DE102005023183B4 - Method and device for rollover detection of a vehicle - Google Patents

Abstract

A method for the rollover detection of a vehicle, in which at least one vehicle dynamic variable (ωx, ωz, ax, ay, az, vx, β) is determined and evaluated, wherein based on the at least one determined driving dynamic variable (ωx, ωz, ax, ay, az, vx, β, ri) a current vehicle state is determined and based on a simplified physical motion model, a temporal extrapolation of the current vehicle state with respect to a yawing motion and / or a rolling motion is performed, wherein a rollover-critical state is detected when the time extrapolated yawing and / or rolling motion with a predeterminable probability exceedable predetermined critical thresholds, characterized in that for the rollover detection in a lower roll angle range the at least one of a comfort system and / or vehicle dynamics control system and / or assistance system detected driving dynamic variable (ωx, ωz, ax, ay, az, vx , β, ri) and / or used by the comfort system and / or vehicle dynamics control system and / or assistance system evaluated and / or preprocessed data for determining the current vehicle state as a starting point for the temporal extrapolation.

Description

State of the art

The invention relates to a method for rollover detection of a vehicle according to the preamble of independent claim 1 and of an associated device for rollover detection.

Figures from the US confirm the importance of passive safety in vehicle rollovers. In 1998, half of all fatal vehicle accidents were due to a rollover. In the entire accident scene, the vehicle rollover takes a share of about 20 percent.

Against this background, methods and concepts for rollover detection are developed which detect vehicle rollovers early on to ensure that occupant protection devices, such as seatbelt pretensioners, window airbags, curtain airbags, roll bars, etc., are activated in good time and thus the risk of injury to vehicle occupants reduced.

• (a) Messen eines Satzes von Eingangssignalen, der mindestens Fahrzeuggeschwindigkeit, Fahrzeuglenkwinkel, Fahrzeugquerbeschleunigung und Fahrzeugwankrate aufweist,
• (b) Integrieren der Fahrzeugwankrate, um das Fahrzeugwankwinkelinkrement zu erhalten,
• (c) Bestimmen des Fahrzeugzustands auf der Basis der Eingangssignale,
• (d) Bestimmen der geschätzten Fahrzeugquerbeschleunigung, die dem echten Fahrzeugwankwinkel entspricht, auf der Basis von mindestens dem Fahrzeugzustand, der Fahrzeugquerbeschleunigung und der Zentrifugalbeschleunigung,
• (e) Bestimmen des geschätzten Fahrzeugwankwinkels auf der Basis von mindestens dem Fahrzeugwankwinkelinkrement, der geschätzten Fahrzeugquerbeschleunigung und dem Fahrzeugzustand und
• (f) Erzeugen eines Ausgangsaktivierungssignals, das die Möglichkeit eines Überschlags des Fahrzeugs bestimmt, als Funktion von mindestens: dem geschätzten Fahrzeugwankwinkel und der Fahrzeugwankrate.

From the DE 60 2004 007 270 T2 is known, a method for detecting a vehicle rollover, the steps comprising:

• (a) measuring a set of input signals having at least vehicle speed, vehicle steering angle, vehicle lateral acceleration and vehicle roll rate,
• (b) integrating the vehicle roll rate to obtain the vehicle roll angle increment;
• (c) determining the vehicle condition based on the input signals,
• (d) determining the estimated vehicle lateral acceleration corresponding to the true vehicle roll angle based on at least the vehicle state, the vehicle lateral acceleration, and the centrifugal acceleration;
• (e) determining the estimated vehicle roll angle based on at least the vehicle roll angle increment, the estimated vehicle lateral acceleration and the vehicle condition, and
• (f) generating an output enable signal that determines the possibility of vehicle rollover as a function of at least: the estimated vehicle roll angle and the vehicle roll rate.

From the DE 10 2004 029 064 B3 It is known that a rotational rate (ω) about a longitudinal axis of the vehicle and an inclination (θ) of the vehicle in the lateral direction are determined as state variables of a vehicle. In this case, if the state variables in the state area spanned by them exceed a triggering threshold characteristic curve (ZHD), a rollover situation is detected.

Known in production vehicles built systems for rollover detection usually evaluate a rolling motion and accelerations in the x-, y- and z-direction of the vehicle. On this basis, in principle, a secure detection of a vehicle rollover is possible.

Modern vehicles are increasingly being equipped with electronic stabilization systems (ESP). These systems have a significant influence on the driving dynamics in the lower roll angle range of <10 °. The impact of these systems on vehicle dynamics makes it very difficult to predict a rollover event. In addition, in this lower roll angle range influencing factors, such. As tire size, tire pressure, vehicle weight, adhesion coefficients and soil properties, very important for the rollover detection, but can not be reliably detected in their entirety by sensors. On the other hand, for the optimal use of time-critical occupant protection products, such. B. the window airbag and / or the head airbag, a rollover detection already at roll angles below 10 ° may be possible.

Advantages of the invention

The inventive method for rollover detection of a vehicle having the features of independent claim 1 and the device according to the invention for rollover detection with the features of independent claim 11 have the advantage that by temporal extrapolation of a current vehicle state, which based on at least one determined and evaluated driving dynamics Size is determined, with respect to a yawing motion and / or a rolling motion in the required roll angle range, in particular in a lower roll angle range <20 °, a rollover detection is possible. For temporal extrapolation, a simplified physical motion model is used, wherein an override-critical state is advantageously recognized if the time-extrapolated yawing motion and / or roll movement exceeds predetermined critical threshold values with a prescribable probability. This ensures that safety devices such as belt tensioners, head airbags, window airbags, roll bars, etc. are activated in good time and thus reduces the risk of injury to the occupants.

By the measures and refinements recited in the dependent claims Advantageous improvements of the specified in the independent claim 1 method for rollover detection of a vehicle and specified in the independent claim 10 device for rollover detection of a vehicle possible.

It is particularly advantageous that a plausibility of the rollover-critical state is determined by a plausibility function based on the at least one determined driving dynamic variable, wherein the occupant protection means are activated only if the plausibility function determines that a detected rollover-critical state is plausible.

The simplified physical motion model is based for example on a Newton equation of motion according to equation (1): f (ω _x , ω _z , a _x , a _y , a _z , β, v _x , r _i ) = 0 (1) where ω _{x is} a roll rate, ω _{z is} a yaw rate, a _{x is} an acceleration in the direction of the vehicle longitudinal axis, a _{y is} an acceleration in the lateral vehicle direction, a _{z is} an acceleration in the vertical vehicle direction, β is a current slip angle of the vehicle, v _{x is} a vehicle speed along the vehicle's longitudinal axis and r _{i represent} characteristic fulcrums of the vehicle in space.

In addition, for the temporal extrapolation of the future yawing motion and / or roll motion at an extrapolation time t starting from the current time t _0, model functions for the acceleration in the direction of the vehicle longitudinal axis a _x (t ₀ , t) and / or for the acceleration in the lateral vehicle direction a _y (t ₀ , t) and / or for the acceleration in the vertical vehicle direction a _z (t ₀ , t) are calculated. The model functions may advantageously be based on physical ground-wheel models well-known from the literature.

For example, the accelerations may be calculated using the model functions according to equations (2), (3) and (4): a _x (t ₀ , t) = a _x (t, ω _x (t ₀ ), ω _z (t ₀ ), a _x (t ₀ ), a _y (t ₀ ), a _z (t ₀ ), v _x (t ₀ ), β (t ₀ )) (2) a _y (t ₀ , t) = a _y (t, ω _x (t ₀ ), ω _z (t ₀ ), a _x (t ₀ ), a _y (t ₀ ), a _z (t ₀ ), v _x (t ₀ ), β (t ₀ )) (3) a _z (t ₀ , t) = a _z (t, ω _x (t ₀ ), ω _z (t ₀ ), a _x (t ₀ ), a _y (t ₀ ), a _z (t ₀ ), v _x (t ₀ ), β (t ₀ )) (4) where t _{0 represents} the current time and applies to the extrapolation time t: t> t ₀ .

Advantageously, the temporal extrapolation of the vehicle movement takes place over a defined period Δt in the range from 200 ms to 2000 ms, depending on the data quality and current vehicle condition. Depending on the evaluation and control unit used, a certain computing time is required for extrapolation. If the temporal extrapolation over a given period .DELTA.t calculated, then with the help of updated sensor data or updated driving dynamics variables, the model functions, adjusted accordingly and carried out a new temporal extrapolation of the vehicle movement.

Furthermore, it is advantageous for the rollover detection in a lower roll angle range of a sensor unit and / or a comfort system and / or a vehicle dynamics control system and / or an assistance system detected driving dynamics variables and / or by the sensor unit and / or the comfort system and / or the vehicle dynamics control system and / or the assistance system evaluated and / or preprocessed data for determining the current vehicle state can be used as a starting point for the temporal extrapolation. The sizes and / or data of the sensor unit and / or the comfort system and / or the vehicle dynamics control system and / or the assistance system are provided, for example, via a bus interface.

In the event of a failure of the data transmission from the sensor unit and / or the comfort system and / or the vehicle dynamics control system and / or the assistance system, the variables provided by the sensor unit and / or the comfort system and / or the vehicle dynamics control system and / or the assistance system are advantageously used and / or data is replaced by predefined values, which although the rollover detection is limited in the lower roll angle range, but rollover detection at higher roll angles is ensured.

The inventive apparatus for rollover detection of a vehicle advantageously comprises an evaluation and control unit, which determines based on the at least one determined dynamic vehicle size a current vehicle state and based on a simplified physical motion model temporal extrapolation of the current vehicle state with respect to a yaw and / or a Rolling motion performs. The evaluation and control unit recognizes a condition that is critical in terms of a rollover when the time-extrapolated yawing motion and / or roll movement exceed a preselectable critical threshold values.

The at least one vehicle-dynamic variable is directly detected, for example, by a sensor unit and / or a comfort system and / or a vehicle dynamics control system and / or an assistance system and / or derived and / or determined by evaluating acquired vehicle dynamics variables.

drawing

Embodiments of the invention are illustrated in the drawings and are explained in more detail in the following description.

Show it

1 a schematic block diagram of an apparatus for performing a method for rollover detection of a vehicle, and

2 a schematic representation of a current vehicle state and an extrapolated vehicle state.

description

How out 1 a device for carrying out a method according to the invention for detecting the rollover of a vehicle comprises an evaluation and control unit 10 which for generating a rollover detection signal driving dynamics variables such as a roll rate ω _x , a yaw rate ω _z , an acceleration in the direction of the vehicle longitudinal axis a _x , a Bescheunigung lateral vehicle a _y , an acceleration in the vertical vehicle direction a _z , a current slip angle β of the vehicle , a vehicle speed v _x along the vehicle longitudinal axis and / or evaluates characteristic pivot points of the vehicle in space r _i . The driving-dynamics variables ω _x , ω _z , a _x , a _y , a _z , v _x , β, r _i can be measured directly via corresponding sensors and / or comfort systems and / or vehicle dynamics control systems and / or assistance systems present in the vehicle measured values are derived or determined. For example, from an electronic stabilization system (ESP system) driving dynamics variables such as the yaw rate ω _z or preprocessed speed data via a bus interface, z. As a CAN interface, be provided for further use.

Based on at least one of the detected driving dynamics variables ω _x , ω _z , a _x , a _y , a _z , v _x , β, r _i determines the evaluation and control unit 10 a current vehicle state and based on a simplified physical motion model performs a temporal extrapolation of the current vehicle state with respect to a yawing motion and / or a rolling motion. The evaluation and control unit 10 recognizes a condition critical to rollover when the time-extrapolated yawing motion and / or roll movement exceeds a critical predetermined thresholds with a predeterminable probability. In addition, the evaluation and control unit executes a plausibility function based on the detected driving-dynamics variables ω _x , ω _z , a _x , a _y , a _z , v _x , β, r _i , which ascertains a plausibility of the condition that is not critical to overturning. If the plausibility function determines that a detected overrun-critical condition is plausible, then the evaluation and control unit generates 10 the rollover detection signal about which occupant protection means, such. As belt tensioner, head airbag, window airbag, roll bar, etc., can be activated.

As a basic complaint for the temporal extrapolation of the current vehicle state with respect to the yawing motion and / or the rolling motion, a corresponding Newtonian equation of motion according to equation (1) and several model functions according to equations (2), (3) and (4) for the accelerations used in two or three spatial directions. The model functions are based on physical ground-wheel models, as they are well known in the literature. f (ω _x , ω _z , a _x , a _y , a _z , β, v _x , r _i ) = 0 (1) where ω _x the roll rate, ω _z , the yaw rate, a _x the acceleration in the direction of the vehicle axis, a _y the acceleration in the lateral direction, a _z the acceleration in the vertical direction, β the current slip angle of the vehicle, v _x the vehicle speed along the Vehicle axis and r _{i represent} the characteristic pivot points in space. a _x (t ₀ , t) = a _x (t, ω _x (t ₀ ), ω _z (t ₀ ), a _x (t ₀ ), a _y (t ₀ ), a _z (t ₀ ), v _x (t ₀ ), β (t ₀ )) (2) a _y (t ₀ , t) = a _y (t, ω _x (t ₀ ), ω _z (t ₀ ), a _x (t ₀ ), a _y (t ₀ ), a _z (t ₀ ), v _x (t ₀ ), β (t ₀ )) (3) a _z (t ₀ , t) = a _z (t, ω _x (t ₀ ), ω _z (t ₀ ), a _x (t ₀ ), a _y (t ₀ ), a _z (t ₀ ), v _x (t ₀ ), β (t ₀ )) (4) where t _{0 is} the current time and applies to the extrapolation time t: t> t ₀ .

The temporal extrapolation of the vehicle movement takes place, depending on the data quality and the current vehicle state, over a defined period Δt, which is generally between 200 ms to 2000 ms. Depending on the design of the evaluation and control unit 10 a certain computing time is needed for extrapolation. If a temporal extrapolation is calculated over a given period of time Δt, the model functions a _x (t ₀ , t), a _y (t ₀ , t) and a _z (t ₀ , t) are adapted accordingly using the updated sensor data, and a carried out a new time extrapolation of the vehicle movement. The extrapolation cycles can be performed until deactivated, for example by the plausibility check function. The data of all sensors or the preprocessed data and quantities serve as initial values for the temporal extrapolation. In the case of partial or complete failure of the data transfer from the ESP system or from a system which provides the required additional information, the missing data can be replaced by predefined values. Although this will limit rollover detection in the lower roll angle range <20 °, rollover detection at higher roll angles can be ensured. The performance of conventional methods for rollover detection is achieved by the inventive method even if the additional information is not available.

2 shows a schematic representation of a current vehicle state of a vehicle 100 at a current time t ₀ = 4.04 s. At the current time t ₀ = 4.04 s, the roll angle is 6.8 °. The vehicle 100 ' represents a temporally extrapolated vehicle state determined according to the method according to the invention for rollover detection at a time t = 4.69 s. The roll angle will therefore be approximately 52 ° at time t = 4.69 s.

The method according to the invention and the device according to the invention make it possible, in particular by means of the use of driving dynamics variables, which are detected and / or evaluated and / or preprocessed by comfort systems and / or vehicle dynamics control systems and / or assistance systems present in the vehicle, to provide cost-effective rollover detection of a vehicle, preferably in one lower roll angle range from 0 to 20 °.

Claims (11)

A method for rollover detection of a vehicle, in which at least one vehicle dynamics variable (ω _x , ω _z , a _x , a _y , a _z , v _x , β) is determined and evaluated, wherein based on the at least one determined driving dynamic variable (ω _x , ω _z, a _x, a _y, a _z, v _x, β, r _i), a current vehicle state is determined, and based on a simplified physical motion model a temporal extrapolation of the current vehicle condition in terms of a yaw and / or roll motion is carried out, wherein an over-critical condition is detected when the extrapolated time yawing and / or rolling movement with a predetermined probability exceedable predetermined thresholds characterized in that the rollover detection in a lower roll angle range, the at least one detected by a comfort system and / or vehicle dynamics control system and / or assistance system driving dynamics Size (ω _x , ω _z , a _x , a _y , a _z , v _x , β, r _i ) and / or evaluated by the comfort system and / or vehicle dynamics control system and / or assistance system and / or preprocessed data for determining the current vehicle state as a starting point for the temporal extrapolation be used. Method according to Claim 1, characterized in that a plausibility of the rollover-critical state is determined by a plausibility function based on the at least one determined dynamic driving variable (ω _x , ω _z , a _x , a _y , a _z , v _x , β, r _i ) wherein occupant protection means are activated when the plausibility function determines that a detected blow-over critical condition is plausible. Method according to claim 1 or 2, characterized in that the simplified physical motion model is based on a Newton's equation of motion: f (ω _x , ω _z , a _x , a _y , a _z , β, v _x , r _i ) = 0 where ω _{x is} a roll rate, ω _{z is} a yaw rate, a _{x is} an acceleration in the direction of the vehicle longitudinal axis, a _{y is} a lateral vehicle acceleration, a _{z is} a vertical vehicle direction acceleration, β is a current vehicle slip angle, v _{x is} a vehicle speed the vehicle longitudinal axis and r _{i represent} characteristic pivot points of the vehicle in space. A method according to claim 3, characterized in that for the temporal extrapolation of the future yawing and / or rolling motion at an extrapolation time (t) starting from the current time (t ₀ ) model functions for the acceleration in the direction of the vehicle longitudinal axis a _x (t ₀ , t) and / or for the acceleration in the lateral vehicle direction a _y (t ₀ , t) and / or for the acceleration in the vertical vehicle direction a _z (t ₀ , t) are calculated. A method according to claim 4, characterized in that the accelerations according to the model functions a _x (t ₀ , t) = a _x (t, ω _x (t ₀ ), ω _z (t ₀ ), a _x (t ₀ ), a _y (t ₀ ), a _z (t ₀ ), v _x (t ₀ ), β (t ₀ )) and / or a _y (t ₀ , t) = a _y (t, ω _x (t ₀ ), ω _z (t ₀ ), a _x (t ₀ ), a _y (t ₀ ), a _z (t ₀ ), v _x (t ₀ ), β (t ₀ )) and / or a _z (t ₀ , t) = a _z (t, ω _x (t ₀ ), ω _z (t ₀ ), a _x (t ₀ ), a _y (t ₀ ), a _z (t ₀ ), v _x (t ₀ ), β (t ₀ )) be calculated. Method according to one of claims 1 to 5, characterized in that the temporal extrapolation over a defined period of time (Δt) in the range of 200 ms to 2000 ms is performed. A method according to claim 1, characterized in that the lower roll angle range includes roll angle in the range of 0 to 20 °. A method according to claim 1 or 7, characterized in that the sizes and / or data of the comfort system and / or vehicle dynamics control system and / or assistance system are provided via a bus interface for further processing. A method according to claim 8, characterized in that in case of failure of the data transmission from the comfort system and / or vehicle dynamics control system and / or assistance system, the sizes and / or data of the comfort system and / or vehicle dynamics control system and / or assistance system are replaced by predefined values, which a rollover detection at higher roll angles. Device for detecting the rollover of a vehicle, which determines and evaluates at least one vehicle dynamic variable (ω _x , ω _z , a _x , a _y , a _z , v _x , β), wherein an evaluation and control unit ( 10 ), which determines a current vehicle state based on the at least one determined dynamic vehicle variable (ω _x , ω _z , a _x , a _y , a _z , v _x , β, r _i ) and based on a simplified physical motion model a temporal extrapolation of actual vehicle state with respect to a yawing motion and / or a rolling motion, wherein the evaluation and control unit ( 10 ) detects a critical condition when the time-extrapolated yawing motion and / or roll movement exceed a predeterminable probability presettable critical threshold values, characterized in that the evaluation and control unit ( 10 ) for rollover detection in a lower roll angle range, the at least one of a comfort system and / or vehicle dynamics control system and / or assistance system detected driving dynamics variable (ω _x , ω _z , a _x , a _y , a _z , v _x , β, r _i ) and / or used by the comfort system and / or vehicle dynamics control system and / or assistance system evaluated and / or preprocessed data to determine the current vehicle state as a starting point for the temporal extrapolation. Apparatus according to claim 10, characterized in that the at least one vehicle dynamic variable (ω _x , ω _z , a _x , a _y , a _z , v _x , β, r _i ) of a sensor unit and / or a comfort system and / or a Vehicle dynamics control system and / or an assistance system detected directly and / or can be derived and evaluated by evaluating acquired dynamic driving variables and / or can be determined.

Images