Method and assistance system for adjusting a warning frequency of warning notifications for a driver of a vehicle

Abstract

The present invention relates to a method for adjusting a warning frequency of warning notifications (104) for a driver of a vehicle, wherein the warning frequency is increased when, without a preceding warning notification (104), a corrective maneuver (108) by the driver is detected, and the warning frequency is reduced when, despite a preceding warning notification (104), no corrective maneuver (108) is detected.

Description

[0001] R. 415141

[0002] Description

[0003] title

[0004] Method and assistance system for setting a warning frequency of warning messages for a driver of a vehicle

[0005] Field of invention

[0006] The invention relates to a method for setting a warning frequency of warning messages for a driver of a vehicle, a corresponding assistance system and a corresponding computer program product.

[0007] State of the art

[0008] A vehicle may have an assistance system that warns the driver, for example, if they approach an upcoming curve at an inappropriate speed.

[0009] DE 10 2014 225 625 A1 describes a method for assisting a driver of a single-track motor vehicle to safely navigate a curve.

[0010] Disclosure of the invention

[0011] Against this background, the approach presented here introduces a method for setting the warning frequency of warning messages for a vehicle driver, a corresponding assistance system, and a corresponding computer program product according to the independent claims. Advantageous further developments and improvements of the approach presented here result from the description and are described in the dependent claims.

[0012] Advantages of the invention

[0013] KU:GR R. 415141

[0014] 2 -

[0015] A vehicle, particularly a single-track vehicle, may be equipped with a curve warning system. This system issues a warning to the driver if the vehicle is approaching a curve at a speed deemed too high by the system. The warning is intended to prompt the driver to reduce speed before the curve to ensure safe passage.

[0016] It can happen that a warning is issued and the driver does not react to it, because due to his driving skills he can safely drive through the curve without corrective intervention.

[0017] Conversely, it can happen that no warning is provided and the driver also has to make a late corrective intervention due to his driving skills, because he was not warned and underestimated the curve.

[0018] Both situations are undesirable and reduce the acceptance of the curve warning system. In extreme cases, the driver will deactivate the system.

[0019] The approach presented here monitors whether the driver fails to respond to warnings. It also checks whether the driver has to perform a corrective maneuver despite not having received a warning. In both cases, a warning threshold, at which a warning will be issued in the future, is adjusted. If there is no response to the warning, the warning threshold is raised, resulting in fewer warnings. If a corrective maneuver is performed despite not having received a warning, the warning threshold is lowered, resulting in more frequent warnings.

[0020] The approach presented here allows the warning frequency to be adjusted to the driver's skill level and driving style. The driver's daily condition can also be taken into account, as the warning frequency is always adjusted slightly when one of these factors is present. Adapting the warning frequency to the driving style and skill level can increase acceptance and thus usage. Increased usage can, in turn, improve overall safety. R. 415141

[0021] 3 -

[0022] A method for setting the warning frequency of warnings for a driver of a vehicle is presented, wherein the warning frequency is increased when a corrective maneuver by the driver is detected without a preceding warning, and the warning frequency is reduced when no corrective maneuver is detected despite a preceding warning.

[0023] Ideas for embodiments of the present invention can be considered to be based, among other things, on the thoughts and findings described below.

[0024] A warning message can alert a vehicle driver to a potential hazard. The warning is intended to encourage the driver to take corrective action to avoid or at least reduce the hazard. For example, if the warning is provided by the vehicle's curve warning system, the driver should reduce the vehicle's speed before the curve as a corrective action. This is because the curve warning system's algorithm, based on information about the curve and assumptions about the driver's skill level, has calculated that without the corrective action, the driver will enter the curve too fast. Excessive speed in the curve could, for example, cause the vehicle to be unable to negotiate a certain radius of the curve and be carried out of the curve. On an inside curve, the vehicle could, for instance, enter oncoming traffic. On an outside curve, the vehicle could leave the roadway.

[0025] The vehicle can be, in particular, a single-track vehicle, i.e., a two-wheeler, especially a motorcycle. With a single-track vehicle, the rider's skill greatly influences the possible cornering speed, as lean angle in the curve limits the achievable speed. A skilled rider will lean more heavily into the curve than an inexperienced rider. Therefore, the skilled rider can also achieve a higher speed through the curve than the inexperienced rider.

[0026] The assumptions used by the algorithm can only incompletely represent driving ability. This can lead to too frequent warnings, i.e., false positive warnings (FP = false positive), or to excessively high warnings. R. 415141

[0027] 4 - ten are warned, i.e. false negative warnings (FN = false negative) are issued.

[0028] The approach presented here reacts to false positive and false negative warnings and adjusts the warning frequency for future situations accordingly.

[0029] The warning frequency can be defined by a stored scaling factor. A larger scaling factor can be assigned if the corrective maneuver is detected without a preceding warning. Conversely, a smaller scaling factor can be assigned if no corrective maneuver is detected despite a preceding warning. The scaling factor can be a numerical value. The scaling factor can indirectly reflect the driver's skill level.

[0030] The scaling factor can also be a multidimensional scaling vector and have different values ​​for different warning situations. In this case, only the corresponding value of the vector needs to be changed in a specific warning situation. Using a scaling vector allows for a more differentiated representation of driving skill than a general scaling factor.

[0031] The warning frequency can remain the same if a corrective maneuver is detected after a preceding warning. The warning frequency can also remain the same if no corrective maneuver is detected without a preceding warning. If a warning is correctly issued, this is a true positive warning situation (TP = true positive). If a warning is not correctly issued, this is a true negative warning situation (TN = true negative). In this case, the warning function has reacted correctly, and there is no need to change the warning frequency.

[0032] The warning frequency can be increased proportionally to the intensity of a detected corrective maneuver. If the corrective maneuver is of high intensity, the warning frequency may be significantly too low. The intensity can reflect the speed or severity of the corrective maneuver. By taking the intensity into account, the warning frequency can be quickly adjusted to the correct value. R. 415141

[0033] 5 -

[0034] The warning frequency can be set cyclically. With each change, the warning frequency can be increased or decreased by a predefined increment. In particular, the scaling factor can be changed by the predefined increment. The increment can define the step size of the change. For example, the increment can be less than 20%, preferably less than 10%, less than 6%, or less than 3%, but more preferably greater than 0.5%, greater than 1%, or greater than 2% of an absolute value of the scaling factor. By adjusting the warning frequency incrementally, an excessively rapid change can be prevented. The warning frequency can adapt to the driver step by step.

[0035] To detect a correction maneuver, the trend of at least one measurement recorded by a vehicle sensor can be analyzed. The correction maneuver can be detected if the trend shows an outlier. An outlier can be an unusual trend and / or magnitude of the measurement. For example, the correction maneuver can be detected if the measurement changes faster than a predefined gradient. The correction maneuver can also be detected if the measurement exceeds or falls below a predefined threshold.

[0036] A vehicle deceleration, for example, can be evaluated as a measured value. The deceleration curve reliably reflects, for instance, an unusually strong or rapidly increasing deceleration, thus enabling a corrective maneuver.

[0037] As measured values, alternatively or additionally, for example, a roll angle and / or a roll rate of the vehicle can be evaluated. The roll angle can represent the lean angle of a single-track vehicle. The roll rate can represent the rate of change of the roll angle. In a normal cornering maneuver, a characteristic profile of the roll angle can be observed in a single-track vehicle. The roll angle increases at the beginning of the curve with an approximately constant roll rate until a maximum roll angle is reached at a vertex of the curve, then the roll angle decreases again with an approximately constant roll rate until the single-track vehicle has exited the curve. During a correction maneuver, both the roll angle and the roll rate can change significantly. R. 415141

[0038] 6 -

[0039] There are fluctuations, and in particular the roll rate also changes its sign, which is easily detectable.

[0040] The method is preferably computer-implemented and can be implemented, for example, in software or hardware, or in a hybrid form of software and hardware, for example, in an assistance system.

[0041] Furthermore, the approach presented here creates an assistance system, whereby the assistance system is trained to carry out, control or implement the steps of a variant of the procedure presented here in appropriate facilities.

[0042] The assistance system can be an electrical device, such as a control unit, with at least one processing unit for processing signals or data, at least one storage unit for storing signals or data, and at least one interface and / or a communication interface for reading or outputting data embedded in a communication protocol. The processing unit can be, for example, a signal processor, a so-called system ASIC, or a microcontroller for processing sensor signals and outputting data signals depending on the sensor signals. The storage unit can be, for example, flash memory, an EPROM, or a magnetic storage device. The interface can be configured as a sensor interface for reading sensor signals from a sensor and / or as an actuator interface for outputting data signals and / or control signals to an actuator.The communication interface can be configured to read or output data wirelessly and / or via a wired connection. The interfaces can also be software modules, such as those found on a microcontroller alongside other software modules.

[0043] Also advantageous is a computer program product or computer program with program code that can be stored on a machine-readable carrier or storage medium such as semiconductor memory, hard disk memory or optical memory and is used to carry out, implement and / or control the steps of the procedure according to one of the above-described R. 415141.

[0044] 7 -

[0045] embodiments are used, in particular when the program product or program is executed on a computer or device.

[0046] It should be noted that some of the possible features and advantages of the invention are described herein with reference to different embodiments. A person skilled in the art will recognize that the features of the assistance system and the method can be suitably combined, adapted, or exchanged to arrive at further embodiments of the invention.

[0047] Brief description of the drawings

[0048] Embodiments of the invention are described below with reference to the accompanying drawings, whereby neither the drawings nor the description are to be interpreted as limiting the invention.

[0049] Fig. 1 shows a representation of a process flow according to an exemplary embodiment.

[0050] Fig. 2 shows a representation of a correction maneuver in a roll angle-roll rate diagram.

[0051] The figures are schematic only and not to scale. Identical reference symbols denote identical or equivalent features.

[0052] Embodiments of the invention

[0053] Fig. 1 shows a representation of a process sequence according to an exemplary embodiment. The process is executed during a vehicle's journey using an assistance system of the vehicle. The vehicle is, in particular, a single-track vehicle. The process is described here using the example of a curve warning function.

[0054] The procedure begins with a first step 101 when approaching 100 a curve. For example, when a distance d of the vehicle from the curve- R. 415141

[0055] 8 - ve smaller than a warning distance d max distance is. The warning distance d max The distance is, for example, 200 meters.

[0056] In response to the approach 100, a calculation 102 is performed in a second step 103 to determine whether a warning message 104 should be issued. The curve warning system calculates whether the warning should be activated based on assumed parameters pi to p. n where pi corresponds, for example, to an assumed reaction time and P2 corresponds, for example, to an assumed target speed for this curve. Furthermore, the activation is calculated using a stored scaling parameter a for a warning frequency.

[0057] In a third step 105, it is checked whether a warning message 104 has been issued by the curve warning system or, more generally, by an assistance system. In both cases, in a subsequent step 107', 107" an observation 106' or 106" of driving behavior is carried out over a predefined time period or period t. The time period or period t can, for example, be set such that the curve or a vertex of the curve is reached within the time period or period t at a known speed. In step 109' or 109" it is observed whether the driver performs a corrective maneuver within the time period or period t.

[0058] If warning message 104 was issued and the correction maneuver 108 is detected in step 109', the warning was justified and a confirmation 110 of the scaling parameter a is issued in step 111'.

[0059] If warning message 104 was issued and no corrective action 108 is detected in step 109', warning message 104 was unjustified or a false alarm (FP). Then, in step 113', the scaling parameter a is changed 112. The scaling parameter a is reduced to decrease the warning frequency in the future. For example, the scaling parameter a can be multiplied by a factor less than 1.

[0060] If warning message 104 was not issued and a correction maneuver 108 is nevertheless detected in step 109, warning message 104 was missed (FN), i.e., incorrectly not issued. Then, in step 113, the following occurs: R. 415141

[0061] 9 - if a change is made to the scaling parameter a. In this case, the scaling parameter a is increased to raise the warning frequency in the future. For example, the scaling parameter a can be multiplied by a factor greater than 1.

[0062] If the warning message 104 was not issued and no correction maneuver 108 is detected in step 109, the absence of the warning message was justified and a confirmation 110 of the scaling parameter a is also issued in step 111.

[0063] Then, in step 115, a saving operation is performed (114) and the confirmed or changed scaling parameter a is stored and used for the next curve.

[0064] In one embodiment, the scaling parameter a is multiplied by 0.95 when it is to be decreased and by 1.05 when it is to be increased. This ensures that the scaling parameter a changes by the same relative increment of 5% with each change. Alternatively, the scaling parameter a can always be increased or decreased by a fixed, predefined absolute increment.

[0065] Fig. 2 shows an example of a correction maneuver 108 in a roll angle-roll rate diagram. The abscissa of the diagram represents the roll angle or bank angle in °. The ordinate of the diagram represents the roll rate or change in bank angle in ° / s.

[0066] Correction maneuver 108 takes place while a single-track vehicle is cornering. During correction maneuver 108, the driver of the single-track vehicle avoids a hidden obstacle in the curve.

[0067] The curve begins with an upright single-track vehicle without any rolling motion, i.e., at zero. 0 Roll angle and zero ° / s roll rate. For cornering, the driver leans the single-track vehicle into the curve, thus increasing the roll rate and therefore the roll angle. Then the driver perceives the obstacle and begins the correction maneuver 108. R. 415141

[0068] 10 -

[0069] The correction maneuver 108 begins with a counter-movement. The driver briefly rights the single-track vehicle to execute an evasive maneuver. During the righting phase, the roll rate changes sign and becomes negative. The roll angle, or lean angle, decreases again.

[0070] To be able to negotiate the curve after the obstacle, the driver quickly increases the roll angle. The roll rate then reverses its sign and rises sharply until the driver reaches the appropriate roll angle for the curve.

[0071] When exiting the curve, the roll rate becomes negative again and the roll angle decreases. After the curve, the driver experiences an overshoot, during which the roll angle briefly becomes negative until the single-track vehicle returns to an upright position.

[0072] The correction maneuver 108 exhibits characteristic roll angle and roll rate profiles that can be easily identified, for example, by pattern recognition.

[0073] Possible embodiments of the invention are summarized below or presented using slightly different wording.

[0074] A curve warning system with automatic learning self-calibration of intensity based on driver reactions is presented.

[0075] The typical challenge for driver assistance systems is balancing false positives (FP) and missed warnings (FN). Almost every warning function can take driver characteristics into account at one or more points or rely on typical model assumptions – at least a reaction time assumption is essential for a warning. In curve warning systems, individual deceleration values ​​and lean angles in curves can also be considered.

[0076] Obviously, it is difficult, if not impossible, to know all the important parameters – after all, there are far more complex criteria, such as the visibility of an R. 415141.

[0077] 11 -

[0078] A curve that is difficult to represent as a simple parameter. It is even more difficult for a familiar driver to correctly estimate the values ​​of these parameters in the respective situation, because factors such as fatigue, weather, surrounding traffic, etc. also have a strong effect.

[0079] Therefore, typical, averaged assumptions are usually used. If these assumptions are chosen to be too conservative or too sporty, the driver will experience too many or too few warnings. This can either decrease their acceptance – in extreme cases, they may ignore or deactivate the system – or a warning may be missed – in extreme cases, resulting in an avoidable accident.

[0080] In summary, every warning system faces the dilemma of achieving a suitable benefit (TP = true positive warnings) by choosing the appropriate parameters, while simultaneously keeping FP and FN low.

[0081] In the approach presented here, the warning system automatically adjusts its warning frequency by analyzing the driver's reaction to a previous warning or even a missed warning. This increases both safety and driver acceptance of the system.

[0082] Increased safety can prevent more accidents and make traffic situations safer – also for the traffic involved.

[0083] Self-calibration, as suggested here, can achieve good acceptance of the curve warning function.

[0084] The present concept describes a system that automatically adjusts the frequency and intensity of warnings to reduce the number of FP (Functional Prevention) and FN (Functional Prevention) warnings, thereby increasing driver acceptance. To achieve this, the driver's reaction after a warning is observed and interpreted to determine whether the warning was justified and helpful. The warning frequency and / or intensity is then adjusted accordingly for future warnings. R. 415141

[0085] 12 -

[0086] The system proposed here fulfills three core tasks. First, the parameters used are scaled according to a scaling parameter a to influence the warning frequency. Then, an analysis is performed to determine whether a response to a previous warning has occurred. In one embodiment, the intensity of the response is analyzed. After the analysis, the scaling parameter a is adjusted.

[0087] In one embodiment of the frequency parameter a, a simple linear scaling is proposed for scaling the scaling parameter a, where a is a scalar value restricted to an interval a \in [a m in, a max ], e.g., a \in [0.5, 2]. Scaling the parameters is then a simple multiplication by a for those parameters that trigger more warnings with increasing value. For example, for the parameter PReactionTime, the scaled value will be pReactionTime = a * PReactionTime. A longer reaction time results in more / earlier warnings.

[0088] In one embodiment, the parameters that trigger more warnings as their value decreases are divided by a. For example, the parameter PCurveSpeed ​​is considered to be the scaled value pCurveSpeed. =The I / O curve speed is used. The curve speed is an assumed speed at which the driver can safely negotiate the curve. At reduced curve speeds, more frequent and earlier warnings are issued to adjust the current vehicle speed.

[0089] A "corrective maneuver" is a driver intervention that indicates the driver misjudged the curve – for example, by entering it too fast, at the wrong angle, or for other reasons. The underlying assumption is that a warning should precede any corrective maneuver – precisely to prevent such maneuvers from occurring in the first place.

[0090] Many different indicators can suggest that the curve was misjudged. Physical factors can include, for example, abrupt braking or throttle input to correct for increased speed, and / or an abrupt change in lean angle. Physiological factors can include, for example, an increased heart rate or perspiration. The focus here is on R. 415141

[0091] 13 - in particular, focus on physical factors that can be easily detected with existing sensors.

[0092] For example, a corrective maneuver can be detected using outlier detection via machine learning. In this AI domain, data is analyzed using unsupervised learning (without labels), and rare events / outliers are automatically found.

[0093] Abrupt braking can be detected using a histogram. The histogram identifies typical braking decelerations and braking gradients that the driver experiences before and during a curve (detectable in the presented basic system). By comparing the current values ​​with the histogram values, an outlier can easily be identified as a deviation from the known pattern.

[0094] An abrupt lean angle correction can be detected via a lean angle-roll angle diagram. The depicted lean angle-roll angle diagram shows an atypical curve during a cornering maneuver with an obstacle. Such a corrective movement can be described as a pseudo-critical maneuver and is algorithmically detectable.

[0095] When adapting the scaling parameter a, in one embodiment, the parameter a is incrementally increased or decreased by linear scaling after each adjustment (FP or FN): For example, the warning frequency is decreased by decreasing a' = a * 0.95. Conversely, the warning frequency is increased by increasing a' = a * 1.05. Subsequently, a is again restricted to the defined limit range. a' = min(a max , rnax(a min , a))

[0096] Self-calibration can be further improved by, for example, using a vector ai, ..., a instead of a single scaling factor for a. n is used, which uses the individual parameters pi, ..., p n individually scaled.

[0097] The scaling parameter(s) a and Oi, respectively, can be scaled additively instead of linearly after an FP / FN warning (e.g., a' = a + 0.1). Alternatively, other functions can be used depending on the R. 415141

[0098] 14 - The criticality achieved in the curve can be used. The criticality can be recognized by the intensity of the correction maneuver.

[0099] The approach described here is not limited to curve warning systems. It can be applied to any type of warning system that aims to warn the driver of excessive speed (e.g., warnings about other hazards such as potholes, etc.).

[0100] Finally, it should be noted that terms such as "comprising," "encompassing," etc., do not exclude other elements or steps, and terms such as "a" or "an" do not exclude a plurality. Reference numerals in the claims are not to be considered as limitations.

Claims

R. 415141 15 - Claims 1. Method for setting a warning frequency of warnings (104) for a driver of a vehicle, wherein the warning frequency is increased when a corrective maneuver (108) by the driver is detected without a preceding warning (104), and the warning frequency is reduced when no corrective maneuver (108) is detected despite a preceding warning (104).

2. Method according to claim 1, wherein the warning frequency is defined by a stored scaling factor (a), wherein an increased scaling factor (a) is stored if the correction maneuver (108) is detected without the preceding warning (104) or a decreased scaling factor (a) is stored if no correction maneuver (108) is detected despite a preceding warning (104).

3. A method according to any of the preceding claims, wherein the warning frequency remains the same if, after a prior warning, (104) a corrective maneuver (108) is detected and if no corrective maneuver (108) is detected without a prior warning (104).

4. Method according to one of the preceding claims, wherein the warning frequency is increased proportionally to the intensity of a detected corrective maneuver (108).

5. Method according to one of the preceding claims, wherein the warning frequency is set cyclically and is increased or decreased by a predefined increment with each change.

6. Method according to one of the preceding claims, wherein, in order to detect the correction maneuver (108), a progression of at least one measured value acquired by a sensor of the vehicle is evaluated, wherein R. 415141 16 - the correction maneuver (108) is recognized when the trend shows an outlier.

7. Method according to claim 6, wherein a deceleration of the vehicle is evaluated as the measured value.

8. Method according to one of claims 6 to 7, wherein a roll angle and a roll rate of the vehicle are evaluated as measured values.

9. Assistance system, wherein the assistance system is configured to execute, implement and / or control the method according to one of the preceding claims in appropriate facilities.

10. Computer program product configured to instruct a processor, when the computer program product is executed, to execute, implement and / or control the method according to any one of claims 1 to 8.

11. Machine-readable storage medium on which the computer program product according to claim 10 is stored.

Images