METHOD FOR OPERATING A DRIVER ASSISTANCE SYSTEM FOR A VEHICLE ON A ROAD AND DRIVER ASSISTANCE SYSTEM

Description

[0001] The present invention relates to a method for operating a driver assistance system for a vehicle on a road and a driver assistance system for a vehicle on a road.

[0002] Modern vehicles are increasingly equipped with driver assistance systems that allow at least partially automated vehicle control. For automated driving, the system must be able to perceive the vehicle's surroundings, interpret them, and make driving decisions based on this interpretation. Intersections are particularly important in this regard, as they often require assessing a particularly complex situation. Traffic regulations, such as those defined by traffic law or specific regulations at the respective intersection, must be taken into account.

[0003] The method proposed in DE 10 2005 0220 429 A1 for assisting a driver when crossing an intersection involves first dividing the intersection into several zones. For each zone, a situation-dependent status value for the navigability is determined based on environmental information, and a suitable driving movement for crossing the intersection is defined before entering the intersection. A driver warning can be issued in the event of danger.

[0004] DE 10 2014 011 117 A1 describes a driver assistance system in which an object located on a digital map can be highlighted with light using a lighting device. Such objects require the driver's special attention; furthermore, areas on the road, such as intersections, can be highlighted.

[0005] The method described in US 2016 / 0293004 A1 for controlling and monitoring traffic lights to optimize waiting times involves traffic lights regulating the passage of vehicles. They can communicate with vehicles and communicate the optimal speed for continuous movement.

[0006] The document DE102013005362 describes a method for detecting a traffic situation.

[0007] The present invention is based on the object of providing a method for operating a driver assistance system for a vehicle on a road and a driver assistance system which enables a safe and rapid interpretation of a traffic situation.

[0008] According to the invention, this object is achieved by a method having the features of claim 1 and a driver assistance system having the features of claim 13. Advantageous embodiments and further developments emerge from the dependent claims.

[0009] In the method according to the invention of the type mentioned above, environmental data about the road and the surroundings of the vehicle are recorded. Based on the recorded environmental data, at least one traffic control for a road section located in front of the vehicle in the direction of travel is recorded and assigned to the vehicle. Furthermore, a passability assessment is generated and output for the road section based on the traffic control assigned to the vehicle. The passability assessment is further generated based on a confidence parameter, which is determined based on the recorded environmental data and comprises a measure of the complexity of a control situation for the road section.

[0010] By taking the confidence parameter into account, a particularly reliable assessment of the situation, particularly at an intersection, can advantageously be achieved, and driving decisions based on the assessment can be made with particular confidence. Furthermore, the method according to the invention allows for particularly rapid processing of the environmental data used to assess the situation. This assessment can be performed, in particular, without requiring a complete interpretation of the data acquired regarding the control situation, or even if the passability assessment cannot be correctly determined. The assessment is therefore also performed particularly quickly and can be carried out in real time when available resources are limited.

[0011] The environmental data is collected in a known manner, for example by means of a camera, a stereo camera, a laser scanner and / or another type of detector. Alternatively or additionally, a method for communication between a vehicle and other facilities of the traffic infrastructure ( Car2X ), whereby, for example, information about traffic light phases or other traffic regulations is transmitted directly to the vehicle via a corresponding interface.

[0012] According to the invention, the environmental data relate to the road and the surroundings of the vehicle. This means that the acquired environmental data are acquired in such a way that properties of the road are recorded within a specific space in the surroundings of the vehicle, which depends on the respective detector used. In particular, the environmental data are recorded within a specific radius around the current position of the vehicle. The environmental data comprise information about traffic infrastructure facilities, for example traffic structures, or traffic control facilities such as traffic lights, traffic signs and road markings, as well as optionally about other characteristics of the environment, such as vegetation, buildings and / or others. Based on the acquired environmental data, at least one traffic control is recorded in a further step.For this purpose, traffic lights, traffic signs, road markings, and / or other devices that signal instructions for the behavior of road users are recognized. It can also be provided that a traffic regulation implemented by a person, such as the police or a traffic marshal directing traffic, is recognized. Detection occurs in such a way that traffic regulations are recorded for a specific section of the road that is located in front of the vehicle in the direction of travel. The traffic regulation can therefore be spatially related to the road section, for example, if a device that signals the regulation is located in the immediate vicinity of the road section or in front of it.

[0013] In particular, the presence of a device signaling traffic regulation is first detected, for example, a traffic light or a traffic sign. In a subsequent step, the signal emitted by the previously detected device is identified, for example, by capturing and evaluating a symbol depicted on a traffic sign and / or the shape and color of a traffic light signal. This allows different categories of traffic regulations to be detected, which differ, for example, in their impact on the passability of the road section.

[0014] The traffic regulation is then assigned to the vehicle. Particular consideration is given to the vehicle's location on the road and the section of the road to which the traffic regulation applies. In particular, lanes on the road are taken into account, and it is determined which lane the vehicle is in and which lane the traffic regulation is assigned to. In a further step, it can also be determined whether the traffic regulation is relevant for the specific vehicle, for example, in the case of traffic regulation for certain vehicle types, such as trucks.

[0015] Based on the analysis of the environmental data and the determination of the traffic regulations relevant to the vehicle and the road section, a passability assessment is generated, which is then output. The output can be provided, for example, via a display, with an output signal being transmitted to a display unit. Output within the meaning of the invention can also be understood as an output signal being transmitted via an interface to another device in the vehicle, for example, to a unit for automatically controlling the vehicle, a navigation unit, or another device for which information about the passability of a road section is relevant to the vehicle.

[0016] The passability assessment includes information about whether the vehicle is permitted to drive on the road section at a specific time. Its generation therefore corresponds to an interpretation of the data recorded about the road, in particular the road section, as well as about traffic control signals. In particular, it also includes information about how the road section may be passed, for example at what speed. The passability assessment can also relate to areas of the road ahead of the road section, for example if the road section may not be driven on, for example because a traffic light prohibits entry, and the vehicle is to be brought to a stop in an area outside the road section. Furthermore, the passability assessment can include further information relating to the vehicle's journey in or through the area of ​​the road section.

[0017] According to the invention, the acquired environmental data includes road data about the section of road ahead of the vehicle in the direction of travel, as well as about one or more lanes of the road. This advantageously allows the traffic situation for the vehicle to be recorded and comprehensively analyzed with particular accuracy. Furthermore, the environmental data can be acquired in such a way that the lane the vehicle is traveling in can be identified particularly reliably and easily, and relevant traffic regulations can be assigned particularly accurately.

[0018] In particular, the road data includes information about the straight or curved course of the road, the positions of intersections or crossings, and other special features of the road course and of a traffic network to which the road belongs. Particular attention is paid to the area of ​​the road section considered in the method, but other sections of the road that may be related to the road section can also be considered. Using the road data, an intersection within the area defined by the road section can be detected or an intersection can be located. Furthermore, properties of the intersection can be determined, for example, the number of intersecting traffic routes and their relative positions and angles.

[0019] The road data also relate to lanes on the road, with particular reference to their number, markings, and / or arrangement. The road data is collected for the road section for which a passability assessment is to be generated using the method according to the invention; however, the data can also be collected for other sections of the road.

[0020] The road data includes information that can be considered for assigning a vehicle or a traffic regulation to a lane. This includes, in particular, data on the spatial arrangement of devices used to signal traffic regulations relative to the road, the individual lanes of the road, and the road section.

[0021] The reliability with which traffic regulations, lanes and the vehicle can be assigned to one another depends heavily on the complexity of the regulatory situation for the road section. For example, assigning a traffic regulation to a lane can be made more difficult if there are multiple lanes and / or multiple devices for signaling the traffic regulation. The complexity typically increases with the number of possible combinations for the lanes and traffic regulations in the area of ​​the road section. In addition, increased complexity can arise because it is more difficult to clearly assign a traffic regulation to a specific lane, for example because a traffic sign or traffic signal is relevant for multiple lanes. Conversely, multiple traffic signs or traffic signals can affect the same lane.

[0022] The method according to the invention therefore provides for a confidence parameter to be generated which reflects a measure of the complexity of the control situation and, in particular, indicates the degree of certainty with which the method leads to correct results for a specific control situation. In particular, the confidence parameter is a quantitative measure of complexity. The confidence parameter can be determined in different ways, with the environmental data being used to arrive at a description of the control situation and to determine the confidence parameter. Furthermore, any other data can be taken into account, for example by receiving pre-processed data from a device external to the vehicle, for example by providing a trust parameter for a specific road section via a central server or by providing data for determining it.

[0023] The confidence parameter can be used to provide further information about the degree of certainty with which the passability assessment is generated by the method. For example, when performing a passability assessment for a particularly critical area, where a conflict with other road users could occur, it may be appropriate to only accept a traffic regulation determined with a particularly high degree of certainty in order to thus implement autonomous control of the vehicle. Alternatively or additionally, the confidence parameter can be used to generate an output that informs a driver of the vehicle that the automatic determination of the passability assessment was carried out with, for example, a particularly high or low degree of certainty.

[0024] The confidence parameter contrasts with a standard deviation or confidence, which can be determined for individual process steps, such as the recognition of traffic regulations, their allocation to lanes, and their allocation to the vehicle. While the uncertainty determined in these steps can be taken into account when determining the confidence parameter, the latter is itself generated based on the overall recorded configuration of the road as well as the facilities surrounding the road that signal traffic regulations. This means that the confidence parameter can be used to estimate the degree of certainty with which a passability assessment, including the associated confidence values, can be generated in the respective situation.In particular, the confidence parameter can be determined before the traffic regulations are assigned to the lanes and before the vehicle is assigned to its lane. In contrast, a standard deviation or confidence for an individual step of the assignment process is only determined when the respective step is performed. According to the invention, the confidence parameter is determined without the specific determination of the assignment of the traffic regulations to the vehicle.

[0025] According to one embodiment of the method, map data about the road in the vicinity of the vehicle is also acquired, wherein the map data includes information about the number of lanes on the road, the position of intersections, and / or the position of traffic control devices. This advantageously allows for an analysis of the situation based on predefined map data, which can be particularly reliable.

[0026] The map data is acquired in a conventional manner, for example, using a vehicle's navigation system. The map data can be acquired from a non-volatile storage medium in the vehicle, a mobile unit, or an external unit, such as an external server. In particular, it includes information about a traffic network that includes the road and, in particular, the road section. The information includes, for example, absolute or relative positions, such as intersections, deviations of the traffic routes from a straight line, the number and arrangement of lanes, and other parameters, such as different types of traffic routes and roads.

[0027] For example, a position can be determined using the map data, in particular using a map-matching method. The method can also take into account any uncertainty in the map data, for example, by recording position information with an indication of its accuracy. The map data can be recorded alternatively or in addition to the street data, and the different types of data can complement each other.

[0028] In a further embodiment, the acquired environmental data includes ego position data, and based on the ego position data, an ego position probability distribution for the ego position of the vehicle is determined. This advantageously allows the accuracy of the determination of the vehicle's position to be taken into account.

[0029] The acquisition of the ego position data is carried out in a manner known per se, for example using a satellite navigation system (GPS), by localizing the vehicle using landmarks that are detected by the vehicle's sensors and are particularly included in the environmental data, or by other environmental data that are acquired, for example, by means of a stereo camera, a camera or a laser scanner (Lidar). Map-Matching- procedures are applied.

[0030] The ego position can be captured as an absolute position within a global coordinate system or as a relative position within a local coordinate system. For example, the ego position can be captured relative to a landmark. In particular, the position is captured relative to a width of the road, where the width is perpendicular to the longitudinal extent of the road.

[0031] The ego-position probability distribution comprises the probability that the vehicle is at a specific position, depending on a number of different potential positions. For example, if the ego position of the vehicle is detected at a specific position, the corresponding probability distribution exhibits a maximum at that specific position, while the probability is lower for other positions in the vicinity, with the distribution depending on the accuracy. Furthermore, the ego-position probability distribution can be determined for discrete values ​​of potential positions, in particular for specific areas of the road.

[0032] In a further development, the road has at least two lanes, and the ego position probability distribution is determined relative to the lanes of the road. This advantageously allows the ego position of the vehicle relative to the lanes to be determined with a quantitatively determined degree of certainty. This determines which lane the vehicle is in and the degree of certainty with which this statement can be made.

[0033] In particular, the lanes of the road as well as their arrangement and location are recorded using the environmental data, such as the recorded road data and / or map data.

[0034] The vehicle's position relative to the lanes is determined in a conventional manner, with a probability for the vehicle's ego position in the respective lane being determined. A distribution of this probability is determined for the various lanes of the road. Furthermore, the vehicle's ego position can be determined in such a way that a position within the lanes is also determined, for example, an ego position in the middle of a lane or offset laterally. A probability distribution can also be determined for this, or the probability distribution can be included in the ego position probability distribution.

[0035] In one embodiment, the detected traffic regulation is assigned to the vehicle based on the ego position probability distribution. This advantageously makes it particularly easy to determine the traffic regulation that must be taken into account for the vehicle, and also determines the degree of certainty with which this assignment is made.

[0036] In particular, in a first step, after the traffic regulation itself has been detected and recognized, it is determined for which lane of the road the traffic regulation is relevant. For example, a traffic signal system can be configured in such a way that it is assigned to only one lane. In this case, the traffic regulation signaled by it is assigned to the vehicle based on the distribution of probabilities for the ego position in the corresponding lane.

[0037] In one embodiment of the method, the acquired environmental data includes traffic control data, and to determine traffic control based on the traffic control data, a traffic control probability distribution is determined for a plurality of predefined traffic control categories. In this way, the traffic control data is interpreted based on a distribution of probabilities for the predefined categories.

[0038] The traffic control data includes, for example, data on traffic control devices, i.e., traffic signs, road markings, and traffic lights. Furthermore, other data relevant to the assessment of the vehicle's driving situation may be included, such as generally prescribed rules for driving situations where, for example, no explicit right-of-way rule is indicated by traffic signs. In particular, information about the regulatory content signaled by a traffic control device may be included, such as a specific regulation regarding permission to enter an intersection or a right-of-way rule. Furthermore, information relating to a specific assignment of the signaled traffic regulation to the vehicle may be included.

[0039] Various known methods can be used to collect traffic control data. In particular, vehicle sensors, such as a camera or a laser scanner, can be used, or data can be transmitted via a suitable traffic infrastructure, such as direct communication between traffic control devices and the vehicle ( Car-to-Infrastructure , Car-to-Roadside ), be used.

[0040] The traffic control categories can, for example, include various types of devices for signaling traffic control, such as road signs, lane markings, or traffic lights. This allows a distinction to be made between different categories to enable reliable evaluation of the different types of signals. The traffic control categories can be specified in various ways, for example, through a system preset or through another vehicle device, such as a driver assistance system.

[0041] In a further embodiment, an assignment probability distribution for assigning traffic control devices to the vehicle is determined based on the acquired environmental data. This advantageously determines the probability with which the signal from a specific traffic control device is relevant for the vehicle. In particular, this probability distribution includes the probabilities determined for different combinations of traffic control devices, lanes, and vehicles.

[0042] In one embodiment, the assignment probability distribution can be determined based on two different probability distributions, namely a first probability distribution for the assignment of a traffic control device to a lane and a second probability distribution for the assignment of the vehicle to a lane. These probability distributions include, in particular, for each lane of the road, the probability that a specific traffic control device or vehicle is assigned to that lane.

[0043] In particular, road data, traffic control data, and / or ego position data are taken into account when generating the assignment probability distribution. Furthermore, the probability distribution can be determined for all individual lanes of the road, whereby the probability of assigning a traffic control device to a lane and the probability of assigning a vehicle to the lanes is determined. The assignment is therefore based on the lanes.

[0044] Alternatively or additionally, the assignment can be made in another way, for example through a vehicle-specific assignment.

[0045] In a further development, the passability assessment includes a passability probability distribution. This advantageously allows for consideration of the degree of certainty with which the passability assessment can be determined.

[0046] The passability probability distribution is generated, in particular, for a plurality of predefined passability states. For example, the assessment can be performed based on passability states provided by a driver assistance system or another vehicle device. For example, the passability states can be used to distinguish whether a particular driving maneuver can be performed or not, and conditions for the maneuver can be determined.

[0047] In one embodiment, for at least one of the determined probability distributions, that is to say for the ego position probability distribution, the assignment probability distribution, the traffic control probability distribution and / or the passability probability distribution, a temporal smoothing is carried out using a smoothing parameter, preferably an exponential moving average ( exponential moving average ), particularly preferably a conditional moving exponential averaging ( conditional exponential moving average ). This advantageously compensates for temporal fluctuations in the detection of environmental data, for example, and allows for a consistent interpretation of the vehicle's surroundings and the road section, even in a temporal dimension.

[0048] Temporal variations in the determination of probability distributions can arise, for example, from a changing traffic situation. Furthermore, temporal fluctuations may need to be taken into account when collecting environmental data, for example, when a camera records data in a detection area that may be temporarily obscured by other road users, traffic structures, or other facilities. Further fluctuations can arise from changes in the vehicle's position relative to traffic control devices, for example, when passing a traffic light whose signal is only detectable by oncoming vehicles.

[0049] Smoothing is primarily achieved using a moving average, but other methods can also be used, such as a filtering method, which uses a smoothing parameter. This smoothing parameter determines how the system handles temporal changes, for example, to compensate for high-frequency fluctuations and give greater consideration to low-frequency fluctuations. In this way, fluctuations caused by short-term changes in the detection conditions can be ignored. On the other hand, actual changes in the current traffic situation, such as when a traffic light signal is switched, can be taken into account.

[0050] The preferred exponential averaging allows for conservative or non-conservative smoothing by choosing a suitable smoothing parameter. With conservative smoothing, a change in the underlying signal is only slowly incorporated, whereas with non-conservative smoothing, a change is incorporated quickly.

[0051] In a particularly preferred conditional moving exponential averaging, the smoothing parameter can be determined based on certain conditions, for example depending on the direction of the change of the signal, for example differently for a decreasing or increasing value.

[0052] Alternatively or additionally, the smoothing parameter can be calculated using different parameters for different conditions, for example, to apply different smoothing for different categories of traffic control devices using a different smoothing parameter. For example, it can be taken into account that a traffic control signaled by a traffic sign or road marking is typically constant over time, while traffic lights can emit alternating signals.

[0053] In a further development, a crossing state is determined based on the recorded environmental data, and the smoothing parameter is determined depending on the crossing state. This advantageously allows the vehicle's driving situation to be particularly well taken into account for the interpretation of the environmental data.

[0054] The crossing state is determined in a known manner for the road section under consideration, in particular an intersection. For example, different crossing states can be considered for approaching an intersection from a greater distance, approaching from a closer distance, and a state while crossing the intersection. Alternatively or additionally, further information, in particular other aspects of the recorded environmental data, can be used to improve the interpretation.

[0055] In a further development, the confidence parameter is determined using a machine learning method. This advantageously allows different information to be used as a basis for a reliable determination.

[0056] The determination can, for example, be made using a system that is trained using a previously known data set and thus recognizes how reliably the procedure, in particular the assignment of traffic regulations to the vehicle, functions in different situations, whereby the situations are characterized by the recorded environmental data and other data. In particular, no model needs to be specified that defines the determination of the confidence parameter through clear rules, but the machine learning method can be used to generalize decisions from the training data set, for example by means of a Support Vector Machine ( SVM ), a neural network or an artificial neural network (ANN) or another machine learning method.

[0057] Furthermore, it can be provided that data acquired during the execution of the method is used to further train the system using machine learning. In this way, a continuous refinement and improvement of the determination of the confidence parameter can be achieved.

[0058] The confidence parameter can be determined depending on different classes of traffic control devices, for example, by different confidence parameters for traffic signs, road markings, and traffic lights. In particular, it can be taken into account that an assignment of certain classes of traffic control devices can be made with varying degrees of certainty. For example, road markings can typically be assigned to a specific lane with a high degree of certainty, whereas this assignment can be made with less certainty for a traffic sign attached to the edge of a multi-lane road or a traffic light. Furthermore, multiple confidence parameters can be determined for different classes of traffic control devices, and / or one confidence parameter can be provided for a combination of different classes of traffic control devices.

[0059] In particular, the confidence parameter is calculated such that it decreases with an increasing number of lanes. Alternatively or additionally, it can also decrease with an increasing number of traffic control devices, especially of the same type. According to the invention, the confidence parameter decreases with an increasing number of possible combinations in which the traffic control devices can be assigned to the lanes. The confidence parameter then represents an estimate of the degree of ambiguity in the assignment.

[0060] During training, the passability rating is generated for a given set of driving maneuvers. This advantageously allows for a targeted assessment of individual driving maneuvers for the vehicle.

[0061] The set of driving maneuvers for which a passability assessment is generated can be specified, for example, by a device of the vehicle, such as a driver assistance system, a navigation device, or a device for at least partially autonomous control of the vehicle. Alternatively or additionally, the driving maneuvers can be provided by another device, in particular a vehicle-external device, such as a mobile device or an external server.

[0062] The set of driving maneuvers specifically includes various categories of interventions in the vehicle's control, which may include longitudinal and / or lateral control. Examples of driving maneuvers that may be included in the specified set are: turning right or left, crossing an intersection in a straight line, stopping at a stop line, changing lanes, accelerating or decelerating, and so on.

[0063] In a further development, a driver assistance system generates a planned trajectory for the vehicle. The specified set of driving maneuvers is generated based on this planned trajectory. Based on the passability assessment, a clearance signal is generated to drive along the planned trajectory. The clearance signal can, of course, also include a signal denying clearance or defining certain additional conditions under which clearance can be granted. This allows the passability assessment to be determined specifically for a specific trajectory and / or planned driving maneuver.

[0064] The driver assistance system can intervene in the vehicle's control in various ways. At the lowest level of automation, a driver can be supported in driving the vehicle by, for example, issuing information such as a warning or authorization, as with a turn assist system. At a higher level of automation, the driver assistance system can generate control signals and transmit them to specific devices to influence the vehicle's control in the longitudinal and / or lateral directions. For example, the vehicle can automatically accelerate or decelerate, or perform driving maneuvers with automatic assistance.At the maximum level of automation, the vehicle is controlled completely automatically, meaning that the driver is only required to monitor the automatic driving along a planned trajectory and can intervene manually if necessary, for example to prevent malfunctions.

[0065] During at least partially automated driving, the passability assessment can be used to determine whether a specific driving maneuver may be performed. In particular, automated execution only occurs if the passability assessment confirms safe passability. The confidence parameter can be used to assess the reliability of determining the passability assessment and, for example, to identify in which situations a driver should check the automatic passability assessment. Furthermore, situations can be identified in which control should be at least partially returned to the driver.

[0066] Furthermore, it can be provided that a driver assistance system detects which driving maneuver a driver of the vehicle intends to perform, while the vehicle is controlled at least partially manually. For example, it can be detected when the driver indicates an intention to turn by activating the turn signal or by making a specific steering angle. In this case, several driving maneuvers can also be determined that are optionally possible for the vehicle. The method according to the invention can then generate and output passability assessments for these various driving maneuvers and, if appropriate, alternative driving maneuvers determined by the driver assistance system.

[0067] The driver assistance system according to the invention for a vehicle on a road comprises a detection unit by which environmental data about the road and about the vehicle's surroundings can be detected, and a recognition unit by which, based on the detected environmental data, at least one traffic control for a road section located in front of the vehicle in the direction of travel can be detected. It further comprises an assignment unit by which the detected traffic control can be assigned to the vehicle, and an evaluation unit by which a passability assessment can be generated and output for the road section based on the traffic control assigned to the vehicle. The passability assessment can also be generated based on a confidence parameter, which can be determined based on the detected environmental data and comprises a measure of the complexity of a control situation for the road section.

[0068] The driver assistance system according to the invention is particularly designed to implement the above-described method according to the invention. The driver assistance system thus has the same advantages as the method according to the invention.

[0069] In one embodiment of the driver assistance system according to the invention, the passability assessment can be received by a control unit, and based on the passability assessment, a control signal can be generated, depending on which an at least partially automatic travel of the vehicle in the area of ​​the road section can be carried out. This advantageously enables particularly safe automatic control of the vehicle.

[0070] The invention will now be explained using embodiments with reference to the drawings. Figuren 1A und 1B show an embodiment of the driver assistance system according to the invention in a vehicle, Figuren 2A, 2B und 2C show various traffic situations in which the method according to the invention can be applied, Figuren 3A bis 3D show an embodiment of the method according to the invention and Figuren 4A, 4B und 4C show examples of conditional exponential averaging.

[0071] With reference to the Figuren 1A und 1B An embodiment of the driver assistance system according to the invention in a vehicle is explained.

[0072] A vehicle 1 comprises a detection unit 2 and a navigation unit 7, both coupled to a control unit 3. The control unit further comprises a recognition unit 4, an assignment unit 5, and an evaluation unit 6.

[0073] In the exemplary embodiment, the detection unit 2 comprises a camera and a laser scanner, as well as interfaces to other units, in particular a unit for retrieving information from a non-volatile data memory. Furthermore, a GPS module is included, which can be used to locate the vehicle 1 using a global navigation satellite system. Alternatively or additionally, localization can be achieved using other methods, for example, using landmarks or other systems.

[0074] Vehicle 1 is on road 10 and is moving in direction of travel F. Road 10 has two lanes 10.1, 10.2, with vehicle 1 in lane 10.2. Road 1 meets another road at an intersection. Road 1 also has lane markings 12, through which a first lane 10.1 is open for straight-ahead travel and a second lane 10.2 is designated as a left-turn lane 10.2. In the area of ​​the intersection, two traffic lights 13 are arranged on the right-hand side of road 1 in the direction of travel, one traffic light 13 relating to both lanes and the other traffic light 13 relating to the left-turn lane 10.2.

[0075] In the exemplary embodiment, the navigation unit 7 is configured to enable at least partially automatic control of the vehicle 1. This is done in a manner known per se, with automatic interventions in the longitudinal and lateral control of the vehicle 1 being performed. For example, the vehicle 1 can be accelerated or decelerated by the automatic control, interventions in the steering can be performed, and / or outputs can be generated that inform the driver of the vehicle 1 about measures for controlling the vehicle 1.

[0076] With reference to the Figuren 2A, 2B und 2C Various traffic situations are explained in which the method according to the invention can be applied. For simplicity, road sections are shown that include an area before a stop line. Other sections of the road are not shown here and can be formed in various ways, for example, as an intersection, pedestrian crossing, or in another way.

[0077] In the Figur 2A In the case shown, road 10 comprises a single lane 10.1, in the vicinity of which a traffic light 13 is located. This case represents a control situation of low complexity, since traffic light 13 is typically uniquely assigned to lane 10.1.

[0078] In the Figur 2B In the case shown, road 10 comprises three lanes 10.1, 10.2, 10.3, which further have lane markings 12, each of which designates a lane 10.1, 10.2, 10.3 for right-turning vehicles, for driving straight ahead, or for either driving straight ahead or turning left. Five traffic lights 13 are arranged in the vicinity of road 10. In the case shown here as an example, four of the traffic lights 13 are of the same design and relate to all three lanes 10.1, 10.2, 10.3, while the fifth traffic light 13 only regulates right-turning traffic, which in the exemplary embodiment is limited to the right lane 10.3.

[0079] The regulatory situation in the Figur 2B The case shown is more complex than the one described above with reference to Figur 2A This is the case explained above, where a vehicle intended to travel on the road section and located in one of the three lanes 10.1, 10.2, 10.3 can detect signals from five traffic lights 13 and must then identify the signal that is relevant from its own perspective. It is not necessarily immediately obvious which traffic light 13 is relevant for vehicle 1 in one of the lanes 10.1, 10.2, 10.3.

[0080] In the Figur 2C In the case shown, road 10 comprises five lanes 10.1, 10.2, 10.3, 10.4, 10.5, in the vicinity of which three traffic lights 13 are arranged. Here, too, a more complex control situation exists than in the case described above with reference to Figur 2A explained case, since for a vehicle similar to the one above with reference to Figur 2B In the case explained, the vehicle's own position on one of the five lanes 10.1, 10.2, 10.3, 10.4, 10.5 must first be detected and then the signals relevant to the vehicle on its lane must be detected.

[0081] In further embodiments, alternatively or in addition to the traffic lights 13 and the road markings 12, further traffic signs are provided, which, for the purposes of the following explanation of the embodiment, are treated in a common category with the road markings 12. In further embodiments, further categories of traffic control devices can be provided and treated in different ways.

[0082] With reference to the Figuren 3A bis 3D An embodiment of the method according to the invention is explained. In particular, the method described above with reference to Figuren 1A und 1B explained embodiment of the driver assistance system according to the invention as well as traffic situations such as those described above with reference to the Figuren 2A, 2B und 2C are explained.

[0083] The exemplary embodiment assumes at least partially automated driving of vehicle 1, with navigation unit 7 generating control signals based on which individual driving maneuvers are carried out automatically. The driver of vehicle 1 nevertheless retains responsibility for the journey and can fully reassume control at any time. Other exemplary embodiments may provide for other degrees of automation, such as fully automated driving along a predetermined route.

[0084] To perform an automated journey or at least individual automated driving maneuvers, the vehicle's surroundings must first be recorded. The recorded scenario is then interpreted, and a suitable strategy for the journey is determined. Scenario interpretation is responsible for recognizing, processing, and interpreting information relevant to the driving situation, especially traffic regulations.

[0085] For example, the passability of an intersection can be specified using three different types of traffic control devices: traffic signs, traffic lights, or no traffic signs, in which case general right-of-way rules apply, such as "right before left." Traffic control can also be specified by a person, such as a hand signal. A central problem in assessing a traffic situation at an intersection, particularly in an urban context, is therefore correctly interpreting the various detected traffic light phases and / or traffic signs in order to plan the driving behavior of vehicle 1 and, if necessary, a subsequent automatically executed driving maneuver.

[0086] In the exemplary embodiment of the method according to the invention, passability is assessed based on probability distributions for various conditions. A passability assessment is generated, which can then be used to automatically determine and execute appropriate driving behavior at an urban intersection.

[0087] In the exemplary embodiment, the detection unit 2 of the vehicle 1 captures environmental data about the surroundings of the vehicle 1. The environmental data here includes, for example, road data, map data, ego position data, and traffic control data. In the example, the environmental data is further preprocessed such that an environmental perception is provided based on the raw data, with recognition already performed for some relevant elements.

[0088] Traffic sign recognition is performed in step S1, and traffic light recognition is performed in step S4. In this way, traffic signs or traffic light phases, for example, are recognized. Furthermore, map data is acquired in step S2, and data on the ego position of vehicle 1 is acquired in step S3. Based on the data acquired in step S3, an estimate of the ego position of vehicle 1 on road 10 is generated. The map data acquired in step S2 can be considered in various ways, for example, to determine the number of lanes present at a specific position on road 10.

[0089] In further embodiments, the detection unit 2 can communicate directly with traffic infrastructure facilities, in particular by means of a Car-to-X Communication, through which, for example, a data connection to a central facility and / or a traffic signal system is established and data about traffic control is transmitted to vehicle 1. This means that in this case, traffic sign recognition in step S1 and traffic light recognition in step S4 can be simplified.

[0090] In the exemplary embodiment, elements of the traffic infrastructure in the vicinity of the vehicle, or rather for the road section for which a passability assessment is to be generated, are first detected based on the acquired environmental data. Alternatively or additionally, other method steps can be performed to acquire and evaluate information about the traffic situation.

[0091] The data acquired and possibly pre-processed in steps S1 to S4 have certain uncertainties, for example due to inaccuracies in the sensors used in the acquisition unit 2, inaccuracies in the acquired map data and / or in the steps for recognition, in particular of signs and traffic lights 13. In the Figur 3A In the diagram shown, this is indicated by probability values, which indicate an uncertainty in the traffic sign recognition P(TSR j ) in step S1, an uncertainty of the map material P(AM) in step S2, an uncertainty in determining the ego position P(Pos) in step S3 or an uncertainty in traffic light recognition P(TLR i ) in step S4. The index refers to i on different traffic lights 13 and the index jto different road markings 12 and traffic signs. The exemplary embodiment assumes that any traffic signs present are viewed in the same way as road markings 12.

[0092] These probabilities are taken into account to assign the vehicle and the detected traffic regulations to individual lanes in steps S5, S6, and S7. This means that it is determined which lane (10.1, 10.2, 10.3, 10.4, 10.5) the vehicle is located in and which traffic regulations, signaled by lane markings (12), traffic signs, or traffic lights (13), are relevant for which lane (10.1, 10.2, 10.3, 10.4, 10.5).

[0093] In a step S5, traffic signs are assigned to the lanes, whereby the results of the traffic sign recognition (step S1), the acquisition of the map data (step S2) and the acquisition of the ego position (step S3) are compared with the respective uncertainties P(TSR j ) , P(AM) and P(Pos) For the resulting assignment of traffic signs, an uncertainty is determined, which is determined by the probability P(TSA jf ) is represented, where the index f for the individual lanes 10.1, 10.2, 10.3, 10.4, 10.5 of road 10.

[0094] Similarly, in a step S7, the detected traffic lights 13 are assigned to the lanes 10.1, 10.2, 10.3, 10.4, 10.5, whereby the results of the traffic light detection (step S4), the acquisition of the map data (step S2) and the acquisition of the ego position (step S3) are compared with the respective uncertainties P(TLR i ) , P(AM) and P(Pos) be taken into account. This involves a probability P(TLA if ) generated.

[0095] In addition, in a step S6, the vehicle 1 is assigned to one of the lanes 10.1, 10.2, 10.3, 10.4, 10.5 of the road 10, wherein in the exemplary embodiment the map data acquired in step S2 are compared with the uncertainty P(AM) and the ego position determined in step S3 with the uncertainty P( Pos ) should be taken into account, whereby an uncertainty P(LA f ) is determined.

[0096] The exemplary embodiment then aims to determine which traffic regulations are relevant for vehicle 1.

[0097] For this purpose, in a step S8, an assignment of the detected lane markings 12 and / or traffic signs to the vehicle 1 is carried out, wherein the assignment of the position of the vehicle 1 to the lanes 10.1, 10.2, 10.3, 10.4, 10.5 determined in step S6 and the assignment of the traffic regulations to the lanes 10.1, 10.2, 10.3, 10.4, 10.5 determined in step S5 are compared with the respective probabilities P(LA f ) and P(TSA jf ) be taken into account.

[0098] In the embodiment of the method according to the invention, in this step S8 a confidence parameter γ j for the assignment of traffic signs, which is determined depending on the complexity of the control situation for the road section whose passability is to be assessed. In the exemplary embodiment, this is done by means of a machine learning method, whereby various known methods can be used. In particular, a Support Vector Machine (SVM, support vector method) can be used. A classifier is trained using suitable training data, for example, provided by the system manufacturer. Alternatively or additionally, another machine learning method can be used, such as a neural network or an artificial neural network (ANN), or another machine learning method.

[0099] This allows, based on the collected environmental data, for the trust parameter γ j to determine a value that lies, in particular, between 0 and 1 and indicates the probability with which the system, under the given circumstances, will reliably assign the traffic signs to lanes 10.1, 10.2, 10.3, 10.4, 10.5. In particular, the complexity of the existing regulatory situation is taken into account, particularly the number and arrangement of lanes 10.1, 10.2, 10.3, 10.4, 10.5 and / or the number and arrangement of traffic signs and road markings 12. Alternatively or additionally, various parameters can be used which describe the control situation independently of the functioning of the actual procedure for determining the passability assessment, for example the number of detected traffic lights 13, the number of lanes 10.1, 10.2, 10.3, 10.4, 10.5, the lateral distance of the detected traffic lights 13 to the center of the various lanes 10.1, 10.2, 10.3, 10.4, 10.5, distance of vehicle 1 to traffic lights 13 and so on.

[0100] The assignment of the traffic signs to the vehicle 1 determined in step S8 is with a probability P(TS l ) which is also determined and whose index / stands for different categories of traffic signs. This means that a probability distribution is determined that indicates the probability with which a traffic sign of a certain category was assigned to the vehicle. An example of such a probability distribution is shown in Figur 3B shown, with a specific probability for different types of traffic signs represented as a bar chart. The example considers the following conditions: "No traffic sign detected," "right before left," "priority road," "give way," "stop sign." Equivalent traffic regulations are included in each category.

[0101] This probability distribution P(TS l ) can therefore be expressed in the following way as a product of the conditional probabilities of the intermediate steps explained above:

[0102] This refers to: P ( TS l ) the probability that a traffic sign of category / must be observed for the vehicle; P ( TS l | LA f , TSA jf ) the conditional probability for a traffic sign of category / with the assignment of vehicle 1 to a lane f and an assignment of the traffic sign j to the lane f ; P ( LA f | Pos , AM ) the conditional probability for the assignment of vehicle 1 to a lane f with the probabilities for determining the ego position and for the captured map data; P ( TSA jf | AM, Pos, TSR j ) the conditional probability for the assignment of the traffic sign j to the lane f with the probabilities for the captured map data, the determination of the ego position and the recognition of the traffic sign j ; γ j the confidence parameter for the assignment of the traffic sign j to vehicle 1; P ( TSR j ) the probability of detecting the traffic sign j from step S1; P ( Pos ) the probability of determining the ego position from step S3; and P ( AM ) the probability of the captured map data from step S2.

[0103] Analogously, in a step S9, the detected traffic lights 13 are assigned to the vehicle 1, wherein the assignment of the position of the vehicle 1 to the lanes 10.1, 10.2, 10.3, 10.4, 10.5 determined in step S6 and the assignment of the traffic lights 13 to the lanes 10.1, 10.2, 10.3, 10.4, 10.5 determined in step S7 are compared with the respective probabilities P(LA f ) and P(TLA if ) be taken into account.

[0104] Instead of the trust parameter γ j For the assignment of traffic signs, a confidence parameter is used θ i for the assignment of traffic lights 13, which is determined in principle in the same way as described above. Here, too, an SVM can be used, which is trained on the basis of suitable data, in particular by the manufacturer of the system, and which is designed in particular to determine a value between 0 and 1 which indicates a probability that the system will achieve a reliable assignment of traffic lights 13 to lanes 10.1, 10.2, 10.3, 10.4, 10.5 under the given circumstances. As explained above, the complexity of the existing control situation is also taken into account here. Here, too, another machine learning method can be used alternatively or additionally, such as a neural network or an artificial neural network (ANN) or another machine learning method.

[0105] Analogous to step S8, for the assignment of the traffic lights 13 to the vehicle 1 determined in the step of the new, a probability assigned to the vehicle 1 P(TL k ) whose index k stands for different traffic light states. The traffic light states are determined based on the traffic light phases and indicate different traffic regulations, which cannot necessarily be assigned to a specific traffic light phase. Different traffic light phases can be assigned to the same traffic regulation, which classifies the passability of an intersection. For example, a green traffic light signal or a green traffic light with an arrow display can be provided, with both alternatives meaning that the vehicle is permitted to pass. In such a case, the specific traffic light phase is less relevant than the resulting traffic light state for deciding whether the vehicle may perform a certain driving maneuver.

[0106] Figur 3C shows examples of the assignment of different traffic light phases to different kdesignated traffic light states. The overview shown is for illustrative purposes only and makes no claim to completeness or correctness under traffic law. Traffic lights are shown, each with three illuminated surfaces, which can illuminate red (upper illuminated surface), yellow (middle) and / or green (bottom). The illuminated surfaces can also be round or arrow-shaped, with the arrow shapes pointing forward (or upwards), right, left, diagonally right or diagonally left. In the drawing, non-activated, i.e. dark, illuminated surfaces are indicated by unfilled outlines, while activated, i.e. illuminated, illuminated surfaces are shown as filled outlines. Of course, traffic light systems with just two or one illuminated surfaces can also be provided; the illuminated surfaces can be arranged in different ways relative to one another, the shapes of the illuminated surfaces can be designed in other ways, and so on.Furthermore, time-varying light signals can be considered as a separate traffic light phase and assigned to a traffic light state, in particular a flashing yellow light area.

[0107] At the Figur 3C shown overview corresponds k =1 a traffic light state in which the traffic light is deactivated, in particular switched off. k =2 shown traffic light phases have one red illuminated area of ​​the traffic light, while the other illuminated areas are deactivated, and are assigned to a traffic light state in which entry into the intersection is not permitted. k =3 assigned traffic light phases with a green illuminated area or simultaneously illuminated red and yellow areas are assigned to a traffic light state in which entry into the intersection is permitted. Traffic light phases in which only the yellow illuminated area is activated are shown in the overview. k=4 and thus a traffic light state in which entry into the intersection is permitted for a limited time. The traffic light state k =5 are assigned to traffic light phases in which the permission to turn right is signaled by a green arrow or by simultaneously illuminated red and yellow arrows. The traffic light state k =6 is analogous to k =5 and includes permission to turn left. Analogous to traffic light conditions k =5 and k =6 include the traffic light state k =7 and k =8 assigned traffic light phases each have a yellow arrow pointing right or left and signal the temporary permission to turn in the respective direction. The traffic light state k=9 in the overview is assigned to a traffic light phase in which a permanent green arrow pointing to the right is arranged at a red traffic light, which corresponds to a conditional permission to turn right.

[0108] Alternatively or additionally, various other traffic light states can be defined in other ways, such as "traffic light state not recognized", "traffic light off", "entry not permitted", "entry permitted", "compatible turning", "conditionally compatible turning", "compatible left / right turning at red light" (e.g.: traffic light with green arrow), whereby alternatively or additionally other traffic light states can be defined.

[0109] A probability distribution is generated which is analogous to that in Figur 3B distribution shown and the probabilities for the different k Such a probability distribution is shown in Figur 3D shown as an example, where along the X-axis the values ​​for k and along the Y-axis the associated probabilities P(TL k ) In the example shown here, a maximum for k =3, meaning that the detected traffic light phase indicates with the highest probability that entry into the intersection is permitted. This probability distribution P(TL k ) can therefore be expressed in the following way as a product of the conditional probabilities of the intermediate steps explained above:

[0110] This refers to: P ( TL k ) the probability that the traffic light state k for vehicle 1; P ( TL k | LA f , TLA if ) the conditional probability for the traffic light state k with the assignment of vehicle 1 to a lanef and an assignment of the traffic light i to the lane f ; P ( LA f | Pos , AM ) the conditional probability for the assignment of vehicle 1 to a lane f with the probabilities for determining the ego position and for the captured map data; P ( TLA if | AM , Pos, TLR i ) the conditional probability for the assignment of the traffic light i to the lane f with the probabilities for the captured map data, the determination of the ego position and the detection of the traffic light i ; θ i the confidence factor for the assignment of the traffic light i to vehicle 1; P ( TLR i ) the probability of detecting the traffic light i from step S4; P(Pos) the probability of determining the ego position from step S3; and P(AM) the probability of determining the captured map data from step S2.

[0111] In further embodiments, the generation of the trust factors γ j , θ i in other ways. In particular, trust factors γ j , θ i be fixed for different conditions, for example for certain categories of known traffic signs, for a certain number of lanes 10.1, 10.2, 10.3, 10.4, 10.5, traffic signs, road markings 12 and / or traffic lights 13 or for a certain ratio of the numbers of lanes 10.1, 10.2, 10.3, 10.4, 10.5 and traffic control devices to one another or for a certain road section.

[0112] In a further step S10, a passability assessment for the road section is generated based on the previously determined probabilities or wash uniformity distributions. In the exemplary embodiment, the passability assessment is output by transmitting a signal to the navigation unit 7, which determines a safe driving maneuver for the situation, in particular a target trajectory of the vehicle 1, and generates control signals such that the longitudinal and lateral guidance of the vehicle is controlled accordingly.

[0113] In the embodiment, it is provided that the passability assessment is a probability P(Pass m ) for certain conditions mthe passability is determined, i.e., a passability probability distribution is determined. A predetermined set of different passability states is considered, and their probability is calculated. For example, the navigation unit 7 can specify states that are relevant for route guidance and / or at least partially automatic travel along the specific route.

[0114] The determination of the passability rating can be done in various ways, for example by means of a Fuzzy-Logic -Procedure, a rule-based procedure or in another way. In further embodiments, additional data can also be recorded and taken into account, for example specifications for a particular road section.

[0115] In particular, a standardization is provided, that is, the sum of all m Probabilities P(Pass m ) is 1. In addition, the probability distributions for possible traffic light states k and the possible traffic signs l standardized.

[0116] The passability assessment is output according to the invention, with transmission to the navigation unit 7 being provided in the exemplary embodiment. The passability assessment is taken into account when executing driving maneuvers 3 of an at least partially automatic control system of the vehicle 1. For this purpose, an enable signal is generated based on the passability assessment, which contains information about whether the driving maneuver can be performed safely or not.

[0117] The method according to the invention can be carried out for traffic lights and / or traffic signs and, alternatively or additionally, in an analogous manner for various types of traffic regulations, for example also for road markings, traffic signs at construction sites or similar regulations.

[0118] With reference to the Figuren 4A, 4B und 4C Examples of conditional exponential averaging are explained that can be used in the above embodiments.

[0119] To ensure the correct interpretation of temporal changes within the recorded environmental data, the recorded probabilities are smoothed. For example, temporary fluctuations in the detection or assignment of a traffic sign, traffic light status, or lane may occur, for example, due to temporary obscuration by other road users or other effects.

[0120] The Figuren 4A bis 4C The solid line shown represents the probability P determined for a specific traffic light state as a function of time. In this simplified example, the probability is initially 0, then jumps to 1, briefly assumes the value 0 and then back to 1, and falls to 0 at the end of the shown range. For example, a red traffic light is not visible at the beginning of the time interval under consideration, is then detected, is subsequently briefly obscured by another vehicle, and is no longer registered at the end, for example after the traffic light has switched to a green signal.

[0121] The Figur 4A The dashed line shown shows the result P* of an exponential averaging ( Exponential Moving Average ; EMA), where a smoothing parameter ∝ takes a value between 0 and 1 and determines the degree of smoothing: P * A k , n = P A k , n ⋅ 1 − ∝ + P A k , n − 1 ⋅ ∝

[0122] This refers to P* ( A k , n) the smoothed probability P ( A k , n ) a specific traffic light signal A k for the times plotted along the abscissa n .

[0123] In the embodiment, it is provided that the value of ∝ is not chosen to be constant, but depends on certain conditions, for example for different traffic light states or traffic signs.

[0124] In particular, a conditional or conditional exponential averaging ( Conditional Exponential Moving Average ; cEMA), where the smoothing parameter ∝ is chosen depending on whether the value to be smoothed decreases (∝ ab ) or increases (∝ zu ). Depending on whether the smoothing parameter for the increasing value (∝ zu ) is greater or smaller than the smoothing parameter for the decreasing value (∝ ab ), a conservative exponential averaging (∝ zu <∝ ab ) or a non-conservative exponential averaging (∝ ab <∝ zu ) can be achieved.

[0125] In the Figur 4B In the case shown (see dashed line), a conservative exponential averaging is used, in which a decrease in the value is slowed down so that, for example, the short-term decrease to 0 becomes less noticeable. Figur 4C In the case shown (see dashed line), however, a non-conservative exponential averaging is carried out, whereby, conversely, an increase in the value is slowed down and, in the example, the short-term decrease to 0 is very clearly maintained.

[0126] The smoothing parameter ∝ can be calculated in various ways, depending on the traffic light state for which the probability is to be smoothed. For example, conservative smoothing can be selected for particularly critical signals, such as a detected prohibition to enter the intersection, such as a red light or a stop sign. In contrast, non-conservative smoothing can be selected for signals that are evaluated with caution in cases of doubt, such as a detected permission to enter the intersection, such as a green light or a "give way" sign. Bezugszeichenliste

[0127] 1Vehicle 2Detection unit 3Control unit 4Recognition unit 5Assignment unit 6Evaluation unit 7Navigation unit 10Road 10.1, 10.2, 10.3, 10.4, 10.5Other lanes 12Road markings 13Traffic light αSmoothing parameter γ j Confidence parameter (traffic sign) θ i Confidence parameter (traffic light) FDirection of travel S1Traffic sign recognition S2Acquisition of map data S3Acquisition of an ego position S4Traffic light recognition S5Assignment of traffic sign - lane S6Assignment of vehicle - lane S7Assignment of traffic light - lane S8Interpretation of relevant traffic sign S9Interpretations of relevant traffic light S10Generation of passability assessment

Claims

1. Method for operating a driver assistance system for a vehicle (1) on a road (10), in which surroundings data about the road (10) and about the surroundings of the vehicle (1) are recorded; wherein the recorded surroundings data comprise road data about the course of the portion of the road (18) that is located in front of the vehicle (1) in the direction of travel (F), about one or more lanes (10.1, 10.2, 10.3, 10.4, 10.5) of the road (10) and about traffic control devices (12, 13); at least one traffic control for a road portion located in front of the vehicle (1) in the direction of travel (F) is detected and assigned to the vehicle (1) on the basis of the recorded surroundings data; and a passability assessment is generated and output for the road portion on the basis of the traffic control assigned to the vehicle (1); characterized in that the passability assessment is further generated on the basis of a confidence parameter (γj, θi) which is determined on the basis of the recorded surroundings data and comprises a measure of the complexity of a control situation for the road portion; wherein the confidence parameter is determined without the concrete determination of an assignment of the traffic control to the vehicle; wherein the confidence parameter (γj, θi) decreases with an increasing number of possible combinations in which the traffic control devices (12, 13) can be assigned to the lanes (10.1, 10.2, 10.3, 10.4, 10.5) of the road (10).

2. Method according to any of the preceding claims, characterized in that furthermore, map data about the road (10) in the surroundings of the vehicle (1) are recorded; wherein the map data comprise information about a number of lanes (10.1, 10.2, 10.3, 10.4, 10.5) of the road (10), the position of intersections and / or the position of traffic control devices (12, 13).

3. Method according to either of the preceding claims, characterized in that the recorded surroundings data comprise ego position probability data; and an ego position probability distribution for the ego position of the vehicle (1) is determined on the basis of the ego-position data.

4. Method according to claim 3, characterized in that the road (10) has at least two lanes (10.1, 10.2, 10.3, 10.4, 10.5) and the ego position probability distribution is determined relative to the lanes (10.1, 10.2, 10.3, 10.4, 10.5) of the road (10).

5. Method according to claim 3 or claim 4, characterized in that the assignment of the detected traffic control to the vehicle (1) is carried out on the basis of the ego position probability distribution.

6. Method according to any of the preceding claims, characterized in that the recorded surroundings data comprise traffic control data; and in order to detect the traffic control, a traffic control probability distribution is determined for a plurality of predefined traffic control categories on the basis of the traffic control data.

7. Method according to any of the preceding claims, characterized in that an assignment probability distribution for an assignment of traffic control devices (12, 13) to the vehicle (1) is determined on the basis of the recorded surroundings data.

8. Method according to any of the preceding claims, characterized in that the passability assessment comprises a passability probability distribution.

9. Method according to any of claims 3 to 8, characterized in that temporal smoothing is carried out using a smoothing parameter (α) for at least one of the determined probability distributions, wherein preferably exponential moving averaging, particularly preferably conditional moving exponential averaging, is carried out.

10. Method according to claim 9, characterized in that a crossing state is determined on the basis of the recorded surroundings data; and the smoothing parameter (α) is determined on the basis of the crossing state.

11. Method according to any of the preceding claims, characterized in that the confidence parameter (γj, θi) is determined by means of a machine learning method.

12. Method according to any of the preceding claims, characterized in that the passability assessment is generated for a predetermined set of driving maneuvers.

13. Driver assistance system for a vehicle (1) on a road (10), comprising a detection unit (2), by means of which surroundings data about the road (10) and about the surroundings of the vehicle (1) can be detected; wherein the recorded surroundings data comprise road data about the course of the portion of the road 18) that is located in front of the vehicle (1) in the direction of travel (F), about one or more lanes (10.1, 10.2, 10.3, 10.4, 10.5) of the road (10) and about traffic control devices (12, 13); a recognition unit (4), by means of which at least one traffic control for a road portion (11) located in front of the vehicle (1) in the direction of travel can be detected on the basis of the recorded surroundings data; an assignment unit (5), by means of which the detected traffic control can be assigned to the vehicle; and an evaluation unit (6), by means of which a passability assessment can be generated and output for the road portion on the basis of the traffic control assigned to the vehicle; characterized in that the passability assessment can further be generated on the basis of a confidence parameter (γj ,θi) which can be determined on the basis of the recorded surroundings data and comprises a measure of the complexity of a control situation for the road portion; wherein the confidence parameter can be determined without the concrete determination of an assignment of the traffic control to the vehicle; wherein the confidence parameter (γj, θi) decreases with an increasing number of possible combinations in which the traffic control devices (12, 13) can be assigned to the lanes (10.1, 10.2, 10.3, 10.4, 10.5) of the road (10).

14. Driver assistance system according to claim 13, characterized in that the passability assessment can be received by a control unit; wherein a control signal is generated on the basis of the passability assessment, on the basis of which at least partially automated travel of the vehicle (1) in the region of the road portion can be carried out.