Method for controlling a passenger restraint means in a vehicle

Abstract

The invention relates to a method for controlling a passenger restraint means in a vehicle, in which at least one impact sensor supplies at least one impact signal, a triggering criterion derived from the impact signal is compared with a triggering threshold, and an ignition stage of the passenger restraint means is triggered immediately at or after a predefined triggering time if the triggering criterion exceeds the triggering threshold, wherein the triggering threshold and / or the timing element is increased or reduced depending on a curve of a crash severity parameter (CSP) derived from the impact signal or from another impact signal of at least one other impact sensor and representing crash severity. According to the invention, the triggering threshold and / or the timing element is increased or reduced as a function of the curve of the crash severity parameter (CSP) in such a way that the greater the crash severity, the later the ignition stage is triggered.

Description

[0001] Method for controlling an occupant restraint system in a vehicle

[0002] The invention relates to a method for controlling an occupant restraint device in a vehicle according to the features of the preamble of claim 1.

[0003] As described in EP 1 107 887 B1, a method for controlling an occupant protection device in a vehicle using dynamic switching thresholds and a control device for this purpose are known from the prior art. In this method, an impact sensor provides an impact signal, a trigger criterion derived from the impact signal is compared with a threshold, and the occupant protection device is controlled depending on whether the threshold is exceeded by the trigger criterion. The threshold contains at least one term that is determined by the impact signal and, depending on the nature of the impact signal, contributes either to increasing or decreasing the threshold.

[0004] German patent DE 10040 111 A1 describes a method for determining the triggering point for restraint systems. The difference between measured acceleration values ​​over time is calculated, and the magnitude of this difference is then integrated. This integral is compared to at least one threshold value. If the integral does not exceed at least one threshold value up to one or more predefined time points during the crash, the position of a trigger threshold for the measured acceleration or for a derived change in velocity is adjusted to reduce the triggering sensitivity.

[0005] The invention is based on the objective of providing a method for controlling an occupant restraint system in a vehicle that is improved compared to the prior art. This objective is achieved according to the invention by a method for controlling an occupant restraint system in a vehicle with the features of claim 1.

[0006] Advantageous embodiments of the invention are the subject of the dependent claims.

[0007] In a method according to the invention for controlling an occupant restraint system in a vehicle, at least one impact sensor provides at least one impact signal. A trigger criterion derived from the impact signal is compared with a trigger threshold, and an ignition stage of the occupant restraint system is triggered immediately or after a predetermined trigger time if the trigger threshold is exceeded by the trigger criterion. The trigger threshold and / or the trigger time is increased or decreased depending on the course of a crash severity parameter derived from the impact signal or from another impact signal of at least one other impact sensor and representing a crash severity, i.e., in particular, a parameter that is calculated proportionally to the crash severity.The trigger threshold and / or the trigger time is increased depending on the course of the crash severity parameter, such that the ignition stage is triggered later the greater the crash severity is, or reduced from a relatively high level initially, so that the ignition stage is triggered earlier the lower the crash severity is.

[0008] In one embodiment, a restraint force-reducing ignition stage of the occupant restraint device is triggered as the ignition stage.

[0009] In one embodiment, the ignition stage activates a switchable belt force limiter of a safety belt or an actively actuated venting device of an airbag.

[0010] In one embodiment, the trigger criterion itself is used as the crash severity parameter. This can be formed, in particular, from the summation or integral of acceleration and / or pressure signals. To prevent positive and negative signal components from canceling each other out, the signals can be aligned before summation or integration.

[0011] In one embodiment, the crash severity parameter (CSP) is determined by a loss of vehicle speed derived from the impact signal from the start of the crash. In another embodiment, the trigger threshold is dynamically raised depending on the crash severity parameter.

[0012] In one embodiment, the lifting is achieved via static threshold stairs. In another embodiment, the lifting is achieved via decreasing threshold stairs.

[0013] In one embodiment, the trigger time is delayed by a timer that is proportional to the crash severity parameter.

[0014] In one embodiment, an initially high trigger threshold is reduced as long as the crash severity parameter remains below a certain limit.

[0015] This method is intended for controlling at least one restraint force-reducing ignition stage of the occupant restraint system after a crash, i.e., after a collision of the vehicle. The ignition stage is, in particular, a switching stage. In the context of this application, the term ignition stage or switching stage is understood to mean, in particular, a triggering stage that can be activated pyrotechnically or otherwise. Thus, in the context of this application, the term ignition stage does not necessarily mean that the triggering is pyrotechnic.

[0016] The method implements an inverse trigger logic to control at least one restraint force-reducing ignition stage of the occupant restraint system.

[0017] The timing of occupant restraint systems is controlled by conventional crash algorithms inversely proportional to the measured impact severity. This means that a more severe crash leads to a faster exceedance of the corresponding trigger thresholds than a less severe crash, because the associated signals, such as acceleration or pressure, or quantities derived from them, reach certain limits more quickly. This logic is suitable for the activation stages of occupant restraint systems that increase occupant restraint or their coupling to a vehicle structure, for example, the activation stages for airbag deployment or the activation of seatbelt pretensioners. This is because a faster forward displacement of the occupants with increasing crash severity must be counteracted more quickly. However, for activation stages of occupant restraint systems that increase the restraint force on the occupants, this logic is not applicable.This control logic is fundamentally unsuitable for systems whose coupling to the vehicle structure is reduced, for example, for the ignition stages of active vents (i.e., airbag vents that can be actively activated by targeted control) or for the ignition stages of pyrotechnically switchable belt force limiters for seat belts, and / or which release a forward displacement path. Therefore, to date, additional crash-independent state parameters of the vehicle and occupants are evaluated to control these ignition stages, in order to still be able to control these ignition stages for occupant restraint systems appropriately for the situation.

[0018] The procedure described below makes this unnecessary, but it can still be used, for example, in combination or for plausibility checks.

[0019] To trigger the activation stages of occupant restraint systems, which reduce the restraint force on the occupants to relieve their strain, a precisely opposite, i.e., inverse, control logic is required. This involves activating these stages later in the event of a higher crash severity. The crash progression must be monitored for a certain minimum duration before the activation stage can be triggered to largely prevent an initially mild crash from developing into a very severe impact. For example, the activation threshold can be dynamically raised based on at least one parameter calculated proportionally to the crash severity, such as acceleration or pressure signals, so that its exceedance is delayed as the crash severity increases. This at least one parameter is referred to here as the crash severity parameter (CSP).

[0020] To calculate this crash severity parameter, it can be particularly advantageous to evaluate an intrusion velocity, for example, via the time difference of the integral of an acceleration signal increase at satellite sensors and at an airbag control unit (ECU), and / or a crash progression or occupant forward displacement, for example, via a single or double integrated acceleration at the satellite sensors or at the airbag control unit with or without a reset component, and / or a time difference of threshold exceedances, for example, ignition decisions of different occupant restraint devices such as seatbelt pretensioners and different airbag stages. Additionally, it can be advantageous to calculate several different crash severity parameters in parallel and relate them to each other, for example, by determining minimums or maximums, summing them, or in other ways.

[0021] In addition, a time element is also possible, which is proportional to the crash severity parameter and is connected after a conventional threshold crossing.

[0022] Both options can be combined. Furthermore, an initially very high trigger threshold, adapted to the most severe design-relevant crash types, can be continuously reduced during the crash, as long as the crash severity parameter remains below a certain limit.

[0023] The magnitude and rate of the increase or decrease of the crash severity parameter-dependent trigger thresholds for the methods mentioned above are selected so that the triggering of the restraint force-reducing ignition stages occurs as early as possible in a typical very minor crash and as late as possible in a typical severe crash. To largely prevent premature triggering of the crash severity parameter calculation due to misuse or similar events, such as when the vehicle is driven on a rough road, a separate start threshold for the crash severity parameter may be useful, or its calculation may only begin with the ignition point of another occupant restraint device or another ignition stage of the occupant restraint device, particularly an airbag or seatbelt pretensioner. Additionally, a combination or plausibility check with crash-independent state parameters is possible.

[0024] The method described here enables a crash pulse-dependent determination of the correct triggering time of restraint force-reducing ignition stages of occupant restraint devices during the crash process, adapted to the impact severity, and a complete or extensive independence from impact-severity-independent state parameters in the control of restraint force-reducing ignition stages of occupant restraint devices.

[0025] The following describes the procedure, in particular using crash-induced velocity reduction as a crash severity parameter. However, other crash severity parameters can also be used for the procedure, such as the velocity difference at the location of an upfront sensor, typically on a front crossmember, and the airbag control unit signals in the center of the vehicle. This allows for significantly earlier signal analyses that are nevertheless dependent on crash severity. The double integral of acceleration, as a measure of the forward displacement of the occupants, also has a very direct relationship to the ignition points of the seatbelt force limiter and active vent.

[0026] As an alternative to the embodiments of the method described below, it is also possible, for example, that the crash severity parameter signal curve, with which the triggers described in more detail below are started or stopped for the various timers and dynamic thresholds, is not identical to the crash severity parameter curve that must exceed the threshold for the eventual activation of the occupant restraint devices.

[0027] Exemplary embodiments of the invention are explained in more detail below with reference to drawings.

[0028] This shows:

[0029] Fig. 1 schematically shows a variant of a method 1 of a procedure for controlling an occupant restraint device in a vehicle,

[0030] Fig. 2 schematically shows another variant of method 1 of the procedure,

[0031] Fig. 3 schematically shows a variant of method 2 of the procedure,

[0032] Fig. 4 schematically shows another variant of method 2 of the procedure,

[0033] Fig. 5 schematically shows a variant of method 3 of the procedure,

[0034] Fig. 6 schematically shows another variant of method 3 of the procedure, and

[0035] Fig. 7 schematically shows a detailed representation of a further variant of method 3 of the process. Corresponding parts are designated with the same reference numerals in all figures.

[0036] With reference to Figures 1 to 7, a method for controlling an occupant restraint system in a vehicle is described below. This method is designed to control at least one restraint force-reducing ignition stage of the occupant restraint system after a crash, i.e., after a collision of the vehicle. The ignition stage is, in particular, a switching stage. In the context of this application, the term ignition stage or switching stage is understood to mean, in particular, a triggering stage that can be activated pyrotechnically or otherwise. Thus, in the context of this application, the term ignition stage does not necessarily mean that the triggering is pyrotechnic.

[0037] The basic idea behind the solution described below is:

[0038] The inverse shift of ignition times is achieved by evaluating the different average changes in crash severity parameters for more severe and less severe crashes, i.e., the slope of the crash severity parameter curve at specific time intervals or in a continuous evaluation until the desired occupant restraint system is triggered. If the velocity reduction is used as the crash severity parameter (CSP), this corresponds to the averaged or low-pass filtered acceleration at the respective time. For this purpose, either the change in the respective crash severity parameter value is determined at one or more predetermined time points, or the respective time period required until one or more predetermined crash severity parameter thresholds are exceeded is determined.The slope of the crash severity parameter curve determined in this way serves as the basis for positioning the trigger threshold for activating the restraint force-reducing ignition stages of the occupant restraint system, or for directly determining the ideal ignition point via an ignition timer, i.e., a timing element. With the described embodiments of the method, the trigger threshold is raised, or the ignition point is delayed, the higher the crash severity parameter slope value.

[0039] General boundary conditions: For the force-reducing ignition stages of the occupant restraint system, a crash load case is known that defines a lower limit of crash severity, above which their activation should occur. This is typically a vehicle crash at approximately 30 km / h with 100% overlap against a rigid wall, hereinafter referred to as a light crash (LC). In this case, activation of the force-reducing ignition stages of the occupant restraint system at typically 35 ms after wall contact is advantageous.

[0040] Simultaneously, a design load case for the upper limit of crash severity is known, up to which activation should occur. This is typically a vehicle crash at approximately 55 km / h with 100% overlap against a rigid wall, hereinafter referred to as severe crash SC. Here, activation at approximately 60 ms after wall contact is advantageous.

[0041] Experience has shown that when using time-varying trigger thresholds, it is advantageous to initiate the threshold change only after a certain speed reduction, typically around 3 km / h, within less than approximately 20 ms, and / or at the point of activation of other occupant restraint systems. This prevents minor driving events, such as potholes or driving over curbs, that precede the actual crash from influencing the trigger threshold position. For the two aforementioned limit load cases, this point in time is typically reached at approximately 15 ms and 10 ms, respectively, after wall contact. Therefore, for the lower limit load case (i.e., the light crash LC), a maximum observation period for the crash severity parameter curve of typically around 20 ms after the start of the threshold calculation is obtained, during which the activation of the restraint force-reducing ignition stages of the occupant restraint system is not required.

[0042] The following describes three basic methods, i.e., embodiments of the process.

[0043] Method 1: Crash signal-dependent inverse trigger threshold increase. As mentioned above, a trigger threshold can be dynamically raised depending on at least one crash severity parameter (CSP), which is calculated proportionally to the crash severity, for example from acceleration or pressure signals, so that its exceedance is delayed as the crash severity increases. Method 2: Threshold-dependent negative timer. As mentioned above, this alternative Method 2 uses a timer that is proportional to the crash severity parameter (CSP) and is connected after a conventional threshold exceedance.

[0044] Method 3: Crash signal-dependent inverse threshold reduction. As mentioned above, this alternative Method 3 involves a trigger threshold that is initially very high and adapted to the most severe design-relevant crash types, but is continuously reduced during the crash as long as the crash severity parameter CSP remains below a certain limit.

[0045] These three methods are illustrated in several variations in Figures 1 to 7. Figures 1 to 7 each show a diagram depicting the crash severity parameter CSP as a function of a crash-progressive parameter CP for the mild crash (LC) and the severe crash (SC). In these examples, the crash severity parameter CSP is a reduction in speed in km / h. The crash-progressive parameter CP is time in seconds.

[0046] The embodiment of the method according to Method 1 will first be described in more detail below.

[0047] In one variant of this method 1, a dynamic increase in the trigger threshold is achieved via static threshold steps, as shown by way of example in Figure 1.

[0048] For the desired end positions ELC and ESC of the dynamic trigger thresholds for the limit load cases, i.e., for the light crash (LC) and the severe crash (SC), typical speed reduction values ​​are known. These values ​​correspond to the aforementioned typical ignition times of approximately 15 km / h for the lower limit load case (LC) and approximately 48 km / h for the upper limit load case (SC). These end positions ELC and ESC are predefined accordingly. To also address the load case scenario between the two limit load cases with adequate ignition times, one or more intermediate trigger threshold levels with their corresponding switching thresholds are optionally defined, for example, at equidistant intervals.To allow for some tolerance to variations in crash severity parameters, an observation period TB is optionally shortened so that there is a difference of at least 3 km / h between the typical velocity reduction value at the end of the observation window and the target value for the lower limit load case, i.e., for the light crash LC. In the example shown, the observation period TB is 18 ms. A starting threshold STA is defined for the start of the observation period TB, in this example at 3 km / h. The observation period TB begins when the respective profile of the crash severity parameter CSP exceeds the starting threshold STA. This exceedance, relative to the wall contact at T=0 or the start time of CP, occurs earlier for the heavy crash SC than for the light crash LC.

[0049] A first switching threshold, SSI, is applied between these two speed values, for example, midway between them. The optional second switching threshold, SS2, is applied such that it is reached by the typical crash severity parameter profile of the upper limit load case, i.e., the severe crash SC, before the end of the observation period. A relevant region RB of the crash severity parameter profile CSP for the dynamic trigger threshold increase thus lies between the start threshold STA and the switching threshold SSI or SS2.

[0050] As shown in Figure 1, the dynamic trigger threshold, starting from a starting level at the end position ELC for the dynamic trigger threshold for the light crash SC, jumps to the next higher level if the corresponding switching threshold SSI, SS2 is exceeded within the observation period TB.In the example shown, the dynamic trigger threshold for the severe crash SC jumps from the starting level at the end position ELC for the dynamic trigger threshold for the light crash SC to an optionally specified first increased level EEN of the dynamic trigger threshold after the first switching threshold SSI is exceeded within the observation period TB, and after the second switching threshold SS2 is exceeded within the observation period TB from this optionally specified first increased level EEN of the dynamic trigger threshold to a second increased level of the dynamic trigger threshold, which corresponds to the end position of the dynamic trigger threshold for the severe crash.If the first increased level EEN of the dynamic trigger threshold is not provided, the dynamic trigger threshold jumps directly to the level of the end position of the dynamic trigger threshold for the severe crash after exceeding the first switching threshold SSI or the second switching threshold SS2, for example.

[0051] If the respective crash severity parameter CSP exceeds the applicable trigger threshold, the restraint force-reducing ignition stage of the occupant restraint system is activated. In the example shown in Figure 1, this occurs for the light crash LC after 37 ms when the end position ELC of the dynamic trigger threshold for the light crash LC is exceeded, and for the severe crash SC, due to the multiple increases in the trigger threshold, only after 55 ms when the end position ESC of the dynamic trigger threshold for the severe crash SC is exceeded, and thus significantly later than in the light crash LC.

[0052] The timing of the crash severity parameter increase is irrelevant within the respective observation period, i.e., within the respective observation duration TB, thus allowing for greater tolerance of pulse variation. This variant of Method 1 enables simple application with, for example, only one switching stage as a "switch" for the dynamic trigger threshold, with only two different trigger threshold levels. Multiple switching stages allow for a smoother transition between early and late triggering of the restraint force-reducing ignition stage and greater ignition time stability in the medium crash severity range. Using non-equidistant levels for switching thresholds SSI, SS2, and / or trigger thresholds allows for a wider spread between ignition times for weaker and stronger load cases and greater application flexibility.

[0053] In another variant of this method 1, a dynamic, in particular decreasing, increase in the trigger threshold is achieved via dynamic threshold steps, as shown by way of example in Figure 2.

[0054] As a measure for the negative slope of the decreasing threshold steps, the average inverted typical slope ST of the upper limit load case, i.e., the severe crash SC, during the observation period is advantageously suitable, which is approximately 13 km / h in 18 ms, or about 200 m / s. 2The level of the upper dynamic trigger threshold is set such that, with the described negative slope of the decreasing threshold step and the point in time when the threshold is exceeded, the desired final velocity position is achieved in a typical crash severity parameter profile of the upper limit load case. Here too, the end positions ELC and ESC of the dynamic trigger thresholds for the limit load cases, i.e., LC for the light crash and SC for the severe crash, are defined in the manner described above for the first variant.

[0055] In the example shown, the observation period TB is again 18 ms. The starting threshold STA is again defined for the start of the observation period TB, in this example again at 3 km / h. The observation period TB begins when the respective course of the crash severity parameter CSP exceeds the starting threshold STA. This exceedance occurs earlier in the case of the severe crash SC than in the case of the mild crash LC.

[0056] The switching thresholds SSI and SS2 are again defined in the manner described above. The relevant range RB of the crash severity parameter CSP curve for the dynamic trigger threshold increase therefore lies between the start threshold STA and the switching threshold SSI or SS2.

[0057] Here too, the starting level at the end position ELC for the dynamic trigger threshold for the light crash SC, the optional first increased level EEN and a second increased level ZEN are specified.

[0058] In this variant, at the beginning of the observation period TB, a level decrease begins from the first elevated level with the inverted typical slope ST. If the first switching threshold SSI is exceeded by the course of the crash severity parameter CSP within the observation period TB, the trigger threshold jumps to the level value of this level decrease NA and remains at this level, as shown in Figure 2.

[0059] Simultaneously, at the beginning of the observation period TB, a level drop begins from the second elevated level with the inverted typical slope ST. If the second switching threshold SSI is exceeded by the course of the crash severity parameter CSP within the observation period TB, the trigger threshold jumps to the level value of this level drop and remains at this level, as shown in Figure 2.

[0060] If the respective crash severity parameter CSP exceeds the applicable trigger threshold, the restraint force-reducing ignition stage of the occupant restraint system is activated. In the example shown in Figure 2, this occurs for the light crash LC after 37 ms when the end position ELC of the dynamic trigger threshold for the light crash LC is exceeded, and for the severe crash SC, due to the multiple increases in the trigger threshold, only after 55 ms when the end position ESC of the dynamic trigger threshold for the severe crash SC is exceeded, and thus significantly later than in the light crash LC.

[0061] In this variant, the threshold level is weighted according to the steepness of the crash severity parameter (CSP). The threshold levels are not constant over time, but decrease linearly or according to a defined rule until the next threshold is exceeded. Therefore, the faster a threshold is exceeded, i.e., the steeper the crash severity parameter curve, the greater the threshold level.

[0062] By using more switching thresholds (SSI, SS2) at smaller intervals, this variant allows for a continuous increase in the dynamic trigger threshold, which is proportional to the slope of the crash severity parameter signal, and thus proportional to the crash severity during the observation period. This variant enables a greater dependence of the trigger threshold increase on the crash severity and a wider spread between ignition points for weaker and stronger load cases.

[0063] The embodiment of the method according to method 2 is described in more detail below.

[0064] In one variant of this method 2, an inversely delayed triggering takes place, as shown by way of example in Figure 3.

[0065] To map the different initial slopes of the typical crash severity parameter curves of the two limit load cases, i.e., the light crash (LC) and the severe crash (SC), inversely to the desired ignition times of the restraint force-reducing occupant restraint stages during the observation period—typically around 55 ms for the upper limit load case (i.e., the severe crash SC) and around 35 ms for the lower limit load case (i.e., the light crash LC)—this method 2 and this variant first define a start threshold (STA) for a crash severity parameter timer and a fixed trigger timer, and a stop threshold (STO) for the crash severity parameter timer. The corresponding typical times of the respective exceedances are then determined. The start threshold (STA) is the beginning of the observation period, and the stop threshold (STO) is the end of the observation period.

[0066] For the reasons already mentioned, the starting threshold STA is advantageously chosen again at approximately 3 km / h, and the stopping threshold STO is the crash parameter value reached at the ignition point of the lower limit load case, i.e., the light crash LC, typically 13 km / h. The relevant range for increasing the threshold is between these two values.

[0067] This results in the relative ignition times related to the respective start of the observation period.

[0068] Th = 55ms - 8ms = 47ms (1) or

[0069] Ti = 35ms - 14ms = 21ms (2)

[0070] Using the typical crash severity parameter timer values ​​determined from the respective time differences for the upper limit load case, i.e. for the severe crash SC, from

[0071] Ath = approx. 22ms - approx. 8ms = approx. 14ms (3) and for the lower limit load case, i.e. for the light crash LC, of

[0072] Ati = Ti = approx. 21ms (4) the respective ignition time T is now determined via the following calculation relationship:

[0073] T = C - F -A t (5) The strain factor F is calculated from the difference in the release time between the upper and lower limit load cases:

[0074] AT = T h -Ti = -F - (A th -A ti) or F = (T h - Ti) / (A ti - A t h ), typically 26 / 7 = 3.7 (6)

[0075] The trigger timer C is a predefined time constant corresponding to the latest possible ignition time in the event of both thresholds STA and STO being exceeded immediately in the same millisecond. The strain factor F stretches the measured crash severity parameter timer value so that the desired ignition time spread between the lower and upper limit load cases is achieved. It is chosen such that ignition occurs immediately after the end of the observation period in the case of the lower limit load case. Therefore, At = At|:

[0076] Ti = A ti= C - F -A ti or C = A ti • (1 + F), typically approx. 21ms • 4.7 = 100ms (7)

[0077] In this variant of Method 2, the crash severity parameter timer A t and the trigger timer C start after the start threshold STA is exceeded. The crash severity parameter timer stops after the stop threshold STO is exceeded. A multiple F of this timer value is subtracted from the trigger timer and yields the ignition time T according to the formula (5) above:

[0078] T = C - F -A t (5)

[0079] After the maximum observation period (C / (F+1)) has elapsed without exceeding the STO stop threshold, triggering may occur immediately or not at all.

[0080] This is a very easy-to-implement methodology with a continuous relationship between threshold exceedance and trigger time.

[0081] In another variant of this method 2, an inverse dynamically delayed triggering with continuous slope prediction is used, as shown by way of example in Figure 4.

[0082] In this particularly advantageous variant of method 2, the spread of the ignition timings for the lower and upper limit load cases, i.e. for the light crash LC and the heavy crash SC, can be increased even further, and at the same time the crash severity parameter pulse profile can be observed for an even longer period in order to become more robust against random variations of the crash severity parameter profile in the first milliseconds of the observation period.

[0083] Here again, the start threshold STA and the stop threshold STO are predefined, specifically in the manner described above. The start threshold STA is set at 3 km / h and the stop threshold at 15 km / h. The start threshold STA is used for the crash severity parameter timer, the fixed trigger timer (for example, 109 ms with a factor of 3.75), and for determining the gradient. It is the same as the start of the observation period.

[0084] First, the crash severity parameter slope S is determined during the initial observation period between the start threshold STA and the stop threshold STO. If the crash severity parameter slope S remains constant at this value, ignition should occur at the desired time, typically 61 ms.

[0085] The present variant is based on the desire that increases in slope, i.e., an unexpected increase in crash severity during the crash, should still be able to lead to a later ignition time after the first observation period, and conversely, unexpectedly flat crash severity parameter curves after the first observation period should be able to advance the ignition time in accordance with the associated lower crash severity. For this purpose, the actual crash severity parameter value is compared at fixed time intervals Tj with the predicted value, which is calculated according to the simple formula.

[0086] CSP (T) = CSP (T — Ti) + S • Ti (8) with

[0087] S = (CSP(Tstop) - CSP(T start)) / (Tstop Tstart ) 0) typical S = (15 km / h - 3 km / h) / (23ms - 8ms) ~ 722. m / s 2(10) can be calculated for the upper limit load case. For fixed start thresholds STA and stop thresholds STO, the corresponding slope values ​​can also be provided via look-up tables to avoid complex calculations during the crash. In the same way, the crash severity parameter prediction values ​​for fixed Ti can be read as tabular values; a typical time interval Ti of approximately 7 ms has proven advantageous here.

[0088] A prediction threshold is defined at the level of the predicted crash severity parameter value, and the system analyzes when the crash severity parameter curve will exceed this threshold. If this threshold is exceeded before the current trigger time, the time difference between Tj and the time of the threshold being exceeded is added to the current trigger time. This value can be positive, meaning the crash severity parameter curve is steeper than predicted for this time interval, or negative, meaning the crash severity parameter curve is flatter than predicted, thus lengthening or shortening the trigger time, respectively. The new prediction interval always starts at the time the actual crash severity parameter curve exceeds the threshold. If the trigger time is reached before the crash severity parameter curve has exceeded the next prediction threshold, the corresponding restraint force-reducing immobilization stage is triggered.

[0089] In advantageous further configurations, the slope value S determined in the first observation period can be increased by a factor to better adapt the forecast curve to the typically increasing slope during a crash against a rigid barrier. A factor of 1.09 has proven particularly advantageous here. For other crash configurations, a dynamic increase of the slope for each new forecast interval may also be useful.

[0090] In this variant of Method 2, as exemplified in Figure 4, gradient predictions are continuously calculated based on the first longest possible gradient evaluation and regularly compared with the actual crash severity parameter curve. If the crash severity parameter curve exceeds the new threshold earlier than predicted, the ignition time is delayed by this value or a multiple thereof. Otherwise, the time difference until the threshold is exceeded is subtracted from the crash severity parameter timer. In the example shown, a starting value SW is 61 ms, and a value VW shifted due to a change caused by the gradient prediction is 66 ms.

[0091] Even in this variant of method 2, the slope of the crash severity parameter curve can be more accurately represented for shorter sections and used as a reference for the ignition time shift.

[0092] In this variant of Method 2, the crash severity parameter profile after exceeding the STO threshold also contributes to the ignition timing determination. A greater ignition timing spread between minor and major crashes is possible through a different weighting of the timer contribution below and above the STO threshold and / or a constant increase, for example 9%, or a continuous increase of the predicted slope values ​​compared to the value during the observation period.

[0093] The embodiment of the method according to method 3 is described in more detail below.

[0094] In one variant of this method 3, a dynamic threshold reduction with a predefined profile takes place, as shown by way of example in Figure 5.

[0095] In this further advantageous methodology for ignition timing adjustment lies the trigger threshold CSP. AInitially not at the lowest possible level, but at the highest possible level CSP m ax, in order to address the worst-case scenario, i.e., a high-speed crash, during the initial crash assessment phase. The CSP trigger threshold A The crash severity parameter curve is now lowered according to a predefined function, for example, based on the progression of the crash, over time. The lowering begins as soon as the crash severity parameter curve exceeds the start threshold STA and stops as soon as the crash severity parameter curve exceeds a fixation threshold CSPmin. The start threshold STA marks the beginning of the observation period. The fixation threshold CSPmin marks the end of the observation period.

[0096] The simplest form of a settling curve is conceivable as a linear progression that starts at the highest possible level CSPmax as soon as the typical upper limit load case, i.e. the severe crash SC, has exceeded the start threshold STA, and ends at the fixing threshold CSPmin when the typical lower limit load case, i.e. the light crash LC, exceeds the fixing threshold CSPmin.

[0097] However, to achieve a certain degree of stability around the ignition time ranges of the two design-relevant limit load cases, it is desirable to represent the flattest possible shape of the settling function in these areas. An S-shaped function, as shown in Figure 5, has therefore proven particularly advantageous; this can be generated, for example, using an inverted sigmoid function.

[0098] CSP A (t) = CSP min + (CSP m ax - CSP min ) / (l+e (t t b / 2) ■ 10 / tb) (11) with tb: desired width of the lowering function can be calculated and stored as a look-up table. For typical limit load cases, the typical values ​​already described have proven advantageous:

[0099] Starting threshold ~ 3km / h (12)

[0100] Fixing threshold ~ 13km / h (13)

[0101] CSPmin ~ 13km / h (14)

[0102] CSPmax = 67 km / h (15) t b ~ 27ms (16)

[0103] For typical extreme load cases, it has proven advantageous to choose the minimum fixing threshold CSPmin to be identical to the lower level of the trip threshold. In specific cases, however, it may also be useful to separate the two thresholds. Experience has also shown that it is advantageous to define the width of the lowering function CSP as appropriate. Ato adapt to the time difference between the typically advantageous trigger time for the lower limit load case of approximately 35 ms and the typical time at which the upper limit load case exceeds the start threshold STA, i.e., approximately 8 ms. In the example shown in Figure 5, the reduction curve for the severe crash SC starts at the highest possible level CSP. m ax, as soon as the course of the severe crash SC has exceeded the start threshold STA. The final position ESC of the dynamic trigger threshold for the severe crash SC is determined on this lowering curve as soon as the course of the severe crash SC exceeds the

[0104] Fixing threshold CSP m The occupant restraint system's force-reducing ignition stage is triggered in a severe crash (SC) when the crash severity parameter (CSP) exceeds the dynamic trigger threshold (ESC) of the SC.

[0105] In the example shown in Figure 5, the lowering curve for the light crash LC starts at the highest possible level CSP. max , as soon as the course of the minor crash LC has exceeded the start threshold STA. The final position ELC of the dynamic trigger threshold for the minor crash LC results on this downward curve as soon as the course of the minor crash LC exceeds the

[0106] The fixing threshold CSPmin has been exceeded. The ignition of the restraint force-reducing ignition stage of the occupant restraint system occurs in a light crash (LC) when the course of the crash severity parameter CSP exceeds this end position ESC of the dynamic trigger threshold for the light crash (LC).

[0107] In this variant of Method 3, instead of fixed switching points, crash-pulse-adapted switching points between early and late ignition of the restraint force-reducing ignition stages of the occupant restraint system are provided. Load cases with higher requirements for ignition time stability can be addressed via a flatter curve shape. Load cases without special requirements can fall into a crash phase with a steeper curve shape. The ignition threshold starts at a high level, i.e., initially based on the severe crash SC, and is only reduced to a lower level when the crash severity parameter CSP remains below the fixed threshold CSP for a relatively long observation period. m in remains

[0108] In one variant of this method 3, a dynamic threshold reduction with a crash pulse-dependent course takes place, as shown by way of example in Figures 6 and 7.

[0109] To achieve an even greater spread in the trigger times, i.e., the ignition times of the restraint force-reducing ignition stage of the occupant restraint system, between the two limit load cases, i.e., between the severe crash SC and the light crash LC, the slope of the crash severity parameter curve after exceeding the start threshold STA is weighted even more heavily by decomposing the crash severity parameter curve into a sequence of slope triangles and assigning the respective crash severity parameter value to the dynamic trigger threshold CSP. A It is added if it exceeds a certain limit G:

[0110] CSP A (t x ) = fs (tx) + CSP(tx) for CSP(tx) > G and fs = lowering function (17) for

[0111] CSP(tx) > G (18) and f s = Lowering function, for example according to formula (11).

[0112] A gradient of approximately 200 m / s has been used to determine the limit value. 2This has proven advantageous for differentiating typical crash load cases with high crash severity, as described above. With a typical evaluation duration ts per gradient interval of approximately 2 ms, this results in a typical limit value G of approximately 1.4 km / h.

[0113] As an advantageous variant that establishes a continuous relationship between the crash severity parameter curve and the trigger threshold, while simultaneously giving greater weight to steeper slopes, squaring the respective crash severity parameter slope values ​​before adding them to determine the dynamic trigger threshold is suitable. However, since this would also broaden the trigger curve for lighter crash load cases and lead to correspondingly later ignition times than desired for these load cases, the reduction function must be narrowed accordingly by shortening tb.

[0114] However, a significantly less complex and advantageous approach involves determining a reference gradient S from the slope of the typical lower limit load case during the observation period, using the advantageous threshold positions already described. For the typical values ​​already described for the starting threshold STA of 3 km / h and the stopping threshold STO of 13 km / h, and the corresponding threshold crossing times for this load case of 14 ms and 35 ms respectively, a typical reference gradient value S of approximately 130 m / s is obtained. 2 . From the respective crash severity parameter values ​​from the determined slope triangles, this crash severity parameter reference value S • Ats of typically 1.0 km / h is subtracted and the result, which can also be negative, is added to the lowering function fs:

[0115] CSP A (t x ) = fs (t x ) + I (CSP (t x ) - S • A t s ) (19)

[0116] To achieve an even greater spread of the trigger times for both limit load cases, categorized factors can be defined for typical crash severity parameter values, which also increase with increasing crash severity parameter value and which correspond to the calculated crash severity parameter difference according to

[0117] CSP A (t x ) = fs (tx) + I ((CSP (t x ) - S • A t s ) • F) (20) are multiplied. As advantageous for A t s For example, the following factors have proven to be relevant for four typical crash severity parameter categories within a timeframe of 2ms:

[0118] F = 0.25 for (CSP (tx) - S - A t s )) < 0.5km / h (21)

[0119] F = 0.5 for 0.5 < (CSP (t x ) - S • A t s )) < lkm / h (22)

[0120] F = 1 for 1 < (CSP (t x ) - S - A ts)) < 1.5 km / h (23)

[0121] F = 2 for (CSP (t x ) - S • A t s )) > 1.5 km / h (24)

[0122] This approach decomposes a steeper, predefined curve into slope triangles (look-up table). The summed crash severity parameter value is subtracted from the respective y-value, and the dynamic threshold is lowered only by the difference between the two values. By incorporating the crash pulse curve into the threshold calculation, the dynamic threshold remains in the upper range for even longer during severe crashes and is simultaneously lowered even more rapidly for weak crash pulses. This allows for a greater spread between ignition timings for weaker and stronger load cases.

[0123] As shown in Figure 7, bottom right, the measured crash severity parameter signal is decomposed into slope triangles, and the respective y-value (i.e., the value on the y-axis of the diagram), and thus the crash severity parameter, is extracted. The respective y-value is only added if it is greater than the limit value. This is schematically illustrated in Figure 7, right, in the middle section using checkmarks and crosses. A checkmark indicates that the addition is performed, and a cross indicates that no addition is performed.

[0124] As shown in Figure 7 at the top right, the addition of the selected large slopes on the lowering curve results in a weakened dynamic threshold lowering for particularly steep crash severity parameter curves, shown in the top right curve in Figure 7.

[0125] Reference symbol list

[0126] CP crash progressive parameter

[0127] CSP Crash Severity Parameters

[0128] CSPmax highest possible level

[0129] CSPmin fixation threshold

[0130] EEN first elevated level

[0131] ELC end position trigger threshold minor crash

[0132] ESC end position trigger threshold severe crash

[0133] LC minor crash

[0134] NA level drop

[0135] RB relevant area

[0136] SC heavy crash

[0137] 551 first switching threshold

[0138] 552 second switching threshold

[0139] ST typical gradient

[0140] STA Starting Threshold

[0141] STO stop threshold

[0142] SW starting value

[0143] TB observation period

[0144] VW shifted value

[0145] ZEN second elevated level

Claims

Patent claims 1. A method for controlling an occupant restraint system in a vehicle, wherein at least one impact sensor provides at least one impact signal, a trigger criterion derived from the impact signal is compared with a trigger threshold, and an ignition stage of the occupant restraint system is triggered immediately or after a predetermined time interval when the trigger threshold is exceeded by the trigger criterion, wherein the trigger threshold and / or the time interval is increased or decreased depending on the course of a crash severity parameter (CSP) derived from the impact signal or from another impact signal of at least one other impact sensor and representing a crash severity, wherein the trigger threshold and / or the time interval is increased or decreased depending on the course of the crash severity parameter (CSP) such that the ignition stage is triggered later the greater the crash severity, characterized in thatthat a restraint force-reducing ignition stage of the occupant restraint system is triggered as the ignition stage.

2. Method according to claim 1, characterized in that the ignition stage activates a switchable belt force limiter of a safety belt and / or an actively actuated venting device of an airbag.

3. Method according to one of the preceding claims, characterized in that the trigger criterion itself is used as the crash severity parameter (CSP).

4. Method according to one of the preceding claims, characterized in that the crash severity parameter (CSP) is determined by a loss of speed of the vehicle derived from the impact signal from the start of the crash.

5. Method according to one of the preceding claims, characterized in that the crash severity parameter (CSP) is determined by an intrusion velocity derived from the impact signals.

6. Method according to one of the preceding claims, characterized in that the crash severity parameter (CSP) is determined from the difference in speed loss between the front of the vehicle and the center of the vehicle and / or the calculated or measured occupant forward displacement and / or the reciprocal time difference of threshold exceedances of the trigger criterion.

7. Method according to one of the preceding claims, characterized in that the trigger threshold is dynamically raised depending on the crash severity parameter (CSP).

8. Method according to claim 7, characterized in that the lifting is carried out via static threshold stairs.

9. Method according to claim 7, characterized in that the lifting is carried out via decreasing threshold steps, the predetermined temporal decrease of which is started and ended depending on the course of the crash severity parameter (CSP).

10. Method according to one of the preceding claims, characterized in that The trigger time is influenced by a time element that is proportional to the difference of a crash-progressive parameter (e.g., time) between two threshold exceedances of the crash severity parameter (CSP).

11. Method according to one of the preceding claims, characterized in that an initially high trigger threshold is reduced as long as the Crash severity parameter (CSP) is below a certain threshold.

12. Method according to one of the preceding claims, characterized in that an initially low trigger threshold is raised by a crash severity parameter (CSP) that increases strongly or rapidly during the course of the crash and / or an initially high trigger threshold is lowered by a crash severity parameter (CSP) that increases weakly or slowly during the course of the crash.

13. Method according to one of the preceding claims, characterized in that the trigger threshold is subject to a predetermined temporal change, which is started and ended depending on the course of the crash severity parameter (CSP).

14. Method according to claim 13, characterized in that the change over time is influenced by the course of the crash severity parameter (CSP).

Images