DE102020213831B4 - Method for determining an existence probability of a possible element in an environment of a motor vehicle, driver assistance system and motor vehicle - Google Patents

Abstract

Method for determining an existence probability (Vη of at least one possible element (14) in an environment (16) of a motor vehicle (10), having the steps:- providing map data (18a) of a map (18) of the environment (16) of the motor vehicle ( 10), which relate to a first surrounding area, which comprises the at least one possible element (14);- providing sensor data (D) by means of at least one surrounding area detection device (26) of the motor vehicle (10), which relate to a second surrounding area, which comprises the at least one possible element (14); and- comparing the map data (18a) and the sensor data (D) with regard to a presence of the at least one possible element (14);- determining the probability of existence (W) depending on a result of the comparison; and- Outputting the probability of existence (W);- the at least one possible element (14) being based at least on the map data (18a) and/or the sensor data (D ) is classified as present; characterized in that a rule database (36) is provided, in which information relating to a traffic area is stored, at least in the event that when comparing the map data (18a) and the sensor data (D) there is a discrepancy with regard to the presence of the at least one possible element (14) is determined, the existence probability (W) is determined as a function of at least one piece of first information provided by the rule database (36), wherein when comparing the map data (18a) and sensor data (D), a comparison is made in an overlapping area of the first and second surrounding area relating to all elements (14) present in the overlapping area based on the map data (18a) and/or sensor data (D), each element (14) of all elements in the overlapping area based on the map data (18a) and/or sensor data (D ) existing elements (14) depending on a result of the comparison by assignment is classified into one of at least the following groups:- a first group of the matching elements, if the element (14) is detected both on the basis of the sensor data (D) and is present in the map data (18a);- a second group of the deviant ones elements, if the element (14) was detected on the basis of the sensor data (D) but is not present in the map data (18a); - a third group of the unevaluated elements, if the element (14) is in the map data (18a) is present and was not detected as being present based on the sensor data (D); wherein the at least one possible element (14) is assigned to the second group or third group.

Description

The invention relates to a method for determining the probability of the existence of at least one possible element in the surroundings of a motor vehicle, wherein map data of a map of the surroundings of the motor vehicle are provided, which relate to a first surrounding area that includes the at least one possible element, sensor data by means of at least one surroundings detection device Motor vehicle are provided that relate to a second surrounding area, which includes the at least one possible element, and wherein the map data and the sensor data are compared with regard to the presence of at least one possible element. Furthermore, depending on a result of the comparison, the probability of existence is determined and output, with the at least one possible element being classified as present at least on the basis of the map data and/or the sensor data. Furthermore, the invention also relates to a driver assistance system and a motor vehicle.

Automatic driving functions in motor vehicles currently rely on driving-relevant information, such as lane topologies or priority rules, which are stored in an HD map, mostly without checking them. However, it can happen that some of the information stored in such an HD map is incorrect, for example if there are errors in map generation, or is no longer up-to-date, for example if the environment changes, as is the case in the case, for example from construction sites may be the case. Vehicle sensors can be used to detect possible errors in such a map. Accordingly, discrepancies between the sensor detection and the map data indicate possible errors, either in the detection of the surroundings or in the map data. However, it is often not possible, or not particularly reliable, to assess whether the error is in the map data or in the detection of the surroundings. As a precaution, however, the relevant elements are then assessed as questionable on a map and, to avoid dangerous situations resulting therefrom, the motor vehicle which is currently performing an automatic driving function is transferred to a safe state, which involves stopping the vehicle. But even stopping the vehicle can be a safety hazard, as it provokes rear-end collisions, for example. In addition, it is very comfort-restricting, since it does not go hand in hand with reaching a desired destination. Accordingly, it would be desirable to be able to evaluate elements in such a map more reliably with regard to their probability of existence.

The WO 2018/127461 A1 describes a method for providing an HD map, in particular a first HD map, which results from the data fusion of sensor data and map data of a second HD map. This allows existing data to be supplemented with additional information provided by the sensors. Map data can also be replaced by sensor data. This assumes that the sensor data is more accurate and reliable. The reliability of the data is limited to the reliability of the vehicle sensors. However, the case that sensor data can also be faulty or not sufficiently accurate is not taken into account here.

Furthermore describes the WO 2018/126215 A1 a method for updating an HD map based on the collected sensor data of a plurality of motor vehicles driving through geographic areas. If there are discrepancies in the recorded sensor data, the online system to which this data is transmitted can carry out an evaluation.

In principle, fleet vehicles can be used to detect possible errors in map data. However, there is often the problem here that a map update based on this cannot be carried out in real time but only takes place on a significantly longer time scale, for example weeks or months.

The DE 10 2019 200 347 A1 describes a backend for an object update system for removing objects that are present in a digital road map. The digital road map with existing objects is stored on the backend. The backend is set up to receive a path and environmental data recorded along this path from a vehicle and to recognize objects in this environmental data. The backend is also set up to remove an object present in the digital road map from the latter if this object cannot be identified by the backend in the environmental data recorded by the vehicle.

The DE 10 2016 215 249 A1 describes a method for supporting a driver assistance system in a motor vehicle, a map being provided, with objects classified in the map being stored at associated positions, environmental data being recorded using at least one environmental sensor system of the driver assistance system, and the environmental data being evaluated using an evaluation device of the driver assistance system, wherein the captured environmental data for object recognition tion can be evaluated depending on the classified objects stored in the map.

The DE 10 2017 207 441 A1 describes a method for checking a digital environment map for a driver assistance system of a motor vehicle, in which sensor data which describe at least one object in the environment of the motor vehicle are received by at least one environment sensor and based on the sensor data a map element in the environment map which the at least one object described, is checked, wherein for checking using the map element an object position which describes a position of the at least one object in the environment is determined, and without detecting the at least one object using the sensor data it is checked whether the sensor data the at least one object describe at the determined object position.

The object of the present invention is therefore to provide a method for determining the probability of the existence of at least one possible element in the area surrounding a motor vehicle, a driver assistance system and a motor vehicle which make it possible to identify a possible element in the area surrounding the motor vehicle as quickly and reliably as possible, preferably in Real-time while driving, to assess as existent or non-existent.

This object is achieved by a method, a driver assistance system and a motor vehicle with the features according to the respective independent patent claims. Advantageous configurations of the invention are the subject matter of the dependent patent claims, the description and the figures.

• - einer ersten Gruppe der übereinstimmenden Elemente, wenn das Element sowohl auf Basis der Sensordaten detektiert als auch in den Kartendaten vorhanden ist;
• - einer zweiten Gruppe der abweichenden, insbesondere neu detektierten, Elemente, falls das Element auf Basis der Sensordaten erfasst wurde und aber in den Kartendaten nicht vorhanden ist;
• - einer dritten Gruppe der nicht bewerteten Elemente, falls das Element in den Kartendaten vorhanden ist und auf Basis der Sensordaten nicht als vorhanden detektiert wurde.

In a method according to the invention for determining a probability of existence of at least one possible element in the surroundings of a motor vehicle, map data of a map of the surroundings of the motor vehicle are provided, which relate to a first surrounding area that includes the at least one possible element. Furthermore, sensor data are provided by means of at least one surroundings detection device of the motor vehicle, which relate to a second surrounding area that includes the at least one possible element. In addition, the map data and the sensor data are compared with regard to the presence of the at least one possible element and, depending on a result of the comparison, the probability of existence is determined and output, with the at least one possible element being classified as present at least on the basis of the map data and/or the sensor data . Furthermore, a rule database is provided, in which information relating to a traffic area is stored, wherein at least in the event that a discrepancy regarding the presence of the at least one possible element is determined when comparing the map data and the sensor data, the probability of existence depending on at least one first information provided by the rule database is determined. When comparing the map data and sensor data, a comparison is carried out in an overlapping area of the first and second surrounding area relating to all elements present in the overlapping area based on the map data and/or sensor data, with each element of all elements present in the overlapping area based on the map data and/or sensor data Elements are classified depending on a result of the comparison by being assigned to one of at least the following groups:

• - a first group of the matching items if the item is both detected based on the sensor data and present in the map data;
• - A second group of deviating, in particular newly detected, elements if the element was detected on the basis of the sensor data but is not present in the map data;
• - a third group of the unassessed items if the item is present in the map data and was not detected as present based on the sensor data.

The at least one possible element is assigned to the second group or third group.

The invention is based on the finding that knowledge relating to a traffic area can be used to resolve discrepancies between sensor data and map data, in particular about aspects of the traffic area relevant to driving and safety, which knowledge can be stored in the form of rules in the rule database. This knowledge, which is explicitly formulated in the rule database, can thus advantageously be used as additional information in order to resolve any contradictions between sensor data and map data. Said knowledge can be specified in particular in the form of rules, which can include traffic rules on the one hand, but on the other hand also allow conclusions to be drawn from certain recorded conditions or from conditions indicated by the map. For example, the rule database allows conclusions, such as that an intersection without traffic signs means "right-before-left".

This allows the right of way situation to be modeled for each arm of the intersection. Furthermore, a speed limit of 50 km/h can be deduced from a town limit sign. The rule database can also specify permissible or impermissible links between elements. In this way, it can be determined, for example, which street signs can coexist in the same place and which cannot. For example, an overtaking ban can be combined with a speed limit. However, a stop sign combined with a give way sign represents a contradiction. Such linking rules can be stored in the rule database and, in the event of discrepancies between sensor data and map data relating to a possible element, can be used to more accurately determine the probability of this element's existence, for example below on the basis of the probability of existence, to decide correctly with a higher probability whether the sensor data or map data are correct with regard to the existence hypothesis of this element. Overall, this knowledge of the rule database can be used to ultimately determine the probability of existence of certain elements, here for example street signs. If, for example, an element is specified by the map that is not recorded on the basis of the sensor data, this can also have several reasons. For example, it may be the case that this element is currently being covered by a road user driving ahead. If, instead, at the position at which the element should be detected, such a road user driving ahead is detected using the sensor data, it can be assumed to be more likely that this element is still present and is only covered by this other road user than to assume that this item does not exist. Such conclusions can also be specified by the rule database. In particular, therefore, the existence of the possible element can be checked for a rule conformity of this existence according to at least one of the rules of the rule database. The result of this checking can therefore be included in the determination of the probability of existence. If the existence of the possible element does not conform to the rule, its probability of existence can be downgraded and vice versa, ie if the existence of the possible element conforms to the rule, its probability of existence can be increased. The provision of the rule database means that an additional instrument is provided, on the basis of which the probability of the existence of elements can be determined in addition to the sensor data and map data. By comparing the environmental information provided by the sensor data and the map data with such a set of rules, more reliable statements about the existence of such elements in the environment of a motor vehicle can be provided overall. In other words, the probability of existence of possible elements in the area surrounding the motor vehicle can be determined more precisely by additionally taking into account the information that can be provided by the rule database. Ultimately, this means that the probabilities of existence determined in this way largely enable a clearer statement to be made. Existence hypotheses can thus be checked and, if necessary, corrected, in particular in real time and thus during the journey or normal operation of the vehicle. This allows possible discrepancies between sensor data and map data to be resolved without having to interrupt the autonomous journey for safety reasons, and thus enables the motor vehicle to continue driving safely autonomously much more frequently in such cases, i.e. when discrepancies occur, which in turn increases road safety and the driving comfort increases.

The probability of existence itself can be determined according to a predetermined metric. Such metrics are sufficiently known and are therefore not specified in more detail within the scope of the invention. To give an example, the probability of existence can be provided as a weighted average of various parameters that influence the probability of existence. In addition to the knowledge provided by the rule database as such a parameter, numerous other parameters can also be included in the averaging of the probability of existence, such as the reliability of the respective sensor data and the map data, a degree of correspondence between sensor data and map data, the type of environment detection device, through which the sensor data are provided, the number of environment detection devices that are used for environment detection and, for example, deliver the same results with regard to the presence of the element, or the like. All of these parameters can accordingly enter into a kind of weighted averaging as an example of such a predetermined metric in order to ultimately provide the existence probability of the relevant possible element.

The possible element is an element which is classified as present at least on the basis of the map data and/or the sensor data, as was defined above. This means that either the map data state that this element exists, in which case this element does not necessarily have to be detected by the environment detection device, for example sensors of the motor vehicle, or Conversely, that this element was detected by the at least one environment detection device, regardless of whether the presence of this element is also documented in the map data. In particular, it can be provided that the probability of existence is only ascertained as a function of information provided by the rule database if a discrepancy with regard to the presence of the at least one possible element is ascertained. This means that the rule database does not have to be used, for example, if an element is classified as present or existing both on the basis of the map data and on the basis of the sensor data. If the sensor data match the map data, it can be assumed that this element has a very high probability of existence, so that no further verification is necessary. However, it is preferred that these map elements are nevertheless checked with the help of the rule database with regard to their (logical) consistency in the overall context, ie when considering all available information or elements.

In addition, the map data provided can represent map data of an HD map described at the outset. Such an HD map can contain a great deal of high-resolution information about the surroundings of the motor vehicle, in particular relating to the course of streets and lanes, and the driving rules applicable on these respective streets and lanes, in particular traffic rules.

In order to provide the sensor data, the motor vehicle can have not only one environment sensor or an environment detection device, but also a plurality of environment sensors. The at least one environment detection device can also represent an ultrasonic sensor and/or lidar (light detection and ranging), for example a laser scanner, and/or a radar and/or a camera. In particular, the motor vehicle can include a plurality of such sensors in any combination and number. In particular, a 360° area around the motor vehicle can be detected by the at least one environment detection device.

The results of the method described with regard to the determined probabilities of existence can be transmitted to a backend, i.e. an Internet server and/or a service provider, for example for maps, which in turn manages map data and can update map data on the basis of the transmitted probabilities of existence. For example, an existing Map, like the map stored in the vehicle, is validated or updated by the service provider on the basis of the existence probabilities determined in this way. If, for example, it is determined on the basis of the method described that a possible element which was classified as present on the basis of the map nevertheless has a very low probability of existence, this element can be removed from the map. Conversely, an element that is not on the map but whose probability of existence was rated very high due to the procedure can be added to the map. The detections of individual vehicles can be aggregated in the backend and checked as to whether they are temporary (e.g. construction site) or permanent (e.g. new sign). This is made possible in particular if several vehicles transmit this information to such a backend independently of one another. The HD map cannot be changed (permanently) solely from the detections of a single vehicle. A temporary change in the map data in the vehicle based on the ascertained existence probabilities is nevertheless conceivable.

Depending on the ascertained probabilities of existence, a currently valid static environment model can be generated in the motor vehicle itself. This means that the HD map as such does not have to be changed, even for a single vehicle. Only a new model of the environment is formed from this information, which can only be valid for the current point in time. The real change, which then results in a map update for all vehicles, can be done in the backend instead.

However, it is particularly advantageous that the probability of existence finally determined can be made available directly to other systems in the motor vehicle. In particular, the existence probabilities determined in this way can also be used directly by other driver assistance systems of the motor vehicle, for example a system for scene interpretation or a system for trip planning as part of automatic driving. This also shows the great further advantage of the method according to the invention, because now it is possible by determining the existence probabilities of a subsystem, such as the scene interpretation mentioned and / or the journey planning, not only a map, in particular an HD map, with corresponding elements that may have been verified on the basis of the method described, but also to indicate the associated probability of existence for each element. The background to this is that not every element or its presence is equally important for every subsystem in the motor vehicle. Each subsystem also has its own requirements Demands on the reliability of the information used in an HD map, which is why the resulting hypotheses, i.e. the assessment of the presence of possible objects, can be assessed system-specifically and possibly element-specifically with regard to their probability of corresponding to the current vehicle environment. For example, information about a footpath parallel to the lane is of greater importance for the scene interpretation subsystem, as this provides support in the classification of dynamic objects as pedestrians, than for the path planning subsystem, since areas outside the driving corridor are not for trajectory planning are taken into account. Which requirements are now made by a respective subsystem on the respective probability of existence of the respective elements, for example which threshold values must be exceeded or fallen below by the probability of existence in order to be able to make reliable assumptions, is then left to the subsystems themselves or can be determined for the respective Subsystems are defined specifically.

Overall, the reliability in evaluating the probability of existence of elements in the area surrounding the motor vehicle can be significantly increased and the functioning of individual driver assistance systems and subsystems can also be significantly improved.

The traffic space, which is covered by the rule database relating to information, preferably does not represent a geographical space, but rather an abstract space with traffic-related elements, rules and other knowledge.

The information stored in the rule database can preferably be broken down as follows: context knowledge, maneuver knowledge, misconduct and traffic rules. Information relating to traffic regulations relates, for example, to the regulations laid down by the road traffic regulations. Context knowledge makes it possible, for example, to define context groups. For specific context groups, such as pedestrian crossings, freeway ramps, roundabouts, intersections and so on, a minimal set of elements that define this context group can be specified by the context knowledge. As such a minimal set of elements, the pedestrian crossing context group includes, for example, a zebra crossing and/or a pedestrian crossing sign and/or a pedestrian traffic light and/or an adjacent pedestrian walkway. If, for example, it is assumed on the basis of the map data that there is a pedestrian crossing at a certain point, which, however, was not detected on the basis of the sensor data, it can nevertheless be concluded that it is very likely that this pedestrian crossing still exists, if, for example, on the basis of the sensor data detected a pedestrian crossing sign and a pedestrian traffic light, i.e. many elements that are assigned to the pedestrian crossing context group, even though the pedestrian crossing or zebra crossing itself was not recognized. Furthermore, knowledge of maneuvers can define how movement patterns of other road users, which can be detected using the at least one detection device of the motor vehicle, can be interpreted. Such courses of movement, also called traces, can be used, for example, to derive the course of lanes, the presence of stop signs or the like. The rules of such conclusions can also be laid down in maneuver knowledge. For example, it can be specified there that stopping other road users means that priority is to be given at this point, that there is a zebra crossing, that there is a stop sign or something similar. Knowledge relating to misconduct can also be stored in the rule database. This means that when determining probabilities of existence, misconduct by other road users can also be taken into account as probable. For example, it can be assumed that many road users exceed speed limits. For example, if many other road users drive 90 km/h, a speed limit of 80 km/h can still be valid and should not be dismissed as improbable because the sensor data has been used to record that many other road users are a little faster than 80 km / hour drive. In other words, in the validation based on maneuvers, a certain tolerance that takes into account the misconduct of other road users can be acceptable.

The knowledge stored in the rule database can also be linked to one another in order to be able to derive further conclusions from it. If, for example, many other road users stop at a certain position, at which there should be a zebra crossing based on the map data, for example, but which was not directly detected on the basis of the sensor data, the stopping of other road users at this position can determine what is recorded on the basis of the sensor data can be concluded that the existence of the zebra crossing is nevertheless very likely. Thus, for example, maneuver knowledge can also be combined with the context knowledge defined above. For example, it is likely that at a particular point where many road users stop, is a zebra crossing and not a stop sign if, for example, other elements of the pedestrian crossing context group were also detected at this position.

These are just a few examples that illustrate how the knowledge or information provided by the rule database can be used to advantage in order to improve the determination of the probability of existence of possible elements in the environment of the motor vehicle, precisely when there are discrepancies between Sensor data and map data there. The rule database can also be stored in a memory of the motor vehicle and/or also be provided outside the vehicle, in which case the information can be retrieved from the vehicle at any time. It is theoretically conceivable that the method steps mentioned can also be carried out externally to the motor vehicle, with the exception of the provision of the sensor data. The result of the method, ie the probabilities of existence, can then be transmitted back to the motor vehicle accordingly. It is preferred, however, if the method is carried out entirely in the motor vehicle, since the sensor data is provided there and transmission of large amounts of data to a device external to the motor vehicle is therefore not necessary. This also advantageously makes it possible to make driving-related decisions based on the existence probabilities while driving in real time, since the calculation time for determining the existence probabilities is sufficiently short and no additional data transmission time occurs. In addition, the results of the method can also be used directly in the motor vehicle, especially in real time, which significantly increases safety in autonomous driving and significantly reduces the probability of having to put the motor vehicle into a “safe state”.

In a further advantageous embodiment of the invention, the discrepancy with regard to the presence of the at least one possible element is determined when comparing the map data and the sensor data if the at least one possible element was classified as a newly detected element. This means that the possible item was detected based on the sensor data and does not exist in the map data. The discrepancy is also determined if the at least one possible element was classified as an unevaluated element that is present in the map data and was not detected as being present based on the sensor data. This does not necessarily mean that this element was not detected at all based on the sensor data, but also, for example, that no evaluation could be made with regard to this element based on the sensor data, for example because this element was not seen by the sensors, for example because it was covered, or due to an error in the algorithm, or poor sensor data quality, e.g. due to bad weather, and so on. Especially in such a situation, in which a discrepancy defined in this way between the sensor data and card data is detected, it is particularly advantageous to have an additional assessment criterion, namely the rule database, which can be used to decide whether the possible element exists or not.

Furthermore, the at least one possible element can represent a physical element that relates to a physical object that may be present in the environment.

Examples of such physical elements are, for example, traffic lights, traffic signs, lane markings, lane markings, sidewalks, crosswalks or the like. Such physical elements can be added or removed, for example, in the case of a construction site. It is precisely in such a situation that lanes are often changed in their course, new traffic signs are erected or the priority rules are changed. The correct recognition of physical elements is therefore of great importance, especially in connection with autonomous driving.

However, not only the probability of existence of physical elements can be determined, but also that of semantic elements. Accordingly, it represents a further advantageous embodiment of the invention when the at least one possible element represents a semantic element, which represents a logical group of physical elements or at least one traffic rule. For example, additional information in the form of semantic elements can also be contained in maps, such as the map mentioned above, in particular HD maps. Such semantic elements can represent traffic rules, for example the right of way rules at an intersection. In addition, individual items, particularly individual physical items, can be grouped logically and assigned to a supergroup. This can also be referred to as composition. Such a supergroup can then also represent a semantic element. Relationships between elements can also be used and derived. For example, a dashed line and a solid line can be combined to form a lane. Furthermore, a traffic island and lanes can be combined to form a roundabout. Correspondingly, the lane and the roundabout represent semantic Elements. Semantic elements also include associations, that is, connections between elements, such as the connection between a traffic light and the lane to which the traffic light is assigned. Another example is the assignment of a stop line to a stop sign. In addition, as mentioned, priority rules also represent semantic elements. For example, priority rules are also stored as semantic elements for each intersection in the map. However, such semantic elements can not only be stored in the map, but they can also be detected by the at least one detection device of the motor vehicle or be derived from the detected sensor data. For example, the individual physical elements that can be assigned to a semantic supergroup can also be detected by the motor vehicle sensors. If, for example, a traffic island and lanes are detected by the at least one surroundings detection device, a roundabout as a semantic element can be considered to have been detected by the at least one surroundings detection device, more precisely to be considered to have been detected by the algorithm that processes the sensor data of the surroundings detection device and, as a result, the roundabout classified as recorded. This means that not only physical elements can be evaluated with regard to their existence, but also numerous semantic elements that are also relevant to traffic and driving.

In addition, according to a further advantageous embodiment of the invention, it is provided that if the possible element was classified as a newly detected semantic element on the basis of the comparison, the probability of existence of the possible element is determined as a function of the presence of newly detected physical elements. An assessment of the newly detected physical elements with regard to their probability of existence may have preceded this, so that only newly detected physical elements that probably exist are taken into account. If new elements are detected on the basis of the sensor data that are not contained in the map data, the newly detected physical elements can first be evaluated with regard to their probability of existence, for which purpose, as described, the rule database can be used. Newly detected semantic elements, which were also detected by the motor vehicle sensor system but are not yet contained in the map data, can also be evaluated with regard to their probability of existence depending on the first information provided by the rule database. In addition, however, the newly detected physical elements, at least those that probably exist, can also be included in the determination of their probability of existence. If, for example, a zebra crossing and a pedestrian traffic light are newly detected as physical elements and assessed as likely to exist according to the method, then the probability for the newly detected semantic element pedestrian crossing can also be regarded as high. As a result, the assessment and determination of the relevant existence probabilities can be further improved.

The procedure as described with regard to the at least one possible element can be implemented analogously in particular for all possible elements contained in a map or a map section in a surrounding area, for which sensor data provided by the at least one surrounding sensor or the at least one surrounding area detection device , present.

The aforementioned first surrounding area and the second surrounding area can be identical, but also different from one another, but have at least one overlapping area, since both relate to a surrounding area that includes the at least one possible element. The second surrounding area is typically smaller than the first surrounding area and the second surrounding area can also be completely surrounded by the first surrounding area. This is due to the fact that very large geographical areas can be covered by a map, while sensor data can typically only be provided in a small area around the motor vehicle, at least at a current point in time. The comparison of the map data and the sensor data can then correspondingly be limited to this overlapping area of the environment for which both map data and sensor data are available.

According to the invention, when comparing the map data and sensor data, a comparison is carried out in an overlapping area of the first and second surrounding area with regard to all elements present in the overlapping area based on the map data and/or sensor data, with each element of all elements present in the overlapping area based on the map data and/or sensor data depending on a result of the comparison, it is classified by being assigned to one of at least the following groups: a first group of the matching elements if the element is both detected based on the sensor data and is present in the map data; a second group of deviating, in particular newly detected, elements if the element is based on the sensor data has been recorded but is not present in the map data; and a third group of the unassessed items if the item is present in the map data and was not detected as being present based on the sensor data. In particular, such unevaluated elements relate only to map elements that could not be detected due to occlusions or errors in the algorithm based on the sensor data. Because here an uncertainty remains as to whether the element exists or not. Instead, a map element that is not detected by the sensors with a "free view" and an error-free algorithm is very likely not (or no longer) existent and therefore counts as a subgroup to the other deviating elements and not to the third group of non-evaluated elements defined here. In this case, too, it applies again that this does not imply that this element is classified as not present on the basis of the sensor data, but on the contrary that the presence of the element cannot be reliably determined on the basis of the sensor data. The element in question can be covered by an object, for example, so that the detection area to be detected, in which the element is located or should be located, can then be detected directly by the detection device.

In this way, the elements present in the map data, as well as all elements recorded on the basis of the sensor data, can be divided into one of these groups. In this case, for example, the group of newly detected elements can represent a subgroup of the deviating elements. In addition, a group can also be provided for other elements that cannot be assigned to one of the groups mentioned above. However, the above groups usually cover all possible cases that may arise. The elements of the second and third group represent possible elements, since elements assigned to these groups were not recorded on the basis of both the map data and the sensor data, so that there is a certain discrepancy between the map data and sensor data with regard to these elements , which may question the existence of these elements. With the elements of the first group, on the other hand, it can be assumed that they also exist in all probability, since their existence is confirmed both on the basis of the map data and on the basis of the sensor data.

The elements contained in the map data and the elements detected on the basis of the sensor data can be filtered according to the above grouping. These element groups with the assigned elements contained therein can then be transferred to the other modules in order to subsequently determine the respective existence probabilities, which is explained in more detail below. In principle, the elements in the respective groups can also be broken down into physical elements and semantic elements, as defined above. According to a further advantageous embodiment of the invention, each element that has been assigned to the second or third group represents a possible element whose probability of existence is determined. The determination of the existence probabilities for possible physical elements is carried out by a first evaluation module, to which all physical elements of the second and third group are transferred, and the first evaluation module calculates the existence probabilities for all newly detected physical elements and all physical elements that have not been evaluated as a function of determined from the transferred elements. The probabilities of existence are determined in particular as already explained above, in particular using at least one piece of first information from the rule database. Furthermore, the determination of the existence probabilities for possible semantic elements is carried out by a second evaluation module, to which all physical elements of the second group and all semantic elements of the third group are transferred, with the second evaluation module determining the existence probabilities for all non-evaluated semantic elements and all newly detected ones semantic elements, i.e. possible new semantic connections, determined depending on the transferred elements. In this case, too, the probability of existence is determined, as already described above, in particular again using the at least one piece of first information from the rule database. In this case, too, it is particularly advantageous to also take into account the previously determined, newly detected physical elements when determining the probability of existence of the possible semantic elements, in particular the newly detected semantic elements. It is precisely through such newly detected physical elements that new semantic connections can arise, which in turn lead to newly detected semantic elements or increase the probability of their existence. Thus, when evaluating the probability of existence or determining it for a specific possible element, not only the map data and sensor data relating to this element and the rule database and the information contained therein play a role, but also the maps and sensor data relating to other elements. In other words, not only each element is considered individually, but the elements in a possible network among themselves. In this way, additional information about elements can be obtained, which allows a statement about their probability of existence.

Finally, a consistency check can also be carried out, which checks the results obtained in this way for possible contradictions. Accordingly, it represents a further advantageous embodiment of the invention if, following the determination of the existence probabilities, a consistency check is carried out by a checking module, with the checking module being given all the elements of the first group and all the elements assessed as likely to exist by the first and second assessment modules whose existence and/or coexistence is checked by the checking module for consistency based on at least one piece of second information provided by the rule database, wherein in the event of a detected contradiction relating to at least one of the elements, the element in question depends on at least one piece of information provided by the rule database third piece of information is reassessed with regard to its probability of existence.

This last step advantageously allows inconsistencies to be uncovered and, as a result, elements that may have been evaluated incorrectly to be discovered. Incorrect assessments can thus advantageously be avoided or at least their extent can be greatly reduced. This also enables a significant increase in the reliability of the results.

Optionally, the input data can also be filtered, in particular with regard to data that may contain errors. For example, the map data provided by the maps can be compared in advance for consistency with the information and rules contained in the rule database, in order to remove logically invalid connections from the HD map, for example. This is due to the fact that, in principle, HD map elements are available whose (logical) connections to one another are not valid per se. Examples of this are associations between traffic lights and lanes. Reasons for errors in existing connection hypotheses can be manual processing, in particular due to carelessness and ignorance, as well as errors in the map creation algorithm. Connections between map elements represent a system-specific safety-relevant aspect. For example, the awareness of the semantic grouping roundabout has an influence on the interpretation of the scene and the trajectory planning, since other traffic regulations apply. The initially existing connection hypotheses between elements provided by the map data can thus be checked for inconsistencies by means of the rule database. This also makes it possible to further increase the reliability of the method. However, it is preferred that such a prior check of the card data is not carried out because it can be assumed that the HD cards initially, ie coming from the card manufacturer, have no logical inconsistencies.

Furthermore, the invention relates to a driver assistance system for a motor vehicle, the driver assistance system being designed to carry out a method according to the invention or one of its embodiments. In addition, the invention also includes a motor vehicle with such a driver assistance system. The driver assistance system can represent, for example, an assistance system for autonomous driving of the motor vehicle.

The invention also includes developments of the driver assistance system according to the invention, which have features as have already been described in connection with the developments of the method according to the invention. For this reason, the corresponding developments of the method according to the invention are not described again here.

The invention also includes the combinations of features of the described embodiments.

• 1 eine schematische Darstellung eines Kraftfahrzeugs mit einem Fahrerassistenzsystem zur Ermittlung einer Existenzwahrscheinlichkeit möglicher Elemente in einer Umgebung des Kraftfahrzeugs gemäß einem Ausführungsbeispiel der Erfindung; und
• 2 eine schematische Darstellung der Funktionskomponenten des Assistenzsystems gemäß einem Ausführungsbeispiel der Erfindung.

Exemplary embodiments of the invention are described below. For this shows:

• 1 a schematic representation of a motor vehicle with a driver assistance system for determining a probability of existence of possible elements in an environment of the motor vehicle according to an embodiment of the invention; and
• 2 a schematic representation of the functional components of the assistance system according to an embodiment of the invention.

The exemplary embodiments explained below are preferred exemplary embodiments of the invention. In the exemplary embodiments, the described components each represent individual features of the invention that are to be considered independently of one another, which also develop the invention independently of one another and are therefore also to be regarded as part of the invention individually or in a combination other than that shown. Furthermore, the described executions games can also be supplemented by other features of the invention already described.

Elements with the same function are each provided with the same reference symbols in the figures.

1 shows a schematic representation of a motor vehicle 10 with a driver assistance system 12 for determining an existence probability W of possible elements 14 in an environment 16 of the motor vehicle 12 according to an exemplary embodiment of the invention. The driver assistance system 12 can in particular represent an assistance system for autonomous driving of the motor vehicle 10 . A card 18, in particular an HD card 18, which includes information relevant to the driver, is used to implement automatic driving functions, but possibly also for other purposes. These include roadways, lanes, in particular lane topologies, directions of travel, which are assigned to the respective lanes, priority rules and so on. In the present case, this map 18 is stored, that is to say stored, in a memory 20 of the motor vehicle 10, in particular a memory 20 which is assigned to a control device 22 of the motor vehicle 10. The control device 22 is otherwise designed to carry out the method described in more detail below, in particular to determine the existence probabilities W of the possible elements 14 in the surroundings 16 of the motor vehicle 10 . Furthermore, the motor vehicle 10 can have a system 24 for self-localization within the HD map 18 . This system 24 may include a GPS receiver. Above all, this system 24 preferably uses the sensor data D provided by at least one surroundings detection device 26 of motor vehicle 10. Several such surroundings detection devices 26, in particular sensors 26 of motor vehicle 10, are shown here as examples. However, the motor vehicle 10 can also include more or fewer such sensors 26 . Such sensors 26 can be embodied, for example, as a camera and/or laser scanner and/or radar and/or ultrasonic sensor. It is advantageous to use a plurality of sensors 26, in particular sensors of different types, for detecting the surroundings. The motor vehicle 10 can roughly determine its own position within the map 18 using the GPS sensor. The motor vehicle 10 , in particular the system 24 , can localize itself very precisely within the HD map 18 on the basis of the environment detection provided by the sensors 26 . This includes the assignment of the current position as well as the current orientation of the motor vehicle 10 within the map 18.

However, the sensor data D provided by the sensors 26 can be used not only for self-localization, but in particular also for checking the information about the surroundings 16 of the motor vehicle 10 provided by the map 18 . As mentioned above, the map contains 18 relevant information about the environment 16 of the motor vehicle 10, which is traffic-related or driving-related in particular. Such information is hereinafter referred to as elements. Examples of such elements are in 1 a traffic sign 14a, in particular for speed limits, and a lane limit 14b illustrated. However, there are many other such elements 14 that are relevant to driving and that can be detected by sensors 26 and can also be stored in map 18 in the form of map data. Such elements 14 can include not only physical elements, as illustrated here, but also semantic elements. In addition to context-dependent relationships between physical elements, such semantic elements can also relate to traffic rules, for example.

Optionally, the motor vehicle 10 can also have a sensor system for detecting and recording traces of other road users, ie for recording the course of movement of other road users. In order to detect these other road users in the surroundings 16 of the motor vehicle 10 and to detect their course, the aforementioned surroundings sensors 26 can in particular also be used. The provision or determination of the course of movement of other road users on the basis of this recorded sensor data D can be carried out in advance by the corresponding sensors 26 and transmitted as a result to the control device 22, or determined by the control device 22 on the basis of the data D provided by the sensors become.

In general, both the sensor data D and the map data provided by the card 18 can be faulty or no longer correspond to the current status. This can occur, for example, when the road layout changes, for example due to a construction site. Correspondingly, it is advantageous to check the correctness of such map hypotheses, which postulate the existence of certain elements 14 in the area 16 of the motor vehicle 10 . For this purpose, map deviations are first detected. This is performed by a system 28 for map anomaly detection and filtering. This system can, for example, only be a functional component of the control device 22 represent. For this purpose, ie for the purpose of detecting map deviations, the sensor data D are compared with the map data provided by the map 18 relating to the same overlapping area of the surroundings 16 . The results of this comparison can be filtered and divided into different groups. If elements in the environment 16 are detected both on the basis of the sensor data D and on the basis of the map 18, then these elements 14 represent matching elements and can be assigned to a first group. If elements 14 are detected on the basis of the sensor data D, but are not present in the map 18, then these can be assigned to a second group, namely the group of deviating, in particular newly detected, elements. If elements 14 are present in the map data of map 18 but were not detected as being present on the basis of the sensor data, elements of a third group, namely the group of non-evaluated elements, are assigned for this purpose. The elements 14 filtered in this way can now be passed on to further modules in order to determine the respective existence probabilities W. A first evaluation module 30 is provided, which can also represent a functional unit of the control device 22 . This first evaluation module 30 assumes the task of determining the existence probabilities W for all newly detected physical elements and all physical elements that have not been evaluated. For this purpose, all physical elements of the second and third group are passed to this first evaluation module.

Furthermore, a second evaluation module 32 is provided, which can likewise represent a functional unit of the control device 22 . All physical elements of the second group and all semantic elements of the third group are transferred to this second assessment module 22, with the second assessment module 32 determining the existence probabilities W for all non-assessed semantic elements and all newly detected semantic elements. Following the determination of the existence probabilities W, a consistency check can also be carried out by a checking module 34 . This checks all elements ultimately assumed to be likely to exist, ie all elements of the first group and all elements assessed as likely to exist by the first and second assessment module 30, 32, with regard to whether their existence and/or coexistence may lead to contradictions. When any contradictions are uncovered, such individual elements 14 can be checked again in a targeted manner, and ultimately the output result and such contradictions can be corrected. As a result, for example, the existence probabilities W determined for the respective elements 14 can then be output. If, for example, it should have been determined during the course of this procedure that the existence of some elements 14 contained in the map 18 was assessed as improbable, this can also be output as a result, and vice versa, if some elements 14 were determined to be likely to exist, which are not contained in the map data of the map 18, however.

The great advantage of the invention is that the determination of the existence probabilities W and the checking of the map hypotheses is not carried out solely on the basis of a comparison of sensor data D with the map 18, but that when determining the existence probabilities W precisely when there are discrepancies , that is to say for the detection of map deviations, a rule database 36, which can also be stored in the memory 20, for example, is used. Such a rule database 36 can also be referred to as a knowledge database or ontology, for example. Such a rule database 36 can also have been created in advance, ie offline. The aspects relevant to driving safety, for example element groups in an intersection, element groups in a roundabout, permissible combinations of elements 14, traffic rules and so on, are summarized in such a knowledge database using a formal description logic. Here, basic elements 14 and possible relationships between these elements 14 are defined in a subject area, ie in the entire traffic area. In addition, rules are established as to how the elements 14 and relationships can be combined for the expansion of the knowledge database 36. The ontology thus represents the tool which enables logical inference, ie the conclusion of new facts about existing elements 14 by comparison with previous knowledge, as well as plausibility checks, ie the examination of the modulated hypotheses for consistency. The information stored in the rule database 36 can be subdivided, for example, into the four groups of context knowledge, maneuver knowledge, knowledge about incorrect behavior and traffic rules. Context knowledge specifies a minimum set of elements 14 for specific context groups, such as pedestrian crossings, freeway entrances, roundabouts, intersections and so on, which define these context groups. Maneuvering knowledge describes the rules according to which conclusions can be drawn from the recorded movements of other road users. If, for example, lanes cannot be detected by the sensors 26 or only poorly, the traces of other road users can be used. from theirs Driving behavior can be derived from the course of lanes, the presence of certain traffic signs such as stop signs, traffic lights or the like. The rules for deriving these conclusions can be defined accordingly in the maneuver knowledge mentioned.

The knowledge of misconduct takes into account that other road users do not necessarily always follow the prescribed traffic rules exactly, as is often the case with speed limits. When deriving conclusions, this can be taken into account accordingly by a certain specified tolerance. The rule database 36 also contains all valid traffic rules, signs and their meanings that apply in the current environment 16 in which the motor vehicle 10 is currently located.

If there is now a discrepancy between the sensor data D and the card data of the card 18 with regard to the existence of a specific element 14, this rule database 36 can now advantageously be accessed in order to make a decision as to whether this element 14 in question now probably exists or non-existent. With the help of this knowledge stored in the rule database 36, a significantly more reliable and more precise determination of existence probabilities W of the elements 14 in question in the surroundings 16 of the motor vehicle 10 can be provided. In particular, the named first and second evaluation modules 30, 32 can fall back on the knowledge stored in this rule database 36 when determining the respective existence probabilities W. The consistency check carried out at the end, which is carried out by the checking module 34, can also be carried out with the aid of this rule database 36. This allows a reliable interpretation of all available information, ie the sensor data D and the map data of the map 18 with regard to their rule conformity and probability of corresponding to the current reality. In addition, this procedure allows the provision of corrected HD map hypotheses in the form of corrected map data at runtime in a particularly advantageous manner, so that these corrected map data can be used directly by the driver assistance system 12 in order to continue to maneuver the motor vehicle 10 correctly autonomously and to carry out this autonomous driving not to abort due to excessive uncertainties regarding the interpretation of the environment and detection.

2 shows again a schematic representation of the functions that can be executed by the driver assistance system 12 according to an exemplary embodiment of the invention. Here, in particular, the individual functional units of the method described are shown.

In particular, map data 18a and sensor data D are first provided here. Map deviations are then detected in function module 28a. This is done by comparing the map data 18a with the recorded sensor data D. The results of this comparison can then be subjected to filtering in a filter module 28b. This filter module 28b can carry out the above-described assignment of individual elements 14 to the groups mentioned. In this way, elements 14 relevant to the subsequent modules can be filtered out, in particular evaluated map elements and connections, which are necessary for aspects relevant to driving and safety. Such filtering reduces the amount of data and computation. Furthermore, these filtered elements 14 are then supplied to the subsequent modules for further processing. First of all, an evaluation of physical elements 14 is carried out by the first evaluation module 30 . The element groups "unassessed physical elements 14" and "newly detected elements 14" are transferred. The following two sub-tasks are then carried out in the first assessment module 30 using the knowledge database 36: on the one hand, the estimation of the existence probability W of non-assessed physical elements 14, and on the other hand the assessment of newly detected physical elements 14 with regard to their existence. As a result, respective probabilities of existence W are then determined according to a specified metric 38, which can be specified and used as a basis for the module 40 for modeling the environment. An evaluation of semantic elements 14 is then carried out in the second evaluation module 32 . For this purpose, individual elements 14, in particular physical elements 14, can be logically grouped and assigned to a supergroup, and relationships between elements 14 can be used and produced. Various elements 14 can also be associated with one another, such as traffic lights and corresponding lanes. Priority rules can also be part of these semantic elements 14 . For example, priority rules are also stored as semantic elements 14 for each intersection in the map 18 . In this way, new connections between map elements can also be established, in particular by reconstructing map elements that are partially covered. With the help of the ontology, it is possible to use the spatial interpolation based on known rules and previous knowledge, for example the reconstruction of a partially covered solid or dashed lane marking to form a lane limit to provide. Connections, in particular new connections, can also be established between map elements by composing map elements. The ontology allows both "high level" grouping of physical elements 14, such as lane markings into lanes, and "low level" assembly of semantic elements 14 themselves, such as traffic islands and lanes into roundabouts. Map elements can also be associated with one another. This also makes it possible to create new connections between map elements. The association of map elements is also made possible by the ontology. Logical relations and contexts explicitly formulated in the ontology, for example the arrangement and prescribed distances between elements 14, allow conclusions to be drawn about the logical association of elements 14, for example traffic signs to lanes or light signals to lanes. Traffic regulations can also be inferred. The ontology allows the logical inference of new information through combinatorics of explicitly formulated relationships, for example an intersection without traffic signs means right before left, and a sign at the entrance to a town means a speed limit of 50 km/h.

The element groups passed to the second assessment module 32 are the non-assessed semantic elements 14 and the subset of the newly detected physical elements 14. The second assessment module 32 performs the following two subtasks using the knowledge database 36: Estimation of the existence probability B of non-assessed ones semantic elements 14 and the estimation of the existence probability W of new semantic traffic rules using the newly detected physical elements. Here, too, the respectively ascertained probabilities of existence W are output in accordance with the specified metric 38 . In the last module 34, a consistency check is also carried out. All correctly mapped elements are transferred to this checking module 34, ie elements that were detected based on the sensor data D as well as on the basis of the map data 18a, as well as elements evaluated by the other two modules 30, 32. In this checking module 34, inconsistencies are to be uncovered and corrected using the knowledge database 36. For example, it can happen in the area of a construction site that two traffic lights are detected, a normal one and a construction site traffic light, and both are classified as very likely to exist. However, in the current situation, the normal traffic light is not valid. However, a corresponding message arranged at the traffic light was not recognized by the sensor system 26 of the motor vehicle. As a result, these two different traffic lights are classified as existing and valid at the intersection. Due to the consistency check with the help of the knowledge database 36, a contradiction can be derived from this, since two traffic lights cannot apply to one lane at the same time. This rule can be formulated and stored by the knowledge database 36, for example. In addition, further knowledge from the knowledge database 36 can be used, which specifies, for example, that a yellow roadworks traffic light has priority over other conventional traffic lights. This information provided by the knowledge database 36 can advantageously be used in this present situation in order to “eliminate” the normal traffic light. In other words, as a result, the normal traffic light can be assumed to be unlikely to exist, at least as far as its validity is concerned.

After this consistency check, the finally provided existence probabilities W and in general the ascertained hypotheses can be made available to other systems, on the one hand inside the vehicle, but also on the other hand outside the vehicle. Inside the motor vehicle, the existence probabilities determined in this way can be used by the respective subsystems 42 that are 2 are illustrated only schematically. Such subsystems can, for example, represent a path planning of the driver assistance system 12 for autonomous driving and/or also a scene interpretation. A map validation and a map update can also be provided in the vehicle 10 . However, the results of the method can also be transmitted to a backend 44, which also uses the transmitted data to update map data, which can then in turn be made available on motor vehicle 10 as a new map. The backend 44 can also collect the data from many vehicles and then update the map data from many vehicles with sufficient probability. However, the card verification made possible by this method in motor vehicle 10, which contains error metrics for compliance with safety goals for automatic driving as an overall system, is particularly advantageous. Because this allows errors or changes in the map data to be determined reliably and quickly, especially in real time, so that safe and autonomous driving is possible. Overall, the invention enables both offline and real-time solutions. The knowledge base 36 is also scene-independent, that is to say it models traffic regulations in their entirety and not just for a sub-area of road traffic, such as only in relation to an intersection. Furthermore, a system-specific metric can also be used Specification of a reliability of the hypotheses generated, ie the existence probabilities W, are provided and used. The map elements can be advantageously divided and assembled into semantic groups such as intersections, roundabouts, and modeled in detail. Furthermore, a clear reference to a map format is made possible, which also allows the modeling of temporary changes in an HD map. In addition, driving and safety-related map elements can be prioritized and the focus can be placed on safe execution of the driving function and the reduction of data and computing effort. The combination of HD map, sensors and explicitly formulated knowledge through the knowledge database also enables inference beyond the sensor range. In addition, this method can also be used in a particularly advantageous manner for temporary and short-term changes, such as at a construction site.

The invention thus advantageously enables an automated process for providing corrected map hypotheses while driving. For this purpose, a knowledge-based model, i.e. the rule database, or more precisely, explicitly formulated knowledge about all aspects relevant to driving and safety, is used to interpret and check the consistency of all available information. This enables the driving function to be maintained even if there are deviations from the map, both in terms of physical and semantic elements. This preferably includes the provision of current and validated map hypotheses including all information required for the driving function, e.g. B. current lane situation and priority regulations.

Reference List

10

motor vehicle

12

driver assistance system

14

element

14a

traffic sign

14b

lane boundary

16

Vicinity

18

Map

18a

card data

20

Storage

22

control device

24

Self-Location System

26

environment sensor

28

System for detection of map deviations and filtering

28a

Module for detecting map deviations

28b

filter module

30

first assessment module

32

second assessment module

34

verification module

36

rules database

38

Metric

40

Environment modeling module

42

subsystem

44

backend

D

sensor data

W

probability of existence

Claims (9)

Method for determining an existence probability (Vη of at least one possible element (14) in an environment (16) of a motor vehicle (10), having the steps: - providing map data (18a) of a map (18) of the environment (16) of the motor vehicle ( 10), which relate to a first surrounding area, which comprises the at least one possible element (14); - providing sensor data (D) by means of at least one surrounding area detection device (26) of the motor vehicle (10), which relate to a second surrounding area, which comprises the at least one possible element (14); and - comparing the map data (18a) and the sensor data (D) with regard to the presence of the at least one possible element (14); - determining the existence probability (W) as a function of a result of the comparison; and - Outputting the probability of existence (W), - the at least one possible element (14) being at least based on the map data (18a) and/or the sensor data ten (D) is classified as present; characterized in that a rule database (36) is provided, in which information relating to a traffic area is stored, at least in the event that when comparing the map data (18a) and the sensor data (D) there is a discrepancy with regard to the presence of the at least one possible Element (14) is determined, the existence probability (W) is determined as a function of at least one piece of first information provided by the rule database (36), wherein when comparing the map data (18a) and sensor data (D), a comparison in an overlapping area of the first and second surrounding area relating to all elements (14) present in the overlapping area based on the map data (18a) and/or sensor data (D), each element (14) of all in the overlapping area based on the map data (18a) and /or sensor data (D) existing elements (14) is classified depending on a result of the comparison by assignment to one of at least the following groups: - a first group of matching elements if the element (14) both on the basis of the sensor data (D) detected and also present in the map data (18a); - a second group of deviating elements if the element (14) was detected on the basis of the sensor data (D) but is not present in the map data (18a); - a third group of the unevaluated elements if the element (14) is present in the map data (18a) and was not detected as being present on the basis of the sensor data (D); wherein the at least one possible element (14) is assigned to the second group or third group. procedure after claim 1 , characterized in that the discrepancy is determined if the at least one possible element (14) was classified as a newly detected element (14) which was recorded on the basis of the sensor data (D) and which is not present in the map data (18a). , or if the at least one possible element (14) was classified as an unevaluated element (14), which is present in the map data (18a) and was not detected as being present by the sensor data (D). Method according to one of the preceding claims, characterized in that the at least one possible element (14) represents a physical element (14) which relates to a physical object possibly present in the environment (16). Method according to one of the preceding claims, characterized in that the at least one possible element (14) represents a semantic element (14) which represents a logical group of physical elements (14) or at least one traffic rule. Method according to one of the preceding claims, characterized in that if the possible element (14) was classified as a newly detected semantic element (14) on the basis of the comparison, the probability of existence (W) of the possible element (14) depending on the presence of newly detected physical elements is determined. Method according to one of the preceding claims, characterized in that each element (14) which has been assigned to the second or third group represents a possible element (14) whose probability of existence (W) is determined, with - the determination of the probability of existence (W) is carried out for possible physical elements (14) by a first evaluation module (30), to which all physical elements (14) of the second and third group are passed, the first evaluation module (30) evaluating the existence probabilities for all newly detected physical elements (14) and determining all unevaluated physical items (14) dependent on the submitted items (14); and - the determination of the existence probabilities (W) for possible semantic elements (14) being carried out by a second evaluation module (32), to which all physical elements (14) of the second group and all semantic elements (14) of the third group are passed, wherein the second evaluation module (32) determines the existence probabilities (W) for all non-evaluated semantic elements (14) and all newly detected semantic elements (14) as a function of the transferred elements (14). Method according to one of the preceding claims, characterized in that following the determination of the existence probabilities (W), a consistency check is carried out by a checking module (34), the checking module (34) receiving all elements (14) of the first group and all are transferred from the first and second evaluation module (30, 32) to elements (14) rated as likely to exist, whose existence and/or coexistence is checked by the checking module (34) for consistency based on at least one second piece of information provided by the rule database (36). in the event of a detected contradiction relating to at least one of the elements (14), the element (14) in question is re-evaluated as a function of at least one piece of third information provided by the rule database (36) with regard to its probability of existence (W). Driver assistance system (12) for a motor vehicle (10), wherein the driver assistance system (12) is designed to carry out a method according to one of the preceding claims. Motor vehicle (10) with a driver assistance system (12). claim 8 .

Images