method.

[0033] Alternatively, the parameters used, such as threshold values, quality, weighting factors and / or decision trees, can be determined empirically. Furthermore, it is conceivable to optimize these parameters using machine learning methods. Such predefined values can be stored in a database and can be saved in the controller.

[0034] A criterion tp can be considered as an amplitude gradient. This utilizes the fact that objects that delimit a narrowing and cannot be driven over typically have high reflectivity and therefore generally provide the highest echo amplitudes. Furthermore, these objects are generally visible to all of the multiple ultrasonic sensors, meaning that each ultrasonic sensor receives at least one ultrasonic echo for each of these objects. The amplitude of the received ultrasonic echoes is also dependent on the field of view of the respective ultrasonic sensor. Ultrasonic sensors positioned to the sides of the vehicle have a more favorable field of view relative to objects that laterally delimit the roadway and receive corresponding echoes with higher amplitudes. Ultrasonic sensors positioned closer to the center of the vehicle, and therefore closer to the center of the roadway, receive corresponding ultrasonic echoes with smaller amplitudes for these laterally delimiting objects. Therefore, an amplitude gradient that increases from the center of the roadway toward the roadway boundary indicates the presence of a narrowing, which is laterally delimited by the object. However, if the amplitude remains largely unchanged or even increases from the outside toward the center of the roadway, this indicates an object located within the roadway.

[0035] Another criterion used is the gradient in the number of received echoes. This utilizes the fact that the corners of extended objects (such as walls or vehicles) often have many small edges that reflect ultrasound waves. The better the sensor's field of view of such edges, the more of these secondary reflections a corresponding ultrasonic sensor can receive. Conversely, low, traversable objects, such as curbs or door sills, generally provide very few reflections or echoes. Therefore, in the presence of a narrowing, it can be determined that the number of received echoes increases from the center of the roadway outward toward the roadway boundary. This characteristic is particularly indicative of the passability of such a narrowing.

[0036] Another criterion tp is the position of the intersection point associated with the distances of the received echoes relative to the roadway. To this end, using propagation time and the speed of sound, distance values are determined for all received echoes, and intersection points are calculated for all distance values and entered into the surroundings map. Strongly reflecting objects that are usually ineligible to drive over are detected by several or even all of the vehicle's ultrasonic sensors, while low objects that can be driven over are usually detected by only a few ultrasonic sensors. This results in a sharp increase in the number of ultrasonic sensors in the area of navigable narrow areas. If the majority of all the intersection points found are outside the driving path, this is a sign of the passability of the narrow area.

[0037] Another possible metric, TP, is the curve of the distance determined for the echoes received by the ultrasonic sensor as the vehicle approaches a reflecting object. In the case of a narrow section with a curb or door sill, the ultrasonic sensor, located more centrally in the front area of the vehicle, receives echoes from low objects that can be driven over, as well as echoes from objects that laterally bound the narrow section. Depending on the distance of the vehicle from the objects in question, the received echoes may overlap, making them difficult to separate. As the vehicle approaches a narrow section, and therefore, particularly when approaching low objects such as door sills or curbs, the distance between the vehicle, or the ultrasonic sensor, and the low object changes more significantly than the distance between the ultrasonic sensor and the laterally bounding objects. Therefore, if the number of echoes determined from the ultrasonic echoes is plotted against the vehicle's travel distance, as the vehicle approaches the low object, the previously overlapping echo separates into two distinct echoes. The distance to the echo that can be assigned to the laterally bounding object does not change or changes only slightly, while the distance to the low object decreases continuously. This dispersion of the distances determined for the ultrasonic echoes is a sign that there are no large, unpassable objects in the center in front of the vehicle, but only small objects directly in front of the vehicle and objects delimiting the vehicle to the sides. This indicates a passable constriction.

[0038] Another possible criterion for identifying a traversable constriction is the detection of an object that represents the constriction. In the case of a garage, the upper edge of the garage can usually be detected by an ultrasonic sensor when entering. In a standard garage, this upper edge is located at a specific height and can be detected as the vehicle approaches the garage, for example, by the curve of the distance determined from the echoes received by the ultrasonic sensor.

[0039] Preferably, the upper edge of the garage and the threshold of the drive-in door are used as characterizing objects.

[0040] If, within the scope of the proposed method, it is known that the driving path is bounded by a passable constriction, for example, an ultrasonic distance warning system can be reconfigured. This can include, for example, shortening the warning distance and adjusting the parameters used for altitude classification. This ensures that the driver of the vehicle does not receive unnecessary warnings or even initiate unnecessary braking interventions when passing through a constriction that has been identified as passable.

[0041] Furthermore, it is possible to reduce the adaptation window within which the ultrasound echoes are processed as echoes of the same object when a passable stenosis is detected.

[0042] In addition, within the scope of the present method, the ultrasound-based system can be reconfigured when the activation condition, ie, the presence of a stenosis, is identified. This reconfiguration can be performed independently of whether the stenosis is classified as passable or impassable.

[0043] For example, when locating an object, during normal operation, a single intersection point is sufficient for determining the object position. In the case of a stenosis, this often results in echoes from different sources intersecting and this leads to an erroneously determined position.

[0044] If a stenosis is detected, the system for determining the position of the object is preferably reconfigured so that each position must be confirmed by at least one multivalued echo in order to be used. A multivalued echo is sensed by two or more sensors. The multivalued echoes are in particular cross-echoes.

[0045] In order to improve object matching in the region of the stenosis and to avoid the calculation of many erroneous objects or incorrectly positioned objects, it can be provided that the matching window is already reduced when the stenosis is detected.

[0046] Preferably, the classification method for the height classification of the object is reconfigured when a stenosis is detected. Utilization is made of the fact that a plurality of different attributes are used in the height classification method, wherein the weighting of these attributes is changed when a stenosis is detected.

[0047] Furthermore, it can be provided that the processing of the ground clutter, ie the processing of the ultrasonic echoes caused by the ground, is adapted when a constriction is detected.

[0048] The detection of narrow spots, and in particular the detection of traversable narrow spots, is preferably used to suggest to the vehicle driver useful assistance functions related to the narrow spot. Such assistance functions are, for example, remote-controlled parking in a garage or a simple guided parking maneuver. This guided parking maneuver can advantageously be started before the parking space is driven through.

[0049] Another object of the present invention is to provide a driver assistance system comprising a plurality of ultrasonic sensors and a control unit. The driver assistance system is designed and / or configured to carry out one of the methods described herein. Accordingly, features described within the scope of one of the methods apply to the driver assistance system, and conversely, features described within the scope of the driver assistance system also apply to the method.

[0050] In the driver assistance system, it is preferably provided that a plurality of ultrasonic sensors are oriented toward the front of the vehicle, ie, toward the normal direction of travel. For example, 3 to 8 ultrasonic sensors are oriented toward the front.

[0051] Another aspect of the present invention is to provide a vehicle comprising such a driver assistance system. In this case, the ultrasonic sensor is preferably arranged in the front part of the vehicle, for example in the bumper of the vehicle.

[0052] The method for determining the passability of a driving lane according to the present invention advantageously allows the echoes of reflecting objects in the vehicle's surroundings to be comprehensively correlated and not evaluated individually. This allows for reliable identification of narrowings and the reliable determination of their passability. Multiple reflections of objects, which cause problems in methods known from the prior art, are not discarded in the proposed method but are advantageously used as a criterion for the presence of a passable narrowing. This advantageously increases safety during classification, preventing false warnings or even unwarranted braking interventions. Furthermore, after a narrowing is detected, the vehicle's corresponding ultrasonic assistance systems can be reconfigured to optimally respond to the identified narrowing. BRIEF DESCRIPTION OF THE DRAWINGS

[0053] Exemplary embodiments of the present invention are explained in more detail with reference to the drawings and the following description.

[0054] The accompanying drawings show:

[0055] Figure 1 Before driving into the garage, a vehicle equipped with the driver assistance system of the present invention,

[0056] Figure 2 For Figure 1 A graphical representation of the echo amplitude for the conditions shown in,

[0057] Figure 3 For Figure 1 A graphic representation of the number of echoes for the conditions shown in FIG.

[0058] Figure 4 For Figure 1 The curve of the distance determined by the middle ultrasonic sensor for the situation shown in FIG. DETAILED DESCRIPTION

[0059] In the following description of the embodiments of the present invention, identical or similar elements are denoted by identical reference numerals, wherein a repeated description of these elements is omitted in individual cases. The drawings merely schematically illustrate the content of the present invention.

[0060] Figure 1 Vehicle 1 is shown when driving into garage 20. Vehicle 1 has a driver assistance system 100 according to the present invention.

[0061] Vehicle 1 or driver assistance system 100 has a plurality of ultrasonic sensors 2 that are connected to control unit 4 . To sense data about the surroundings of vehicle 1 , ultrasonic sensors 2 emit ultrasonic pulses and receive echoes 10 , 12 from objects in the surroundings of vehicle 1 .

[0062] exist Figure 1 In the situation shown in FIG, vehicle 1 is moving forward, that is, in its normal direction of travel toward garage 20. Garage 20 is open and a driveway 6 opens into the interior of garage 20, defining the space in which vehicle 1 will later travel. Driveway 6 is bounded on both sides by boundaries 61 and 62, with driveway 6 being bounded on the left by a first boundary 61 and on the right by a second boundary 62.

[0063] Garage 20 in Figure 1 The visible detail comprises a portion of a wall 32 of the garage 20, wherein the driveway 6 opens into the interior of the garage 20 through the drive-in opening 24 of the garage 20. In this case, the driveway 6 crosses a threshold 30 of the garage 20, which is located in the region of the drive-in opening 24. This threshold 30 is a very low obstacle that can therefore be driven over, but which is nevertheless detected by the ultrasonic sensor 2 of the vehicle 1.

[0064] On either side of the driving path 6, the front corner 321 and the rear corner 322 are located on the wall 32 of the garage 20. These corners 321 and 322 of the wall 32 are high obstacles that cannot be driven over. These high obstacles that cannot be driven over are also detected by the ultrasonic sensor 2 of the vehicle 1.

[0065] With the proposed method, ultrasonic pulses are emitted by the ultrasonic sensor 2 of the vehicle 1 , the reflected echoes 10 , 12 are received again by the ultrasonic sensor 2 , and the received echoes 10 , 12 are evaluated by the control unit 4 . Figure 1 , only the echoes 10 , 12 are marked, but the ultrasonic pulses emitted by the individual ultrasonic sensors 2 are not shown.

[0066] As Figure 1 As can be seen from the diagram, when the threshold 30 is a low obstacle that can be driven over, only the direct echo 10 appears, which is received by the intermediate ultrasonic sensor 22. The direct echo 10 should be understood as the ultrasonic echo received by the ultrasonic sensor 2 that emitted the original ultrasonic pulse. The cross echo 12 should be understood as the echo received by an ultrasonic sensor 2 different from the ultrasonic sensor 2 that emitted the original ultrasonic pulse. Figure 1As can be seen from the diagram, the front corner 321 and the rear corner 322 of the wall 32 reflect ultrasonic waves well, so that both the direct echo 10 and the cross echo 12 are received from the corners 321 and 322. Here, the direct echo 10 and the cross echo 12 from the corners 321 and 322 are received not only by the central ultrasonic sensor 22 but also by the ultrasonic sensor 21 arranged on the side of the vehicle 1.

[0067] Using a known speed of sound, the distance between the corresponding ultrasonic sensor 2 and the object that reflected the ultrasonic pulse can be determined by measuring the signal propagation times of the direct echo 10 and the cross echo 12, respectively. However, a single received echo 10, 12 is not sufficient to determine the relative position of the object that reflected the ultrasonic wave, or the point on the object that reflected the ultrasonic wave. Using the least squares method, the relative position relative to vehicle 1 or relative to the corresponding ultrasonic sensor 2 can be calculated using two or more received echoes 10, 12 that can be assigned to the same reflecting object. The problem here is that, in order to implement a correct least squares method, each received echo 10, 12 must be assigned to the correct reflecting object.

[0068] In the method according to the present invention, however, an intersection point is formed, similar to the least squares method, between two received echoes 10, 12 or the distances assigned to these echoes 10, 12. This results in intersection points that may correspond to real objects in the surroundings of vehicle 1. However, other intersection points may also occur that do not correspond to real objects in the surroundings of vehicle 1.

[0069] exist Figure 1 In the example shown in FIG, the corners 321 and 322 are respectively seen by multiple ultrasonic sensors in the ultrasonic sensor 2, so that a clear intersection point is obtained outside the driving channel 6. The door sill 30 is only visible to the middle ultrasonic sensor 22 and is also recorded there only by the direct echo 10, so that significantly fewer intersection points occur in the area within the driving path 6. Figure 1 In the case schematically shown in FIG, two intersection points appear, which are located on both sides of the driving path 6 and indicate a navigable constriction.

[0070] Figure 2 A diagram is shown in which the number of echoes # is plotted against the distance S from the center of the driving path 6 (see Figure 1 ) relationship. Figure 1 In the situation shown in FIG, a small number of ultrasonic echoes are obtained in the central area of the travel channel 6 and a large number of echoes are obtained outside or at the edge of the travel channel 6. This gradient increase in the number of echoes from the center of the travel channel 6 to the edge of the travel channel 6 indicates a traversable narrow area.

[0071] Figure 3 Shows the Figure 1 The situation shown is that the echoes 10, 12 detected by the ultrasonic sensor 2 (see Figure 1 ) changes in amplitude A relative to the distance S from the center of the travel channel 6. Figure 3 The diagram shows that the echo detected by center ultrasonic sensor 22 in the center region of driving path 6 has a smaller amplitude A than the echo detected by side ultrasonic sensors 21 at the edges of driving path 6 . This indicates that only objects with low reflectivity are present within driving path 6 , which generally represent low objects that can be driven over, while objects with high reflectivity are located at the edges of driving path 6 . This indicates a navigable constriction.

[0072] Figure 4 Shows the Figure 1 The situation shown in FIG. 1 shows the distance D determined by one of the central ultrasonic sensors 22 when the vehicle 1 approaches the garage 20. As time T passes, the vehicle 1 approaches the garage 20, wherein the measured distance D generally decreases due to the decreasing distance. Figure 4 The diagram shows that the received echoes separate over time T, resulting in a first distance curve 50 for obstacles immediately adjacent to vehicle 1 and a second distance curve 52 for objects further away from vehicle 1. First distance curve 50 can be assigned to direct echo 10 from door sill 30, since the distance D between door sill 30 and vehicle 1 decreases continuously as the vehicle approaches. In contrast, the values of second distance curve 52 can be assigned to corners 321, 322 of wall 32 of garage 20, since the distance D between wall 32 and vehicle 1 decreases initially and then remains essentially constant once vehicle 1 has reached garage 20. This separation of distance values as vehicle 1 approaches indicates a narrow, navigable area.

[0073] The present invention is not limited to the embodiments described herein and the aspects emphasized therein, but rather a number of modifications are possible within the scope of the claims and within the scope of the usual technical means of a person skilled in the art.

Claims

1. A method for determining the passability of a driving passage (6), wherein: The method comprises emitting ultrasonic pulses using a plurality of ultrasonic sensors (2) and receiving echoes from reflecting objects in the surroundings of a vehicle (1), and determining a probability P of permeability of a stenosis E, wherein the stenosis E is bounded on both sides by objects, characterized in that the probability P of permeability of the stenosis E is determined by taking into account at least two criteria tp, which are selected from: - an amplitude gradient of the received echo starting from the center of the driving channel (6) to the boundaries (61, 62) of the driving channel (6), a quantitative gradient of the received echoes starting from the center of the driving path (6) to the boundaries (61, 62) of the driving path (6), - the position of the intersection point associated with the spacing of the received echoes relative to the driving path (6), - a curve of the distance determined for the echo received by the ultrasonic sensor (2) when the vehicle (1) approaches a reflecting object, and - an object representing the stenosis E is detected, For each of the considered criteria tp, the probability P(E|tp) of the permeability of the stenosis E is determined by the following relation: P(E|tp)=P(tp|E)·P(E) / P(tp) Wherein, P(tp|E) represents the probability of the standard tp appearing in relation to the stenosis site E. P(E) represents the probability that the narrow part E can be passed by the vehicle (1), and P(tp) represents the attention probability for the standard tp.

2. The method according to claim 1, characterized in that When determining the probability P of the permeability of the stenosis E, the amplitude gradient, the number gradient, the position of the focal point, the ascertained course of the distances and the identified characteristic objects of the received echoes are taken into account.

3. The method according to claim 1 or 2, characterized in that The probabilities P(E|tp) determined for the considered criteria tp are combined by weighted summation to give the probability P of the permeability of the stenosis E.

4. The method according to claim 1 or 2, characterized in that The probability P(tp|E) of the occurrence of the standard tp in relation to the stenosis E and the probability of attention P(tp) of the standard tp are determined empirically and provided via a database.

5. The method according to claim 3, characterized in that The weighting parameters for the weighted summation and / or the probability P(tp|E) and the attention probability P(tp) are determined by a machine learning method, wherein a database is used as training data, which has received echoes for a plurality of different scenarios with a stenosis E together with information on the permeability of the corresponding stenosis E.

6. The method according to claim 1 or 2, characterized in that The probability P(E) of the permeability of the stenosis E is determined empirically and predetermined by a database, or is estimated by predetermined relationships.

7. The method according to claim 1 or 2, characterized in that The recognition of the object representing the narrow area E includes the recognition of the upper edge of the garage (20) and the threshold (30) of the driveway (24).

8. The method according to claim 1 or 2, characterized in that When a passable stenoses E are detected, the ultrasound-based distance warning system is reconfigured.

9. A driver assistance system (100), comprising a plurality of ultrasonic sensors (2) and a controller (4), characterized in that: The driver assistance system (100) is configured to carry out a method according to any one of claims 1 to 8.

10. A vehicle (1) comprising a driver assistance system (100) according to claim 9.

Images