METHOD FOR OPERATING AN OCCUPANT PROTECTION SYSTEM FOR A MOTOR VEHICLE

Description

[0001] The invention relates to a method for operating an occupant protection system for a motor vehicle according to the preamble of claim 1.

[0002] Occupant protection systems typically consist of various airbags, seatbelt pretensioners, etc., and are now designed and evaluated according to legally defined load cases. These load cases determine, for example, the impact speed, the object of the collision, and the occupant's position in the vehicle. Generally, the seating positions of an average-sized man and a small woman, as they would likely assume when driving a motor vehicle, are taken into account. These positions are also assumed for occupants in the front passenger seat.

[0003] In crash simulation and the design of the occupant protection system and vehicle, the evaluation is carried out by comparing the load values ​​measured in different body regions, such as accelerations, forces, changes in position or indentations, etc., with legal limits.

[0004] The assessment of occupant stress is essentially carried out using only a few (1 to <10) sample or test configurations.

[0005] The evaluation is based on individual load values ​​in the different body regions. Comparing different restraint systems becomes more difficult with an increasing number of test configurations, as a compromise must always be found regarding the probability of injury in each body region.

[0006] In contrast, for the triggering behavior during the driving operation of a motor vehicle, only a much smaller number of parameters are recorded, and the occupant protection devices are triggered via fixed, pre-calibrated trigger paths without taking the load values ​​into account.

[0007] DE 10 2004 037016 B4 describes a method for controlling motor vehicle occupant protection systems in a motor vehicle during an accident.

[0008] A primary algorithm calculates a combined accident severity factor based on signals from the vehicle's impact sensors, which detect acceleration, pressure, structure-borne sound, and / or vehicle deformation. This factor characterizes the severity of injuries to a vehicle occupant in the accident. The combined accident severity factor is calculated as a weighted average of the ratios of the loads on specified body parts to the maximum load on those body parts. Additionally, a separate supplementary algorithm, also based on the impact sensor signals, predicts the most likely future position of the vehicle occupant and determines the optimal activation times for the vehicle occupant protection systems based on this position.Based on a combination of the main algorithm with the additional algorithm, the vehicle occupant protection systems are ultimately controlled or triggered based on the calculated value of the common accident severity factor and the determined optimal triggering times.

[0009] The object of the invention is to more accurately evaluate the expected severity of injuries in various crash configurations (or seating positions of the occupant), such as those that can occur on the passenger side during autonomous driving or in a driver-controlled vehicle (e.g., occupants in a reclining position or with the seat moved backward), and to determine and activate the best possible protection concept for each configuration while keeping the computational effort in the control unit reasonable. This object is achieved by the features of claim 1. Advantageous further developments are described in the dependent claims.

[0010] This describes a method for operating an occupant protection system for a motor vehicle, in which the occupant protection system includes at least one control unit and various occupant protection devices, such as airbags, seat belt tensioners or other actuators, for example on the vehicle seat, to optimize the seating position or position of upholstery surfaces in relation to the occupant.

[0011] In addition, sensors are provided to detect the type and severity of an impending or occurring accident, such as acceleration sensors, yaw rate sensors or crash contact sensors, or preferably also predictive environmental sensors, such as radar, lidar or camera sensors, ultrasonic sensors or the like, for evaluation even before the actual collision.

[0012] In addition, at least one occupant detection device is provided to recognize an occupant situation, for example an interior camera, whereby the position of the occupant in the vehicle as well as the position of the seat can also be detected via seat mats or other sensor types.

[0013] The control unit triggers the occupant protection system based on the signals from all these sensors.

[0014] However, whereas previously the sensor signals were fed into a few trigger paths, which were indeed optimized and calibrated in advance based on load values, but the load values ​​were no longer taken into account during the actual triggering, the approach in the present procedure is different.

[0015] Thus, through crash tests or, preferably, simulations for a predetermined number of different sample occupant situations (first dimension) and a predetermined number of sample accident situations (second dimension), at least a number of trigger configurations for the occupant protection system are defined (third dimension), and the influence of at least the relevant trigger configurations on the hazardous situation is evaluated. Multiplying the number of sample occupant situations by the number of sample accident situations by the number of trigger configurations to be considered naturally increases the storage requirements, but also, of course, the quality of the data and ultimately the trigger decision.

[0016] In addition, a further crucial refinement is made by determining at least one stress value for each of the occupant's body parts and normalizing it against a target value.

[0017] This allows for the evaluation of multiple different stress values ​​for each body part, such as acting forces or accelerations, as well as resulting displacements, deformations, etc. Normalization against a target value creates a relative quantity, which, when normalized, can be evaluated much more easily across physically diverse parameters.

[0018] From the standardized load values ​​or derived values ​​for the majority of body parts, a maximum value and an average value are also calculated.

[0019] The maximum value represents the highest value of a normalized load value as a measure of the local load on a body part, while the averaging process also determines a further, independent decision parameter regarding the overall load.

[0020] All this data is stored in a suitable format, such as tables, in memory. The effort required for simulation and evaluation is considerable, but it only occurs during the preliminary simulation and does not need to be performed continuously during operation. Nevertheless, this refined evaluation allows for a significantly improved trigger decision.

[0021] During normal vehicle operation, when an impending or occurring accident situation is detected, the current situation is compared with a predefined number of sample accident situations, and at least one of the most closely matching sample accident situations is determined. This can be exactly one of the predefined sample accident situations directly, or, for example, an evaluation can be made by interpolation from a multiple of sample accident situations.

[0022] On the other hand, the current occupant situation is compared with the predefined number of sample occupant situations, and at least one of the best-matching sample occupant situations is determined. Again, this can be directly one of the predefined sample accident situations or obtained through interpolation from a plurality of sample accident situations.

[0023] The algorithm then only needs to select from the number of trigger configurations of the occupant protection system the trigger configuration for which the maximum value is initially the lowest and, if there are multiple remaining trigger configurations, the average value is also the lowest.

[0024] In one embodiment, the most suitable sample accident situation for the current accident situation can be determined from the given number of sample accident situations by calculating the relative deviation of the parameters currently recorded by the sensors for detecting the type and severity of an impending or occurring accident from each of the target values ​​assigned to the sample accident situation, and by identifying the sample accident situation with the smallest overall deviation across all parameters.

[0025] Similarly, in a preferred embodiment, the best matching sample occupant situation(s) for the current occupant situation can be determined from the given number of occupant situations by determining the relative deviation of currently recorded parameters from the target value assigned by the at least one occupant detection device to each of the sample occupant situations and by identifying the sample occupant situation with the smallest total deviation across all parameters.

[0026] InIn further training, it is also conceivable that at least some or every parameter is assigned a weighting factor, whereby at least one parameter is assigned a weighting factor that differs from the others, and the sum of the weighted relative deviations is calculated. This takes greater account of the unequal relevance of the individual parameters for triggering the alarm and may, in particular, also allow for some parameters to be set to a weight of "zero," meaning they are effectively disregarded for the specific situation.

[0027] A further development takes this approach even further. For at least one individual occupant protection device or a subgroup of the occupant protection system's trigger configurations, only one subgroup of parameters needs to be considered. This is preferably the parameter most relevant to the trigger decision, which is then considered for that specific occupant protection device or subgroup. Therefore, it is not necessary to consider all parameters for every occupant protection device; instead, specific subgroups can be formed from the occupant protection system's trigger configurations, for which the evaluation is performed using a significantly smaller number of parameters related to the occupant situation and / or accident scenario.However, the trigger configuration to be selected from the number of trigger configurations of the occupant protection system must always be the one that meets the two-dimensional evaluation with regard to both the maximum value and the mean value.

[0028] Therefore, for the current occupant situation, the best matching pattern occupant situation and / or for the current accident situation, the best matching pattern accident situation is determined only on the basis of this subgroup of the number of parameters, at least the parameter most relevant for the triggering decision.

[0029] InIn a further embodiment, it is also provided that a weight factor is assigned to the majority of body parts for the standardized load values ​​or quantities derived therefrom, whereby a weight factor differing from the others is provided for at least one load value, thus determining a weighted average. Therefore, not all parameters necessarily contribute equally to the averaging; rather, their relevance to the triggering of the system can also be taken into account.

[0030] Another approach involves categorizing the standardized stress values ​​for most body parts into a predefined number of hazard levels. The maximum value used is the highest value within each hazard level for the body parts, determined during the triggering configuration and in the sample occupant and accident scenarios. These hazard levels can be effectively visualized using a color scheme, where values ​​that are not critical for the occupant are represented by green, more critical values ​​by yellow to orange, and extremely dangerous or even fatal values ​​by red, etc. In an algorithm, this can also be implemented using simple level values, for example, from 1 to 5.

[0031] In this case, the maximum values ​​within a hazard level are preferably not further differentiated, but rather the average values ​​have a somewhat stronger influence on the triggering decision.

[0032] Another preferred configuration arises if, for the current accident situation, instead of the one best-matching model accident situation, a plurality of closest model accident situations are determined, and / or for the current occupant situation, instead of the one best-matching model occupant situation, a plurality of closest model occupant situations are determined, and then, for the current accident situation and / or current occupant situation, for at least individual occupant protection devices, at least for the subgroup of parameters most relevant for their activation, these parameters are determined from the interpolation of the values ​​of the plurality of closest model accident situations and / or plurality of closest model occupant situations.Such interpolation can make the triggering decision even more precise or allow for significantly fewer predefined and therefore fewer stored sample accident or occupant scenarios. The effort involved in interpolation remains manageable even when limited to just a few parameters. This approach is particularly justifiable when considering only a limited number of parameters for individual occupant protection systems, with only the most important ones being interpolated. For different occupant protection systems, only subgroups of parameters are considered, always including the maximum value or the derived hazard level, as well as the average value.

[0033] The invention is implemented in a control unit for an occupant protection system for a motor vehicle, which has at least one input interface for connecting sensors for detecting the type and severity of an impending or occurring accident, as well as at least one occupant detection device, and at least one output interface for connecting occupant protection devices. According to the invention, an algorithm for carrying out the procedure is stored with corresponding standardized load values ​​for the body parts in the respective predefined sample occupant situations and sample accident situations for the respective trigger configurations provided for each.

[0034] The invention will be described in more detail below with reference to figures and exemplary embodiments.

[0035] Figure 1This section first outlines the basic hardware structure of an occupant protection system for a motor vehicle. The core and key, independently executable unit is a control unit (ACU), which shows the connections (indicated by arrows) to the input interfaces for connecting sensors for detecting the type and severity of an impending or occurring accident, here designated SC1...SCm, as well as the connections (also indicated by arrows) to the input interfaces for connecting occupant detection devices SP1 to SPn. The connections to the output interfaces for connecting occupant protection devices R1...Ro are also shown as arrows.The control unit or occupant protection system differs from a conventional system only in the algorithm stored in the memory, in particular the stored sample occupant situations and sample accident situations as well as the standardized load values ​​for the body parts determined for the respective parameters in the considered trigger configurations.

[0036] The Figure 2 Now outlines these three dimensions of the decision: the occupant situation as the "position" axis, the accident situation as the "crash" axis, which of course also includes the impending accident situations based on environmental perception, i.e., "precrash", and the third dimension the various trigger configurations.

[0037] The signals from sensors SP1 to SPn are incorporated into the occupant situation as the "position" axis for detecting the occupant situation.

[0038] The accident situation is considered as a "crash" axis against the direct accident sensors, such as acceleration and yaw rate sensors, as well as the signals from environmental detection, such as camera, radar or lidar and also ultrasonic sensors.

[0039] This results in a current occupant situation and accident situation, represented here as points P(act),C(act), to which the various trigger configurations are added as a third dimension. As indicated by the thick line A1..An, some trigger configurations are conceivable, while others are not, and the line is therefore broken.

[0040] The boxes are intended to illustrate that only specific, selected sample occupant situations and sample accident situations are stored, and their size can vary, meaning they can directly cover different sized areas of the occupant situation and accident situation. Furthermore, the space does not have to be completely filled; gaps, or free spaces, remain between the individual boxes. While no single sample directly covers these spaces, the best matching sample is then selected according to the described rules, or an interpolation is even determined from the neighboring samples.

[0041] It should be clarified once again that, in order to minimize storage requirements and algorithmic effort, it is by no means necessary to store all theoretically conceivable trigger configurations for all occupant protection devices, let alone all parameters or standardized load values ​​for them. Instead, for individual occupant protection devices or specific subgroups, only the most relevant triggering parameters need to be stored, and for other occupant protection devices, other, or rather, from the perspective of current accident and occupant protection situations, only a significantly smaller set of trigger configurations and parameters are available or even considered.

[0042] Each such box conceptually represents a trigger configuration for a stored sample occupant situation and sample accident situation, as well as the standardized load values ​​stored for this purpose as parameters.

[0043] This is intended to be in Figure 3 This can be illustrated using the table, with virtually every box consisting of Figure 2 at least one such spreadsheet is sufficient.

[0044] Stored there for the most relevant parameters, referred to here abstractly as Head1, Head 2..., Arm1, Arm2 etc., are the load values ​​Value(Ist) determined by tests or simulation for this trigger configuration in the defined and stored sample occupant situation and sample accident situation, which, however, are normalized to a normalized load value % compared to reference values ​​of the load Value(Ref).

[0045] This means that a percentage load is preferably determined, through which physically completely different quantities, such as deformations, body displacements or dimensions, etc., are converted to a comparable scale, namely the respective percentage risk to the occupant.

[0046] From these different normalized load values ​​% or from these derived values, two decision-relevant values ​​are derived: on the one hand, the value Max(x) as the maximum value of the normalized load values, and on the other hand, the mean value Ø(x).

[0047] The trigger configuration selected from the number of trigger configurations of the occupant protection system is therefore the one for which the maximum value is initially the lowest and, if there are multiple remaining trigger configurations, the average value is also the lowest.

[0048] The visualization in the Figures 2 and 3This is purely a sketch, and each box preferably contains several smaller worksheets, separated by occupant protection device subgroups, with only the necessary parameters for each subgroup. Since different occupant protection device subgroups may not be relevant for some combinations of occupant and accident situations, or may cover entirely different value ranges (i.e., they may be of different sizes in the box analogy), several such boxes are provided to simplify and reduce memory usage in storing the required values.

[0049] The occupant loads measured in the individual body regions of the dummy are thus combined into a weighted occupant load index, which indicates the percentage of the limit values ​​for each body region to which the occupant is subjected on average. Individual load values ​​for specific body regions that are close to or even exceed the legal limit are preferably considered separately in a hazard level, for example, visualized as a color scale (green, yellow, red). This allows the occupant load for a multitude of load cases to be assessed with regard to the expected severity of injury based on a single numerical value (occupant load index) and the corresponding hazard level, i.e., the color code.

[0050] InWith regard to an autonomously driving vehicle in which numerous seating positions are possible, the expected severity of injury for a specific seating position, which can be determined in simulations or tests, can then be plotted in the form of a matrix that shows the occupant's risk of injury in different seating positions (with different restraint system configurations).

[0051] An interpolation between the individual support points with suitable functions then provides an indication of the expected risk of injury between the support points.

[0052] Different crash configurations (impact speed, occupant position, occupant size, restraint system configuration (airbags, belts,...), etc.) can be evaluated and easily compared using a load value, both with regard to the maximum point hazard based on the maximum value or derived hazard level, and additionally as a further criterion of the average load.

[0053] This offers advantages in all applications where a variety of configurations need to be considered, e.g., in autonomous driving.

[0054] In principle, this applies to all cases where an evaluation or optimization is to be carried out in a multitude of possible configurations with a multitude of evaluation criteria.

[0055] The Figure 4The text now outlines a matrix of stress values ​​determined from a simulation of a vehicle for a specific occupant and accident scenario, and of course a defined triggering strategy. Six levels are provided, ranging from 1 for absolutely non-critical to level 6 for a significant exceedance of the limit values ​​and thus a risk of life-threatening injuries.

[0056] It is clearly recognizable that the moderate mean value (Ø(x)) across all stress values, which would otherwise be assigned to Level 4, but which, due to the one unacceptably high chest stress value (Resultant Acc. 3ms exceedance (g)) in Level 6, leads to the overall Level 6 assessment of the situation.

[0057] This overall assessment can now be transferred into a decision matrix that, in addition to a standard trigger, also shows alternative trigger strategies or adaptations of the parameters, such as the influence of a changed angle of the backrest.

[0058] Although this is only an example, it should be possible to achieve hazard level 5 for an adaptive system and therefore choose this triggering variant.

Claims

1. A method for operating an occupant protection system for a motor vehicle, having a control unit (ACU) and occupant protection devices (R1... Ro) as well as sensors (SC1... SCm) for identifying the type and severity of an imminent or occurring accident, using at least one occupant sensing means (SP1... SPn) for detecting an occupant situation, - wherein the control unit (ACU) triggers the occupant protection system on the basis of these signals, characterised in that - at least a number of triggering configurations of the occupant protection system are respectively specified for a specified number of different model occupant situations and a specified number of model accident situations, and the influence of at least the possible trigger configurations on the hazard situation is assessed, - by determining, for a plurality of body parts of the occupant, at least one stress value (Value(Ist)) for each body part and normalising it with respect to a default value (Value(Ref)), - wherein, from the normalised load values (%) or a variable derived therefrom for the plurality of body parts - on one hand, a maximum value (Max(x)) is formed, which represents the highest value of a standardised loading value as a measure of a local load on a body part, and - in addition, on the other hand, a mean value (Ø(x)) is formed, which forms a further, independent decision variable relating to an overall load, and during ongoing driving operation of the motor vehicle, upon an imminent or occurring accident situation being detected - on one hand, the current accident situation (C(akt)) is compared with the specified number of model accident situations (P(x)) and at least one best matched model accident situation is determined, - on the other hand, the present occupant situation (P(akt)) is compared with the prescribed number of model occupant situations (P(x)) and at least one model occupant situation which matches best is determined, - and then that trigger configuration (A(x)) for which the maximum value (Max(x)) is initially the lowest and, in the case of a plurality of remaining deployment configurations, the average value (Ø(x)) is also the lowest, is selected from the number of deployment configurations of the occupant protection system.

2. The method as claimed in claim 1, characterised in that the model accident situation with the best match for the current accident situation is determined from the specified number of model accident situations, by determining the relative deviation of the parameters currently sensed by the sensors for detecting the type and severity of an imminent or occurring accident with respect to a respective default value assigned to the model accident situation, and determining the model accident situation with the smallest deviation in sum total across all parameters.

3. The method as claimed in claim 1, characterised in that the model occupant situations that best match the current occupant situations is determined from the specified number of occupant situations, by determining the relative deviation of currently detected parameters of the default value assigned by the at least one occupant sensing means to a respective one of the model occupant situations, and determining the model occupant situation with the smallest deviation over all parameters.

4. The method as claimed in claim 2 or 3, characterised in that, wherein a weight factor is assigned to each parameter, wherein a weight factor which differs from the others is provided for at least one parameter, and the sum of the relative deviation thus weighted is determined.

5. The method as claimed in claim 2 or 3, characterised in that, for at least one individual occupant protection device or a subgroup of the number of deployment configurations of the occupant protection system, in each case only a subgroup of the number of parameters, at least of the parameters most relevant to the triggering decision, is taken into account, and for the current occupant situations, the model occupant situations that best correspond and / or for the current accident situation, the model accident situation best matching is determined only on the basis of this subgroup of the number of parameters, at least that parameter which is most relevant to the triggering decision.

6. The method as claimed in claim 5, characterised in that, for the plurality of body parts, a respective weight factor is assigned for the standardised loading values or variables derived therefrom, wherein a weight factor that deviates from the others is provided for at least one loading value, and thus a weighted mean value is determined.

7. The method as claimed in claim 1, characterised in that, for the plurality of body parts, the normalised exposure values are divided into a specified number of danger levels and the maximum value of the danger levels of the body parts that was determined during the triggering configuration and the model occupant situations and the model accident situations is used as the maximum value.

8. The method as claimed in any of the preceding claims, characterised in that for the current accident situation, instead of the one model accident situation which matches best, a plurality of closely matching model accident situations, and / or for the current occupant situation, instead of the one model occupant situation that matches best, a plurality of closely matching model occupant situations are determined and for the current accident situation and / or current occupant situation for at least individual occupant protection devices, at least for the subgroup of parameters most relevant for the triggering thereof, these parameters are determined from the interpolation of the values of the plurality of closest model accident situations and / or the plurality of closest model occupant situations.

9. A control unit (ACU) for an occupant protection system for a motor vehicle having - at least one input interface for connecting sensors (SC1... SCm) for identifying the type and severity of an imminent or occurring accident, and at least one occupant sensing means (SP1... SPn) and - at least one output interface for connecting occupant protection devices (R1... Ro), characterised in that an algorithm for carrying out the method as claimed in one of the preceding claims is stored with corresponding normalised loading values for the body parts in the respectively specified model occupant situations and the model accident situations in the respectively provided trigger configurations for them.