Driver assistance system for lane detection with backup function in case of incomplete lane detection or sporadic disruptions in lane detection, as well as method for generating a GPS-based reference route as a temporary backup in case of sporadic disruptions in optical lane detection

Abstract

Driver assistance system for a vehicle (1) with a lane recognition function acting in the direction of travel, in particular for detecting lane boundaries relating to the lane travelled by the vehicle (1), comprising at least a) an environment detection system (3) acting in the direction of travel, and b) a GPS-based positioning system for determining a route, and / or a navigation system, and c) a storage medium for storing a GPS-based target lane course (TARGET), and d) an evaluation unit for carrying out the required test steps (Px), whereby e) the lane course according to the left and right lane boundaries (FSB) is / is defined as the actual lane course (IST), or f) the lane course according to the left or right lane boundary (FSB) is / is defined as the actual lane course (IST), characterized in that g) in the event of a temporary fault and / or sporadic interruptions of the environmental detection system (3), with regard to h) lane detection (FBA, FBB, FBC), and / or i) the detection of the lane boundaries (FSB) associated with the lanes (FBA, FBB, FBC), j) if possible, an alternative lane prediction is used, whereby k) during the prediction, a GPS-based target lane course (TARGET) is reconstructed from the data stored in a storage medium, whereby 1) the data of the GPS-based target lane course (TARGET) are trained route data (AS), where m) the trained route data (AS) must undergo a qualification cycle so that the GPS-based target lane course (TARGET) stored as a reference route achieves a temporary positive confidence level status as a trained route, whereby n) to achieve a temporary positive confidence level status, it is necessary that a qualification counter has a certain number x of individual increments (n), where o) the specific number x of the individual required increments (n) of the qualification counter are each not older than a specific predeterminable time (x).

Description

[0001] The present invention relates to a driver assistance system for lane detection with a backup function in the event of incomplete lane detection or sporadic disruptions in lane detection, as well as a method for an optical lane detection driver assistance system for generating a GPS-based reference route as a temporal backup in the event of incomplete lane detection or sporadic disruptions in lane detection, as well as an assistance system. Generally:

[0002] Reliable lane detection is essential for autonomous driving. Reliable lane detection is also required for driver assistance systems, such as a Lane Departure Warning (LDW) system.

[0003] Front cameras (mono or stereo) are used for this purpose.

[0004] If lane detection temporarily fails or is not possible, the aforementioned systems very quickly reach their system limits, so that a) the driver (in manual driving mode without monitoring function or without LDW) is left to his own devices, or b) the driver (in autonomous driving mode) must take over control of the vehicle relatively quickly as a fallback level.

[0005] Such disruptions or outages are easily possible, for example, in winter when snow or slush covers the road markings or lane lines (see Fig. 1).

[0006] An electric horizon as a backup also offers only a limited solution, since these solutions have only limited accuracy and are ultimately too inaccurate for autonomous driving.

[0007] From the document DE 10 2010 049 215 A1 a method for determining a vehicle environment is known in which current local environmental data are determined by means of an environmental sensor, wherein objects in the environment of the vehicle are determined from the current local environmental data and a current local environmental map of the vehicle is generated from a sequence of the current local environmental data, wherein a lane course and a position of the objects with respect to the lane course are determined based on a comparison of the current local environmental map with map data read from a digital road map.

[0008] From the document DE 10 2006 000 640 A1 a method for controlling a driver assistance system is known, comprising at least one environmental sensor for the acquisition of environmental data, wherein the data acquired by the environmental sensor are stored, wherein in the event of a disturbance in the acquisition of the environmental data the stored data is used in order to derive estimated values ​​for currently missing environmental data. Object of the invention:

[0009] The object of the invention can therefore be seen in achieving an improvement over the prior art in such a way that Temporary disruptions or sporadic failures in lane detection can still be safely bridged by the driver assistance system (at least for a short time). Solution to the task:

[0010] The task is solved in such a way that, if lane detection based on a left and a right lane boundary is not possible, lane detection in autonomous driving mode and / or lane prediction in LDW (Lane Departure Warning) is carried out alternatively by using trained GPS-based target reference routes where possible. In other words:

[0011] The task is solved by training a GPS-based target reference route for each route travelled multiple times, whereby this must have a certain confidence level so that the trained GPS-based target reference route is sufficiently resilient to a) to be able to reliably maintain a monitoring function (LDW) to support the driver in manual driving mode, and / or b) to be able to continue driving the vehicle safely as a fallback level in autonomous driving mode. Short explanation:

[0012] Trained route data means that the route has already been driven several times by the corresponding vehicle, and a high level of agreement was achieved when documenting the driving trajectory using GPS support during repeated journeys.

[0013] If there is a high degree of agreement (e.g., a route is driven five times, resulting in an identical driving trajectory to the recorded one), then there is a high confidence level, which ultimately enables the vehicle to be controlled autonomously along this trained driving trajectory, at least for a limited time, or this trained driving trajectory can be used as a monitoring reference for the LDW function.

[0014] Traffic sign recognition is understood in particular to mean whether, for example, a construction site sign indicates a change in traffic routing or whether it indicates a change in traffic routing with changed markings.

[0015] The invention is based on the fact that most journeys during driving (in the vehicle's life) are repetitive journeys, such as the daily journey to work and back, whereby these journeys are driven relatively often (e.g. ≥ 5) within a short period of time (e.g. 1 month).

[0016] The driver assistance system according to the invention for a vehicle with a lane recognition function acting in the direction of travel, in particular for the recognition of lane boundaries relating to the lane travelled by the vehicle, consists of at least (a) an environment detection system acting in the direction of travel, and b) a GPS-based positioning system for determining a route, and / or a navigation system, and c) a storage medium for storing a GPS-based target lane path, and d) an evaluation unit for carrying out the necessary test steps, whereby e) the lane course according to the left and right lane boundaries (FSB) is / is defined as the actual lane course (IST), or f) the lane course is / is defined as the actual lane course (IST) according to the left or right lane boundary (FSB), where g) in the event of a temporary fault and / or sporadic interruptions of the environmental detection system, with regard to h) lane detection, and / or i) the detection of the lane boundaries associated with the lanes, j) if possible, an alternative lane prediction is used, whereby k) during the prediction, a GPS-based target lane course is reconstructed from the data stored in a storage medium, whereby l) the data of the GPS-based target lane course are trained route data, whereby m) the trained route data (AS) must undergo a qualification cycle so that the GPS-based target lane course (TARGET) stored as a reference route achieves a temporary positive confidence level status as a trained route, whereby n) to achieve a temporary positive confidence level status, it is necessary that a qualification counter has a certain number x of individual increments (n), where o) the specific number x of the individual required increments (n) of the qualification counter are each not older than a specific predeterminable time (x).

[0017] An advantageous embodiment of the invention provides (as already mentioned above) that the trained route data must undergo a qualification cycle so that the GPS-based target lane path stored as a reference path achieves a temporary positive confidence level status as a trained route. The time limit, which can be, for example, 1 month or other values, is due to the fact that the trained and stored route data is always up-to-date and that no outdated route data can be used.

[0018] A further advantageous embodiment of the invention provides that the value for the specific number x of the individual increments (n) of the qualification counter is 5, and the specific predeterminable time (x) is set to 1 month, wherein the number x of the individual increments (n) of the qualification counter and / or the specific predeterminable time (x) can also assume other values.

[0019] A further advantageous embodiment of the invention provides that the driver assistance system is a system for autonomous vehicle guidance and / or a lane monitoring system for warning the driver if the lane (FBA, FBB, FBC) is not maintained by the driver within the specified tolerance range.

[0020] Furthermore, the invention comprises a method for an optical driver assistance system, wherein a further advantageous embodiment of the invention provides that, in order to generate trained route data of a GPS-based target lane course as a reference course, the determined GPS-based actual lane course (IST) is corrected by the deviation and / or delta in order to obtain a GPS-based target lane course as a trained route after multiple matching and to store it accordingly in a storage medium, whereby the trained route data (AS) must undergo a qualification cycle so that the GPS-based target lane course (TARGET) stored as a reference route achieves a temporary positive confidence level status as a trained route, where to achieve a temporary positive confidence level status it is necessary that a qualification counter has a certain number x of individual increments (n), where the specific number x of the individual required increments (n) of the qualification counter are each not older than a specific predeterminable time (x).

[0021] A further advantageous embodiment of the method according to the invention provides that, based on the determined GPS-based actual position lane profile data and the knowledge of the deviation which is determined via the environment detection system, desired position lane profile data are reconstructed at a later time by guiding the first vehicle along the GPS-based desired position lane profile data in such a way that the earlier GPS-based actual position lane profile data are corrected with the knowledge of the deviation, wherein the correction can already be carried out when the determined position lane profile data are stored, or only when the determined position lane profile data are read out from the storage medium.

[0022] A further advantageous embodiment of the method according to the invention provides that system-related "offset errors" are automatically compensated during position detection, since offset-related errors of the positioning system during data determination are automatically compensated when the original lane course is reconstructed / generated using these offset-affected data and the positioning system with offset errors, which is to be corrected using the deviation originally determined by the environment detection system (a measurement that is measured with an "incorrect reference" results in the original measurement during reconstruction if the reconstruction is carried out with the same "incorrect reference").

[0023] A further advantageous embodiment of the method according to the invention provides that the lane guidance during autonomous driving and / or warning message during manual driving is carried out on the basis of a lane definition, wherein in determining the lane definition and / or evaluating the quality of the lane recognition, in addition to the a) where available, GPS-based target lane path reconstructed from the storage medium and / or trained route data as parameters, b) further parameters are taken into account, whereby at least one of these parameters is given a weighting factor in order to allow this parameter a more dominant influence.

[0024] Further features, effects, and advantages of the invention will become apparent from the description of preferred embodiments of the invention, with a combination of the features of the individual figures also being included within the scope of protection. All figures are only schematic representations (not to scale). They show: Fig. 1 schematically shows an example in which the vehicle 1, or the environment detection system located in the vehicle 1, cannot correctly detect a lane boundary. Fig. Figure 2 shows a schematic flow chart which is used to check how to proceed in the event of an irregularity according to Fig. 1 procedure is followed. Fig. 3 schematically shows a vehicle driving on a roadway with an actual lane course, whereby this actual lane course has a deviation Δ compared to the desired lane course. Fig. 4 schematically shows a flow chart by means of which, as proposed according to the invention, a reliable target lane course can be generated as a reference course for a certain route.

[0025] The Fig. 1 schematically shows a practical example (state of the art), in which a first vehicle (1) is driving behind a second vehicle (2) on the middle lane B (FBB) of a three-lane roadway (FBA, FBB, FBC), or rather follows the second vehicle (2). As can also be seen from the figure, the first vehicle (1) has an environmental detection system (3) acting in the direction of travel, by means of which the roadway course (FBB) ahead / the driving trajectory (FBB) can be detected. As can also be seen from the figure, the three roadways (FBA, FBB, FBC) or lanes (FBA, FBB, FBC) of the three-lane roadway are separated from one another by lane markings (FSB), and are also marked off on the far left and far right by a lane marking (FSB) each.As can be further seen (schematically) from the figure, there is a "slush" (SM) in the area of ​​the lane boundary (FSB) between the roadway A (FBA) and the roadway B (FBB), which results in the environment detection system (3) located in the first vehicle (1) not being able to correctly detect the lane boundary (FSB) at this point or between the roadway A (FBA) and the roadway B.

[0026] The Fig. Figure 2 shows a schematic flow chart which is used to check how to proceed in the event of an irregularity according to Fig. 1 is carried out, or how a decision is made if a lane boundary cannot be detected in sufficient quality by the environment detection system (3) located in the vehicle (1).

[0027] Regardless of whether a vehicle (1) is operated autonomously or manually and travels along a roadway (FBA, FBB, FBC), the lane course (FBA, FBB, FBC) is detected by the vehicle's (1) environmental detection system (3). In a first test step (P1), it is checked whether a right AND a left lane boundary (FSB) are detected. If both a right AND a left lane boundary (FSB) are detected, the lane course according to the left and right lane boundaries (FSB) is defined as the actual lane course (IST), and the value S is set to 1. The value S is a parameter for evaluating the quality of lane detection.

[0028] If it is not possible to detect a right AND a left lane boundary (FSB), the value S is set to 0 and the next test step (P2) checks whether a right OR a left lane boundary (FSB) is detected by the vehicle's (1) environmental detection system (3). If either a right OR a left lane boundary (FSB) is detected, the lane course according to the left or right lane boundary (FSB) is defined as the actual lane course (IST) and the value S is set to 0.5, and the next test step (P3) is proceeded to. If it is not possible to detect a right OR a left lane boundary, the value S is set to 0 and the next test step (P7) is proceeded to.

[0029] In the next test step (P7) it is checked whether (already) trained route data (AS) are available in the storage medium of the assistance system according to the invention for the currently travelled route (the currently travelled route is determined by means of a navigation system or with a position-determining system according to the state of the art) (this is explained in the description of the Fig. 4 discussed in more detail).

[0030] If trained route data (AS) are available, the value AS is set to 1 and in the next test step (P8) it is checked whether the trained route data (AS) have a high confidence level (e.g. n ≥ x - see Fig. 4), whereas if no trained route data (AS) are available, the value AS is set to 0 and the value G1 is set to 1 and the test proceeds directly to the next test step (P9).

[0031] If trained route data (AS) are available and the test step (P8) results in a high confidence level, the value G1 is set to 1 and the process continues with the next test step (P10), whereas if the test step (P8) does not result in a high confidence level, the value G1 is set to 1 and the process continues directly with the next test step (P9).

[0032] In the next test step (P10), the system checks whether traffic sign recognition (V) is possible. If the result is negative, the weighting factor G1 is increased by multiplying the weighting factor G1 by the factor x > 2 (x is also increased from the original 1) to G1' (G1' = x * G1). Following this, as with a positive result of test step (P10), the system proceeds directly to the next test step (P9).

[0033] In the next test step (P9) it is checked whether orientation to a vehicle in front (FS) is possible.

[0034] If this result is positive, the value FS is set to 1, and the weighting factor G2 is increased by multiplying the weighting factor G2 by the factor x > 2 (x is also increased compared to the original 1) to G2' (G2' = x * G1), in order to subsequently calculate S for the track definition. If, however, the result of the test step (P9) is negative, the value FS is set equal to the value S, and the weighting factor G2' is set to 1, in order to subsequently calculate S for the track definition.

[0035] The track definition (S) is calculated using the following mathematical formula: S=(FS*G2'*50%+AS*G1'*50%) / (FS*G1*50%+AS*G1*50%)

[0036] Following this calculation, the following procedure is followed depending on the result: For S ≥ 1: The lane guidance (in autonomous driving mode) or warning message (in manual driving mode) is carried out according to the lane or lane definition according to the GPS-based target lane course (TARGET) (reconstructed from the storage medium); If S < 1: A warning message is issued or, if autonomous driving was in progress, autonomous driving is terminated.

[0037] Then you return to the beginning to repeat the process steps.

[0038] The Fig. 3 shows analogous to Fig. 1 schematically shows a first vehicle (1) which is driving on a three-lane roadway (FBA, FBB, FBC), on the middle lane B (FBB), behind a second vehicle (2), or following the second vehicle (2). As can be seen from the figure, the second vehicle (2) is driving centrally (b / 2 to b / 2) to lane B (FBB) / in lane B (FBB). This lane course is referred to in the light of the invention as the desired lane course (SOLL). As can be seen from the figure, the first vehicle (1) is driving off-center to lane B / in lane B. This lane course is referred to in the light of the invention as the actual lane course (IST). As can be seen from the figure, there is a difference or a distance (A) between the actual road course (ACTUAL) and the desired road course (DESIRED), which in the light of the invention is referred to as delta (Δ) or distance (A). In other words:

[0039] The first vehicle (1) travels along a roadway (FBB) / lane (FBB) with an actual lane path (ACTUAL), wherein this actual lane path (ACTUAL) exhibits a deviation (A) or a delta (Δ) compared to the desired lane path (DEFINITE). As is not clearly visible in the figure, the first vehicle (1) has a GPS system (or a system with the same function but a different name) to record the lane path using position data and to use this data to store the lane path in a storage medium. The term "GPS" also stands for systems with the same function but a different name, so that any type of satellite-based positioning system is to be regarded as equivalent and is included in the scope of protection, or the scope of protection also includes such systems.

[0040] According to the invention, the determined GPS-based actual lane path (ACTUAL) is corrected by the deviation (A) or delta (Δ) in order to obtain (after repeated matching) a GPS-based desired lane path (TARGET) as a trained route and store it accordingly in a storage medium. This allows driver-dependent driving deviations to be eliminated. In other words:

[0041] Based on the determined actual position lane progression data and the knowledge of the deviation (A) which is determined via the environment detection system (3), a desired position lane progression data can be reconstructed at a later time by guiding the first vehicle (1) along the desired position lane progression data in such a way that the earlier actual position lane progression data is corrected with the knowledge of the deviation (A), whereby the correction can already be carried out when the determined position lane progression data is saved (A), or only when the determined position lane progression data is read out (the decisive factor is that the deviation (A) is taken into account once).) A further advantage of this concept is that system-related "offset errors" during position detection are automatically compensated for, since offset-related errors of the positioning system during data determination are automatically compensated for when the original lane course is reconstructed / generated using this offset-affected data and the positioning system with the offset error, which is to be corrected using the deviation (A) that was originally determined by the environment detection system (3). Since an offset-related error of the positioning system is automatically compensated for in this system, it is only mentioned in the invention, since the invention primarily concerns the determination of a desired lane course (TARGET), which is determined taking a deviation (A) into account, orif necessary (in case of irregularities in the environment detection system) it is reconstructed for driving purposes in order to be able to orientate oneself.

[0042] By means of this stored information, which describes the GPS-based target lane course (TARGET), it is possible that at a later point in time the route can be reconstructed from the storage medium using this stored information of the GPS-based target lane course (TARGET) (taking into account the deviation (A)), so that in the event of a sporadic malfunction of the environment detection system (3) or in the event of problems in the detection of the lane boundaries (FSB) by the environment detection system (3), the target lane course (TARGET) can be reconstructed as a "back-up" / as a replacement from the storage medium.

[0043] The Fig.4 schematically shows a flow chart by means of which, as proposed according to the invention, a reliable GPS-based target lane course (TARGET) can be generated as a reference course for a certain route.

[0044] Regardless of whether a vehicle (1) is operated autonomously or manually and travels along a roadway (FBA, FBB, FBC), the lane course (FBA, FBB, FBC) is detected by the vehicle's (1) environment detection system (3). A first test step (P1) checks whether a right AND a left lane boundary (FSB) are detected. If both a right AND a left lane boundary (FSB) are detected, the lane course is defined as the actual lane course (ACTUAL) according to the left and right lane boundaries (FSB).

[0045] If it is not possible to detect a right AND a left lane boundary (FSB), the next test step (P2) checks whether a right OR a left lane boundary (FSB) is detected by the vehicle's (1) environment detection system (3). If either a right OR a left lane boundary (FSB) is detected, the lane path according to the left or right lane boundary (FSB) is defined as the actual lane path (ACTUAL). If it is not possible to detect a right OR a left lane boundary, the system returns to the beginning to repeat the process steps.

[0046] If, however, an ACTUAL lane path (ACTUAL) could be defined from one of the two previous checks, the next test step (P3) uses the image information from the environment detection system (3) to determine whether there is a deviation (A) between an ACTUAL lane path (ACTUAL) currently being traveled by the first vehicle (1) and a TARGET lane path (TARGET) that the first vehicle (1) should currently be traveling in. If a deviation (A) between the ACTUAL lane path (ACTUAL) and the TARGET lane path (TARGET) is detected, the GPS-based ACTUAL lane path determined by the GPS data position determining system is corrected with the size / value of the deviation (A) to form a GPS-based TARGET lane path (TARGET). This correction can be used to correct, for example, driving uncertainties on the part of the driver and / or overtaking maneuvers initiated by the driver. In order to follow up on this correction orIf no deviation between the actual lane course (ACTUAL) and the desired lane course (TARGET) could be detected (did not exist), this (corrected) GPS-based desired lane course (TARGET) is saved as a reference course in a storage medium.

[0047] In the next test step (P4), it is checked whether a reference course as a GPS-based target lane course (TARGET) is already available or has been saved in the memory for this route / lane course due to a previous travel of this route.

[0048] If this check is positive, the next test step (P5) checks whether the new (currently) determined GPS-based target lane path (TARGET) is identical to the previously stored GPS-based target lane path (TARGET) serving as a reference. If this check is positive, a qualification counter (n) is incremented by 1 (n = n + 1). If this check is negative, the qualification counter (n) is reset to zero (n = 0).

[0049] Independent a) whether the qualification counter (n) has been increased or reset, or b) whether it was determined on the basis of a previous check step that no reference course is present or stored in the memory as a GPS-based target lane course (TARGET), and on the basis of which the qualification counter was set to 1 (n = 1), the next check step (P6) checks whether the qualification counter (n) has a value ≥ x, or whether this value is composed of individual increments (n) that are not older than a certain predefined time (x), where, for example, 1 month is set as the time and, for example, 5 is set as the predefined value for x, enabling reliable qualification, although values ​​for x and / or (x) that deviate from this are also possible.If this check results in the qualification counter (n) having a value ≥ x and none of the increments (n) being older than the predeterminable time (x), the GPS-based target lane course (TARGET) stored as a reference course acquires a temporary positive confidence level status as a trained route, which can be used as a backup / replacement if necessary if the environment detection system (3) located in the first vehicle (1) cannot correctly detect the lane boundary (FSB), in order to continue to provide an assistance function of the assistance systems equipped with the method according to the invention / to be able to provisionally maintain the assistance function of these assistance systems. The value x for the required increments (n) can be, for example, three, or five, or another numerical value.However, if the result of the test step (P6) is that the qualification counter (n) does not have the required value ≥ x, or if this value is composed of individual required increments (n) that are partly older than a certain predefined time (x), the time-limited positive confidence level status is reset.

[0050] Following the withdrawal of the temporary positive confidence level status, or following the acquisition of the stored GPS-based target lane course (TARGET) as a reference course with a temporary positive confidence level status, the system returns to the beginning in order to carry out the procedural steps / test steps again. List of reference symbols: 1 First vehicle 2 Second vehicle 3 Environmental detection system, e.g. camera A Deviation (Δ between TARGET & ACTUAL) AS trained route data b / 2 Half lane width FSB lane restriction FBA Roadway A / Lane A FBB Roadway B / Lane B FBC Lane C / Lane C FS orientation to the vehicle ahead G1 weighting factor 1 G1' corrected weighting factor 1 G2 weighting factor 2 G2' corrected weighting factor 2 SM Snow-Mud TARGET TARGET lane path / GPS-based TARGET lane path ACTUAL ACTUAL lane layout S value / size for evaluating the quality of lane detection n Increment for the qualification counter Px test step V Traffic sign recognition

Claims

[1] Driver assistance system for a vehicle (1) with a lane recognition function acting in the direction of travel, in particular for detecting lane boundaries relating to the lane travelled by the vehicle (1), comprising at least a) an environment detection system (3) acting in the direction of travel, and b) a GPS-based positioning system for determining a route, and / or a navigation system, and c) a storage medium for storing a GPS-based target lane course (TARGET), and d) an evaluation unit for carrying out the required test steps (Px), whereby e) the lane course according to the left and right lane boundaries (FSB) is / is defined as the actual lane course (IST), or f) the lane course is / is defined as the actual lane course (IST) according to the left or right lane boundary (FSB), characterized by , that g) in the event of a temporary fault and / or sporadic interruptions of the environmental detection system (3), with regard to h) lane detection (FBA, FBB, FBC), and / or i) the detection of the lane boundaries (FSB) associated with the lanes (FBA, FBB, FBC), j) if possible, an alternative lane prediction is used, whereby k) during the prediction, a GPS-based target lane course (TARGET) is reconstructed from the data stored in a storage medium, whereby 1) the data of the GPS-based target lane course (TARGET) are trained route data (AS), where m) the trained route data (AS) must undergo a qualification cycle so that the GPS-based target lane course (TARGET) stored as a reference route achieves a temporary positive confidence level status as a trained route, whereby n) to achieve a temporary positive confidence level status, it is necessary that a qualification counter has a certain number x of individual increments (n), where o) the specific number x of the individual required increments (n) of the qualification counter are each not older than a specific predeterminable time (x). [2] Driver assistance system according to claim 1, characterized by that the value for the specific number x of the individual increments (n) of the qualification counter is 5, and the specific predeterminable time (x) is set to 1 month, whereby the number x of the individual increments (n) of the qualification counter and / or the specific predeterminable time (x) can also assume other values. [3] Driver assistance system according to one of claims 1 to 2, characterized bythat the driver assistance system is a system for autonomous vehicle guidance and / or it is a lane monitoring system to warn the driver if the lane (FBA, FBB, FBC) is not maintained by the driver within the specified tolerance range. [4] Method for an optical driver assistance system according to one of claims 1 to 3, characterized bythat in order to generate trained route data (AS) of a GPS-based target lane course (TARGET) as a reference course, the determined GPS-based actual lane course (IST) is corrected by the deviation (A) and / or delta (Δ) in order to obtain a GPS-based target lane course (TARGET) as a trained route after multiple matching and to store it accordingly in a storage medium, whereby the trained route data (AS) must undergo a qualification cycle so that the GPS-based target lane course (TARGET) stored as a reference course achieves a time-limited positive confidence level status as a trained route, where to achieve a temporary positive confidence level status it is necessary that a qualification counter has a certain number x of individual increments (n), where the specific number x of the individual required increments (n) of the qualification counter are each not older than a specific predeterminable time (x). [5] Method for an optical driver assistance system according to claim 4, characterized by that, based on the determined GPS-based actual position lane profile data (IST) and the knowledge of the deviation (A) which is determined via the environment detection system (3), desired position lane profile data are reconstructed at a later point in time by guiding the first vehicle (1) along the GPS-based desired position lane profile data in such a way that the earlier GPS-based actual position lane profile data are corrected with the knowledge of the deviation (A), wherein the correction can already be carried out when the determined position lane profile data are stored (A), or only when the determined position lane profile data are read out from the storage medium. [6] Method for an optical driver assistance system according to one of claims 4 to 5, characterized by that system-related "offset errors" are automatically compensated for during position detection, since offset-related errors of the positioning system during data determination are automatically compensated for when the original lane course is reconstructed / generated using these offset-affected data and the positioning system with the offset error, which is to be corrected using the deviation (A) that was originally determined by the environment detection system (3). [7] Method for an optical driver assistance system according to one of claims 4 to 6, characterized by that the lane guidance during autonomous driving and / or warning message during manual driving is based on a lane definition (S), whereby in determining the lane definition (S) and / or evaluating the quality of the lane recognition (S), in addition to the a) if available, GPS-based target lane course (TARGET) reconstructed from the storage medium and / or trained route data (AS) as parameters (AS), b) further parameters (S, FS, V) are taken into account, whereby at least one of these parameters (AS, S, FS, V) is provided with a weighting factor (G1, G1', G2, G2') in order to allow this parameter (AS, S, FS, V) a more dominant influence.

Images