DE102013211243A1 - Method for determining a vehicle mass - Google Patents

Abstract

The invention relates to a method in which the vehicle mass of a moving motor vehicle is determined, wherein a speed signal, a longitudinal acceleration signal, a brake signal and a drive signal are considered, and a force balance of the longitudinal dynamics of the motor vehicle is evaluated and the longitudinal acceleration signal is measured by an inertial sensor. According to the invention, the force balance of the longitudinal dynamics is evaluated both during acceleration and braking processes, a number of raw mass values being calculated, and the vehicle mass being determined on the basis of a statistical evaluation of the raw mass values, which comprises at least averaging. The invention further relates to an electronic control device for a brake system of a motor vehicle.

Description

The invention relates to a method according to the preamble of claim 1 and an electronic control device according to the preamble of claim 11.

The driving behavior of motor vehicles, in particular of commercially used vehicles for the transport of loads, can be considerably influenced by the loading state or the current vehicle weight. Therefore, various methods for estimating the vehicle mass have already become known, for cost reasons, many approaches use signals already existing for a vehicle dynamics control and / or brake slip control sensors. Considering vehicle longitudinal dynamics, Newton-Euler's law on force balancing in the longitudinal direction of the vehicle is evaluated and the vehicle mass is calculated as the quotient of the difference between the driving and braking forces in the numerator and the resulting acceleration in the denominator.

For example, is from the DE 42 28 413 A1 a method for determining the mass of a motor vehicle moved by propulsive forces in its longitudinal direction, wherein at least two longitudinal accelerations are detected at least two different times and the propulsive forces present at these times, and wherein the vehicle mass is determined by the difference of the propulsive forces by the difference the longitudinal acceleration is divided. Since the vehicle acceleration is determined based on a change over time of the vehicle speed or a time derivative of the reference speed of a brake slip control determined from Raddrehzahlsensorsignalen, there is the problem of an unmeasurable road gradient angle, which are obtained on a roadway with slope wrong values.

In the WO 03/023334 A1 An approach is disclosed in which vehicle mass and pitch angles are estimated in parallel. However, because of the underdetermination of the problem, the robustness of the algorithm or the correctness of the detected mass is questionable.

From the EP 1 858 736 B1 It is known to provide a measure of acceleration with an acceleration sensor to account for the influence of a slope in the vehicle mass estimate. However, the problem of noise of the acceleration signals and the engine torque signals due to the vibrations and vibrations in the drive train, which disturbs the considered relationship between the force and the acceleration occurs. Filtering the signals, for example, with a low-pass filter has the disadvantage of introducing a phase delay, which causes a strong distortion of the calculated mass values in dynamic cases.

The previously known methods for estimating the vehicle mass thus have problems that have been solved in many cases only inadequate. Therefore, the object of the present invention is to provide a robust against disturbing procedures for determining the vehicle mass.

This object is achieved by a method according to claim 1.

Thus, a method is provided in which the vehicle mass of a moving motor vehicle is determined, wherein a speed signal, a longitudinal acceleration signal, a brake signal and a drive signal are considered, wherein a force balance of the longitudinal dynamics of the motor vehicle is evaluated, and wherein the longitudinal acceleration signal measured by an inertial sensor becomes. According to the invention, the force balance of the longitudinal dynamics is evaluated both during acceleration and braking processes, a number of raw mass values being calculated, and the vehicle mass being determined on the basis of a statistical evaluation of the raw mass values, which comprises at least one averaging.

The method according to the invention has the following advantages: Because the longitudinal acceleration signal is measured with an inertial sensor, the mass estimation can be used independently of the road gradient and thus also in mountains. The inertial sensor as an inertial sensor with a movably mounted test mass can be realized in particular in the form of a micro-electro-mechanical system and is therefore inexpensive and compact. Moreover, due to the use of the longitudinal acceleration sensor, there is no impairment of the mass estimation by slip, tire pressure loss, and deformation of the tires by the load. Since a number of gross mass values of the vehicle calculated during a plurality of successive braking and drive operations are statistically evaluated, no falsification of the determined vehicle mass occurs due to the pitching of the vehicle during the individual Acceleration or deceleration phases. Furthermore, the method is robust against single outliers, since the averaging reduces scattering caused by noise of the input signals. The preferably recursive averaging can be carried out with little computational effort. Thus, a robust determination of the vehicle mass or the load of the vehicle is provided.

According to a preferred embodiment of the invention, the longitudinal acceleration signal is corrected by an additive displacement constant, in particular in order to compensate for the influence of an uneven loading on the longitudinal acceleration signal. This compensation of the acceleration sensor offset caused by asymmetric loading has the advantage that the accuracy of a mass estimation is significantly increased.

Particularly preferably, a new value of the additive displacement constant is determined when a stable standstill has been detected on the plane. By resting on the plane is to be understood here that the road has no (or at most a slight) inclination. Stable here means that the vehicle does not wobble. By determining the additive displacement constant only in defined driving or holding situations, the remaining residual shift or the resulting systematic error in the mass estimation can be minimized.

• • Innerhalb eines vorgegebenen Zeitraums war eine Bremsbetätigung erfolgt.
• • Momentan erfolgt keine Bremsbetätigung.
• • Das Geschwindigkeitssignal unterschreitet einen vorgegebenen Halteschwellenwert.
• • Eine zeitliche Änderung des Längsbeschleunigungssignals unterschreitet einen vorgegebenen Schwankungsschwellenwert.

Very particularly preferably, a stable standstill on the plane is recognized when one or more, in particular all of the following conditions are met:

• • Within a given period of time, a brake application had taken place.
• • There is currently no brake application.
• • The speed signal falls below a preset holding threshold.
• • A temporal change of the longitudinal acceleration signal falls below a predetermined fluctuation threshold value.

Thus, a deceleration of the vehicle is detected in the standstill. The fact that only holding situations are evaluated, in which the vehicle stops without permanent brake actuation by the driver or a traction help, a pitch clearance of the road is ensured. By using only constant (or very slowly changing) signals from the inertial sensor for the offset calculation, even after a fast braking the vehicle can not flick the shift constant.

Appropriately, an evaluation of the balance of the longitudinal dynamics only takes place when a stable driving condition is detected. This prevents influences that are difficult to take into account in the vehicle mass estimation, such as B. a impaired by the intervention of a brake slip control speed signal, distort the vehicle mass estimate.

• • Ein Betrag einer Beschleunigungsanforderung des Fahrers liegt in einem vorgegebenen Beschleunigungsintervall, wobei insbesondere entweder eine Gaspedalbetätigung oder eine Bremsbetätigung durch den Fahrer einen vorgegebenen Schwellenwert überschreiten.

It is particularly expedient if the stable driving state is detected when one or more, in particular all, of the following conditions are met:

• • An amount of an acceleration request of the driver is in a predetermined acceleration interval, in particular, either an accelerator operation or a brake operation by the driver exceeds a predetermined threshold.

• • Eine Fahrgeschwindigkeit des fahrenden Kraftfahrzeugs liegt in einem vorgegebenen Geschwindigkeitsintervall.

This ensures that a noticeable drive or braking acceleration of the vehicle takes place.

• • A driving speed of the moving motor vehicle is within a predetermined speed interval.

• • Ein Betrag eines vom Fahrer eingestellten Lenkwinkels und/oder einer gemessene Gierrate liegt unterhalb eines vorgegebenen Lenkschwellenwerts und/oder eines vorgegebenen Gierratenschwellenwerts.
• • Ein Betrag einer gemessenen Querbeschleunigung liegt unterhalb eines vorgegebenen Kurvenschwellenwerts

Here, the interval may have a standstill or a predetermined minimum speed of the vehicle as the lower limit or lower threshold, while an upper limit is suitably between 50 km / h and 100 km / h.

• • An amount of a driver-set steering angle and / or a measured yaw rate is below a predetermined steering threshold and / or yaw rate threshold.
• An amount of measured lateral acceleration is below a predetermined threshold value

• • Eine zeitliche Änderung des Antriebssignals unterschreitet einen vorgegebenen Antriebsschwellenwert.
• • Der Betrag des Längsbeschleunigungssignals überschreitet einen vorgegebenen Mindestschwellenwert.
• • Es ist weder eine Fahrdynamikregelung noch eine Bremsschlupfregelung aktiv.

As a result, driving situations are excluded in which a significant lateral dynamics or steering movements by the driver can distort the vehicle mass estimate.

• • A time change of the drive signal falls below a specified drive threshold value.
• • The amount of the longitudinal acceleration signal exceeds a predetermined minimum threshold.
• • Neither a vehicle dynamics control nor a brake slip control is active.

This ensures that there is a defined and well-evaluated driving situation.

Preferably, raw mass values are calculated continuously during a braking or acceleration process, in particular at predetermined time intervals, average and variance being calculated for the raw mass values of a respective braking or acceleration process. This makes it possible to determine a plurality of raw mass values during a driving operation.

According to a particularly preferred embodiment of the invention, in the statistical evaluation, the vehicle mass is determined by means of an optimization method, in particular a least squares method, from the mean values and variances of a plurality of braking or acceleration operations. An optimization method may adjust one or more parameters of a known functional relationship such that a cost function becomes minimal. In this case, a least squares method known per se can be used, in which the smallest quadratic deviation or the smallest possible sum of the squares of the deviations between measured values and adapted estimated values is aimed for. This stochastic procedure has the advantage that a reliable value for the vehicle mass is determined and, in addition, a quality factor, such as a variance or a confidence interval, for the determined mass can be stated. Preferably, by performing the stochastic estimations recursively, only a limited amount of memory is needed.

It is particularly advantageous that calculated raw mass values of a braking or acceleration process are rejected if they deviate from the mean value of the braking or acceleration process by more than a predetermined multiple of the variance, and thus do not go any further into the determination of the vehicle mass. The factor or the multiple is here as any real value, d. H. not just an integer. Alternatively or additionally, it is also possible to specify a predefined deviation threshold value, by means of which raw mass values lying outside are discarded or at least less weighted. The determined vehicle mass is thus more robust against outliers, since these can be detected on the basis of a histogram comparison. The fact that the raw mass values are not always normally distributed and are distributed bi-modally up to multi-modal (eg due to variable engine power) depending on the type of traction (drive or braking) can also be explained by An optimization procedure is identified by a histogram and treated better.

Advantageously, the speed signal is determined based on at least one wheel speed sensor and / or the brake signal determined by a brake pressure sensor and / or a pedal travel sensor on the brake pedal and / or the drive signal from a reported by an engine control unit of an internal combustion engine and / or an electric drive torque and / or an determined by a motor shaft measured speed.

Preferably, in addition to the balance of forces of the longitudinal dynamics, a balance of forces of the lateral dynamics is considered, wherein a lateral acceleration signal and at least two wheel speed signals are considered and wherein the coordinates of the current center of gravity are determined in a vehicle-fixed coordinate system.

• • Verwendung von unabhängig verfügbarer Sensorinformation
• • Weniger Rauschen bzw. weniger durch Ableitung verursachte Signalverzögerung
• • Unabhängigkeit vom Steigungswinkel der Straße, da der Steigungswinkel γ gemäß a_sensor = v . – g·sin(θ) bereits in Beschleunigungssensorsignal enthalten ist. Die Steigung hat somit keinen direkten Einfluss auf die Kraftbilanzierung. Die Massenschätzung kommt auch robust mit Gebirgen zurecht.

The invention further relates to an electronic control unit, in particular for a brake system of a motor vehicle, having interfaces for the connection of at least one wheel speed sensor, at least one brake actuation sensor, at least one inertial sensor and an interface to a vehicle data bus, which comprises a computing unit which is a method according to at least one of preceding claims. The electronic control unit uses the signal of an inertial sensor as a measured longitudinal acceleration instead of the gradient of the vehicle speed, because it has the following advantages:

• • Use of independently available sensor information
• • Less noise or less signal delay caused by leakage
• • Independence of the slope angle of the road, since the pitch angle γ according to a _sensor = v. - g · sin (θ) already included in acceleration sensor signal. The slope thus has no direct influence on the force balance. The mass estimate also copes robustly with mountains.

In a preferred embodiment of the control device according to the invention this has a non-volatile memory, wherein a determined vehicle mass and at least one Gütegröße, in particular a confidence interval, and wherein the at least one goodness variable is reset to a value indicative of inaccurate vehicle mass when the ignition of the vehicle has been inactive for at least a predetermined period of time. Thus, a charging process that significantly changes the vehicle mass can be detected and taken into account in the estimation of the vehicle mass.

It is advantageous if the electronic control unit comprises an actuator for the driver-independent construction of braking force on one or more vehicle wheels, in particular an electrically operated hydraulic pump and at least one solenoid valve, and a computing unit that performs a vehicle dynamics control, wherein at least one parameter of the vehicle dynamics control, in particular the Activation threshold, depending on the currently determined vehicle mass is changed. The vehicle dynamics control can be a yaw moment control for maintaining the driver-specified course, a roll prevention to prevent tilting about the longitudinal axis of the vehicle, a trailer stabilization, a brake or traction control or hill start assist. The vehicle can thus be safely and comfortably controlled by the driver even in a heavily laden condition.

Further preferred embodiments will become apparent from the subclaims and the following description of an embodiment with reference to figures.

Show it

1 a schematic representation of forces acting on the vehicle,

2 a schematic system architecture of the estimation method according to a preferred embodiment of the invention,

3 various applications of the learning algorithm for the acceleration sensor offset, and

4 a schematic structure of the vehicle mass estimator according to a particularly preferred embodiment of the invention.

The method according to the invention is based on the evaluation of a balance of forces of the longitudinal dynamics of the motor vehicle or a consideration of engine torque and driving resistances in order to carry out a determination of the vehicle mass.

The basis of this method is thus a balance of the forces acting on the (force) vehicle longitudinal forces, which must apply in the event of an acceleration of the vehicle under the influence of the vehicle engine: F _closely - F _Accel + F + F _{Steig Roll} + F _air

The driving force F _eng is expediently calculated from the torque T _{eng of} the vehicle engine, a factor η for taking account of the transmission ratio and the efficiency of the transmission, and the dynamic wheel radius r _dyn :

The inertia force F _{accel of} the vehicle is determined by the vehicle mass m _Fzg , the acceleration of the vehicle a _Fzg , the moment of inertia of the wheels J _wh and the moment of inertia of the drive train J _pwt :

The temporal change ω. _wh the wheel speed and the time change ω. _closely The engine speed (or wheel and engine speed) are linked via the transmission ratio of the transmission.

The gradient force F _Increases when driving on a (longitudinally) inclined roadway (eg a mountain pass) depends on the pitch angle θ and the gravitational acceleration g: F = m _{climbing veh} · g · sin (θ)

The rolling resistance F _{Roll of} the vehicle can be described by a similar formula, knowing the rolling resistance coefficient f: F _Roll = f · m _vehicle · g

The air resistance F _air is calculated via the drag coefficient c _W , the cross-sectional area A of the vehicle, the air density ρ and the vehicle speed v:

The equation for the balance of forces can be changed by m _Fzg and are evaluated in the presence of a suitable driving situation, such as the acceleration on dry straight track.

1 shows a side view of an exemplary vehicle, wherein the forces acting and the influence of asymmetric loading are shown schematically.

The engine torque T _{eng of} the drive motor, z. As an internal combustion engine is transmitted in this example on the wheels of the front axle and causes an accelerating force F _eng on the vehicle. The moment of inertia J _{wh of} the wheels indicates that a part of the power applied by the drive motor is also converted into rotational energy of wheels and drive train.

When the driver operates the brake pedal with a certain force F _b , the braking system _{applies a} braking force F _brk = F _b1 + F _b2 (typically increased with respect to the actuating force) to the vehicle which, in accordance with the installed brake force distribution, is applied to the force F _{b b1 distributed} at the front axle and F _b2 at the rear axle.

In the example shown, a charge with the mass m has been stowed in the rear of the vehicle, which is why the contact force F _{n2 of} the rear axle has increased more than the contact force F _{n1 of} the front axle. The center of gravity G is shifted relative to the unloaded vehicle by dx. This causes the vehicle longitudinal axis is at an angle θ to the roadway longitudinal axis. Thus, a longitudinal acceleration measured by a vehicle-mounted sensor no longer corresponds to the acceleration force, but the sensor signal has an offset or an offset.

The constant offset caused by the center of gravity displacement dx or the shift in the measured longitudinal acceleration signal can lead to an overestimation or underestimation of the vehicle mass. If the load is placed in the front area of the hold, a negative offset results, which could lead to an overestimation of the mass. If the load is placed in the rear of the hold, there will be a positive offset of the acceleration signal, which could lead to underestimation of the mass. According to a preferred embodiment of the invention, it is therefore provided to detect the currently present displacement or the offset upon detection of a stable standstill on the plane and to correct the measured longitudinal acceleration signal by the additive displacement constant.

Furthermore, the center of gravity with respect to the unloaded vehicle is shifted upwards by dz. This higher center of gravity means that overturning about the longitudinal axis can already take place at a lower lateral acceleration compared to the unloaded vehicle. It is therefore advantageous if a vehicle dynamics control, in particular a roll-over prevention such as ARP (Active Rollover Protection), is adjusted depending on the determined vehicle mass.

Furthermore, the vehicle nods backwards or forwards due to the weight shift during acceleration or braking. These driving situation variable offsets of the acceleration signal lead in drive case to an underestimation of the mass or in case of braking to an overestimation of the mass and are even more difficult to detect. By carrying out an estimation of the vehicle mass during acceleration as well as during braking, the method according to the invention compensates for the influence of this drive pitch.

In addition, some required to evaluate the balance of forces of the longitudinal dynamics of the motor vehicle sizes are not known exactly, such as the rolling resistance coefficient, which may change due to its dependence on the road surface and within a drive.

The present invention therefore pursues the approach of calculating a number of raw mass values based on a longitudinal vehicle model and estimating the vehicle mass by averaging or further statistical evaluation of the gross mass values.

Preferably, the inventive method is performed by an electronic control unit of a brake system of the motor vehicle, as this often already provides a vehicle dynamics control and / or a brake slip control. Thus, it expediently has one or more computing units, in particular a kernredundanten microcontroller, and evaluation circuits for signals of connected sensors and one or more interfaces connected to a vehicle data bus such as CAN or FlexRay.

The electronic control unit is preferably connected to each wheel associated wheel speed sensors on all wheels that provide a speed signal. From the speed signals of z. B. 4 wheel speed sensors, a vehicle speed and (in particular averaged) wheel acceleration can be determined. A brake actuation by the driver can be detected in a hydraulic brake system based on the pressure in the master cylinder, alternatively or additionally, the signal of a brake pedal angle sensor or a Betätigungswegsensors be considered. Furthermore, the electronic control unit expediently has or is connected to a longitudinal acceleration sensor and a lateral acceleration sensor and / or a yaw rate sensor. As an acceleration sensor, an inertial or inertial sensor is preferably used, which comprises a movably mounted test mass and in particular is realized as a micro-electro-mechanical system.

Thus, it is expedient to detect signals of the sensors already present in the vehicle and / or to read out information such as a drive signal of the engine control via the CAN bus for determining the vehicle mass. For example, via a sensor connected to the engine control sensor on the crankshaft, the engine speed can be determined, and determined by comparison with the wheel speed, the gear ratio or the gear engaged. A control unit according to the invention can estimate the vehicle mass quickly and robustly despite offsets of the longitudinal acceleration signal due to unbalanced loading, drive and brake nodding and also despite outliers of the calculated mass values.

2 showed a schematic system architecture of the estimation method according to a preferred embodiment of the invention, wherein one or more steps of the method can be implemented as software modules in an electronic control unit and the determination of the vehicle mass expediently comprises an interconnection of two estimators.

Estimator 1 makes it possible to determine an additive displacement constant of the longitudinal acceleration signal, the input signals being the measured uncorrected longitudinal acceleration, a drive signal such as engine speed, at least one speed signal such as one or more wheel speed signals and a brake signal such as the master cylinder pressure measured with a pressure sensor. Estimator 1 can be divided into three modules:
In module 1 the detection of a steady state standstill takes place, in which case the speed signal falls below a predetermined threshold value and preferably the engine speed is within a predetermined interval (around the idling speed). Preferably, the settled standstill is characterized in that all wheels are stationary (all wheel speeds fall below a predetermined threshold value) and the drive train oscillation has decayed, which is preferably recognized by the time variation of the longitudinal acceleration sensor signal falling below a predetermined threshold value.

module 2 serves to detect a slope freedom of the road (at least in the longitudinal direction), with a flat road can be recognized by it, the speed signal falls below a predetermined threshold even after the vehicle has been decelerated, if no braking operation by the driver. Alternatively or additionally, the following method may also be provided: First, the current values of predetermined state signals (longitudinal acceleration signal and / or engine torque and / or engine speed) are stored in the presence of a steady-state standstill; then the slope freedom of the road is determined by the values of all given state signals lie within a predetermined interval around the stored signals when the driver puts the vehicle back in motion by accelerator operation.

In module 3 the determination of the offset is carried out. Conveniently, a number of instantaneous values (eg, 10) of the longitudinal acceleration sensor signal are stored for a time interval after the road clearance has been determined and before the driver moves the vehicle by accelerator operation. This is useful because the instantaneous acceleration sensor signal is caused by loading in this time interval alone. It is therefore particularly appropriate, by a stochastic estimator such. For example, an averaging estimator estimates the displacement of the longitudinal acceleration signal along with a quality score, such as a confidence interval. In principle, it would also be possible to use a single measured value in the case of a detected standstill on the plane for the determination of the offset or the displacement of the acceleration sensor signal. An averaging has the advantage of a higher robustness against outliers and thus a better accuracy.

3 visualizes three possible applications of estimator 1, wherein in the left image (a) the previously running vehicle is stopped in front of a red traffic light, in the middle image (b) the vehicle stops in front of the traffic light, and in the right image (c) the vehicle starts again. A movement of the motor vehicle is indicated by an arrow in front of the illustrated vehicle, and longitudinal forces acting by means of arrows under the wheels represent braking (arrows to the left) or acceleration (arrows to the right) of the vehicle, the length of the arrows being the amount of acting Force indicates.

Line 1 shows the case of a downhill vehicle (down hill). To stop the vehicle a significant braking force is needed, which can lead to a pitching of the vehicle, so that the front axle loaded and the rear axle is relieved. While the vehicle is at the traffic lights, in module 1 a standstill of the vehicle recognized by the fact that the speed is below a threshold and preferably no powertrain vibrations present. When checking in module 2 However, no freedom of pitch is detected because the vehicle is in motion in the moment in which the brake is released. Thus, no shift of the acceleration sensor signal is determined in this case, since this would be corrupted by the present slope.

Line 2 shows the case of an uphill vehicle. For stopping a relatively low braking force is needed because the vehicle is already delayed by the slope force. Again in module 1 a stoppage detected, with module 2 recognizes again no freedom of pitch, since a brake operation is required to hold the vehicle at the traffic light. Also in this situation, no (falsified) acceleration signal offset is determined.

Line 3 shows the case of an on-plane motor vehicle. Here is in module 1 a vibration-free standstill found and cached the acceleration sensor signal. When driving off the vehicle is in module 2 ascertained the pitch, whereupon module 3 the displacement of the longitudinal acceleration sensor signal is determined. At least up to a longer standstill of the vehicle, during which a change in the load state and thus the vehicle mass may have taken place, thus caused by the load sensor signal offset is known.

Referring again to 2 will be in module 4 , a sum or difference element, the signal of the longitudinal acceleration sensor corrected by the determined displacement, ie it finds offset compensation of the acceleration sensor by means of the module 3 estimated Ax offset instead. Thus, a misaligned due to uneven loading acceleration sensor signal is avoided.

die – wie oben für den Fall einer Beschleunigung diskutiert – aus dem Longitudinaldynamikmodell des Fahrzeugs abgeleitet wurde, wobei die Symbole folgende Bedeutung aufweisen:

• • T_wh: auf die Räder umgerechnetes antreibendes Motormoment
• • F_brk: Summe der Bremskräfte
• • r_dyn: dynamischer Radradius
• • J_pwt: Trägheitsmoment des Antriebsstrangs ohne Räder
• • J_wh: Trägheitsmoment der Räder
• • ω ._eng : zeitliche Änderung der Motordrehzahl
• • ω ._wh : Radbeschleunigung
• • f: Rollwiderstandskoeffizient
• • g ≈ 9,81 m / s² : Erdbeschleunigung
• • a_x = a_sensor – a_offset: um Beladungsoffset kompensierter Beschleunigungssensorwert
• • c_w: Luftwiderstandsbeiwert
• • A: Frontalfläche des Fahrzeugs
• • ρ: Luftdichte
• • v: Fahrzeuggeschwindigkeit

The estimator 2 for estimating the vehicle mass can usefully be subdivided into six modules:
In module 5 the raw mass values m _{roh are} calculated, preferably according to the formula

which, as discussed above for the case of acceleration, was derived from the longitudinal dynamics model of the vehicle, the symbols having the following meaning:

• • T _wh : driving motor torque converted to the wheels
• • F _brk: total braking forces
• • r _dyn : dynamic wheel radius
• • J _pwt : moment of inertia of the drive train without wheels
• • J _wh : Moment of inertia of the wheels
• • ω. _closely : time change of the engine speed
• • ω. _wh : Rad acceleration
• • f: rolling resistance coefficient
• • g ≈ 9.81 m / s² : Acceleration of gravity
• • a _x = a _sensor - a _offset : Acceleration _{sensor value} compensated for loading offset
• • c _w : drag coefficient
• • A: frontal area of the vehicle
• • ρ: air density
• • v: vehicle speed

The sizes mentioned are either known from the construction of the vehicle or generally as constants or, as in module 4 corrected longitudinal acceleration sensor signal measured with sensors arranged in the vehicle and corrected if necessary.

module 6 preferably carries out an assessment of the driving situation on the basis of one or more of the driving state data of the vehicle or the output signals of the connected sensors, such as pedal position and / or actuation travel of the brake pedal and / or accelerator pedal, pressure present in the master brake cylinder by the driver's brake actuation, current vehicle speed, current engine torque, measured yaw rate, measured lateral acceleration, from which a number or a quality value as a measure of the quality c _{m of} the module 5 calculated raw mass value m _{raw is} determined. The quality value c _m can in particular be expressed as a probability, ie assume values between zero and one.

In module 7 It is intended to carry out a determination of the statistical characteristics of the calculated raw mass values for a considered braking or acceleration process. It is particularly useful to determine the statistical moments in particular mean and variance of Rohmassenwerte {m} _raw, taking into account their quality values {c} _m is. Thus, therefore, a weighted average, and preferably a weighted variance may be formed, which is a best estimate of the vehicle mass for the considered braking or acceleration process.

Estimator 2 introduces module 8th preferably a histogram of the distribution of the weighted with quality values _{m} c Rohmassenwerte {m} _raw, which is updated continuously.

Under evaluation of statistical features from module 7 and / or the reliability of the offset estimation from module 3 and / or the quality values {c _m } and / or the correspondence of the raw mass values {m _raw } with the histogram from module 8th be in module 9 the raw mass values {m _raw } are expediently weighted, with each raw mass value or a quantity of raw mass values being assigned an uncertainty value R. module 9 Thus, outliers can be assigned a greater uncertainty value at raw mass values {m _raw }, and thus a falsification of the determined vehicle mass can be reduced by an excessive weighting of random deviations.

In module 10 An estimate of the vehicle mass is made by means of a least-square method from the average values of the raw mass values {m _raw } and their uncertainty values {R}. Preferably, the least-square method is recursively designed in a manner known per se in order to enable a real-time capability. On the basis of the uncertainty values and / or the number of raw mass values considered, the determined vehicle mass is expediently assigned a product size. This may in particular be a confidence interval, which can be transmitted as an accuracy measure of the determined vehicle mass to further software modules or control units.

4 shows a schematic structure of the vehicle mass estimator according to a particularly preferred embodiment of the invention. Here, the method for implementation in an electronic control unit (ECU), in particular a brake control unit is optimized and can be performed online while driving.

The vehicle mass estimator receives the signals from sensors present in the vehicle and / or information from electronic control devices such as engine torque and / or acceleration sensor signal and / or vehicle speed and / or admission pressure or pressure in the master cylinder and / or pedal travel and / or yaw rate. The supplied signals comprise an offset-compensated longitudinal acceleration.

module 401 provides an a-priori evaluation of the sensor signals. Based on various criteria is checked whether an appropriate acceleration or braking phase is present. The driving speed should be constant within a given accuracy even during braking, advantageously, the gear ratio of the transmission (or the gear engaged) not to be high and constant, and should be straight ahead on a lane with a not too high grade. It can be provided that, additionally or alternatively, an evaluation is carried out by means of fuzzy classifiers. When the conditions are met, and when the braking or driving acceleration is within a predetermined interval in magnitude, a high quality value c _{m is} output and, as a result, a learning phase is detected.

module 402 serves to Rohmassenschätzung, wherein the balance of forces in the longitudinal direction (as explained above) are evaluated and a Rohmassenwert m is outputted _raw.

In module 403 the statistical properties of the gross mass values m _{raw of} a learning phase are considered taking into account the received quality values c _m , wherein in particular the weighted average μ, the weighted variance σ ² or weighted standard deviation σ are output. Furthermore, the duration of the learning phase Δt and an averaged or summed quality value A are output. This lower stage of a two-stage stochastic mass estimate thus serves for data processing for the upper level, in particular by recursive calculation of the statistical moments such as mean and variance.

module 405 allows an a posteriori analysis of the various learning phases, in particular a histogram of the previously determined raw mass values is performed. Based on the analysis, an uncertainty value R of the raw mass values determined during a learning phase is output. This a posteriori analysis is preferably performed using a recursive histogram and also allows rejection of outlier outliers, with detected outliers being less heavily weighted or ignored.

module 406 provides a statistical evaluation of the values for the vehicle mass obtained during several braking or acceleration phases with an optimization method. In this upper stage of a two-stage stochastic mass estimation, preferably a recursive mass calculation takes place, taking into account the information from the lower stage. In particular, a recursive least-squares method known per se is used to estimate the vehicle mass. The estimated vehicle mass m _est and a goods quantity P corresponding to a standard deviation or a confidence interval of the estimated vehicle mass are output.

module 404 is a suitably designed as a state machine management module, which determine the beginning and end of a learning phase, given the beginning of a statistical evaluation in the second stage and can also re-initialize the mass estimate. For this purpose, it can send signals to the other modules, which are indicated by arrows in the figure. For a reinitialization it is with module 407 which gives a new start value m _init for the estimated vehicle mass and a new (low) goods size P _init to module 406 pretends. In addition, it may also be provided that module 404 predetermines a forgetting factor λ to the second stage of the stochastic mass estimation, that is to say that previous learning phases or their raw mass values obtained are less heavily weighted with increasing time. This prevents a permanent falsification of estimated vehicle masses due to outliers far behind.

It is advantageous if estimated vehicle mass and preferably also the product size are supplied to a system known per se for vehicle dynamics control, in particular a system for preventing lateral overturning of the vehicle, and if one or more parameters of the vehicle dynamics control system are adjusted in accordance with the determined vehicle mass , Thus, a safe ride is guaranteed even with a heavily laden vehicle.

QUOTES INCLUDE IN THE DESCRIPTION

This list of the documents listed by the applicant has been generated automatically and is included solely for the better information of the reader. The list is not part of the German patent or utility model application. The DPMA assumes no liability for any errors or omissions.

Cited patent literature

• DE 4228413 A1 [0003]
• WO 03/023334 A1 [0004]
• EP 1858736 B1 [0005]

Claims (13)

A method in which the vehicle mass of a moving motor vehicle is determined, wherein a speed signal, a longitudinal acceleration signal, a brake signal and a drive signal are considered, and wherein a force balance of the longitudinal dynamics of the motor vehicle is evaluated, wherein the longitudinal acceleration signal is measured by an inertial sensor, characterized in that an evaluation of the force balance of the longitudinal dynamics takes place both during acceleration and braking processes, wherein a number of raw mass values is calculated, and wherein the vehicle mass is determined on the basis of a statistical evaluation of the raw mass values, which comprises at least one averaging. A method according to claim 1, characterized in that the longitudinal acceleration signal is corrected by an additive displacement constant, in particular to compensate for the influence of an uneven loading on the longitudinal acceleration signal. A method according to claim 2, characterized in that a new value of the additive displacement constant is determined when a stable standstill has been detected on the plane, preferably a stable standstill on the plane is then detected if one or more, in particular all of the following conditions the following conditions have been met: • A braking operation has taken place within a specified period of time • No braking action is currently taking place • The speed signal falls below a specified holding threshold value • A temporal change in the longitudinal acceleration signal falls below a predefined fluctuation threshold value Method according to at least one of the preceding claims, characterized in that an evaluation of the balance of forces of the longitudinal dynamics only takes place when a stable driving condition is detected. A method according to claim 4, characterized in that the stable driving state is detected when one or more, in particular all, the following conditions are met: • An amount of an acceleration request of the driver is in a predetermined acceleration interval, in particular either an accelerator pedal operation or a • A driving speed set by the driver and / or a measured yaw rate is below a predetermined steering threshold and / or yaw rate threshold • A change in the timing of the drive signal falls below a preset drive threshold value • The longitudinal acceleration signal exceeds a predetermined minimum threshold value • An amount of a measured lateral acceleration l is below a specified threshold • There is neither a vehicle dynamics control nor a brake slip control active Method according to at least one of the preceding claims, characterized in that a calculation of raw mass values during a braking or acceleration process is carried out continuously, in particular at predetermined time intervals, wherein average and variance are calculated for the raw mass values each of a braking or acceleration process. A method according to claim 6, characterized in that in the statistical analysis, the vehicle mass by means of an optimization method, in particular a least-squares method, from the average values and variances of several braking or acceleration operations is determined. A method according to claim 6 or 7, characterized in that calculated raw mass values of a braking or acceleration process are discarded if they differ by more than a predetermined multiple of the variance from the mean value of the braking or acceleration process, and thus not further in the determination of the vehicle mass received. Method according to at least one of the preceding claims, characterized in that the speed signal is determined based on at least one wheel speed sensor and / or the brake signal is determined based on a brake pressure sensor and / or a pedal travel sensor on the brake pedal and / or the drive signal from one of an engine control unit of an internal combustion engine and /or an electric drive reported torque and / or measured at a motor shaft speed is determined. Method according to at least one of the preceding claims, characterized in that in addition to the balance of forces of longitudinal dynamics and a balance of forces of the lateral dynamics is considered, with a lateral acceleration signal and at least two wheel speed signals are considered and wherein the coordinates of the current center of gravity are determined in a vehicle-fixed coordinate system. Electronic control unit, in particular for a brake system of a motor vehicle, with interfaces for the connection of at least one wheel speed sensor, at least one brake actuation sensor, at least one inertial sensor and an interface to a vehicle data bus, characterized by an arithmetic unit, which performs a method according to at least one of the preceding claims. Electronic control unit according to claim 11, characterized in that it has a nonvolatile memory, wherein a determined vehicle mass and at least one Gütegröße, in particular a confidence interval, is stored, and wherein the at least one Gütegröße is reset to a inaccurate vehicle mass indicative value when the Ignition of the vehicle for at least a predetermined period of time was not active. Electronic control unit according to one of claims 11 or 12, characterized in that it comprises an actuator for driver-independent construction of braking force on one or more vehicle wheels, in particular an electrically operated hydraulic pump and at least one solenoid valve, and a computing unit that performs a vehicle dynamics control, wherein at least a parameter of the vehicle dynamics control, in particular an activation threshold value, is changed as a function of the currently determined vehicle mass.

Images