Ascertaining an offset of an inertial sensor

Abstract

A method of providing an additive offset of a longitudinal acceleration signal of a traveling motor vehicle. The signal being measured by an inertial sensor is ascertained. At least the longitudinal acceleration signal, a braking signal, and a drive signal are detected. A force balance of the longitudinal dynamic of the motor vehicle is analyzed. The signals are detected both during at least one acceleration process as well as during at least one braking process. The signals during the acceleration processes are detected and/or analyzed separately from the signals during the braking processes, and the additive offset is ascertained by comparing the signals detected during the acceleration processes or the values calculated therefrom with the signals detected during the braking processes or the values calculated therefrom. The invention further relates to an electronic controller.

Claims

1. A method for measuring an additive displacement of a longitudinal acceleration signal of a traveling motor vehicle with an inertial sensor comprising: detecting at least the longitudinal acceleration signal, a brake signal and a drive signal; analyzing a balance of forces of longitudinal dynamics of the motor vehicle; detecting the signals for at least one acceleration process and for at least one braking process, wherein the signals for acceleration processes are detected separately from the signals for braking processes; determining the additive displacement from a comparison of the signals detected during acceleration processes or values calculated from the signals detected during acceleration process, with the signals detected during braking processes or values calculated from the signals detected during braking processes, wherein analysis of the balance of forces of longitudinal dynamics includes a determination of a longitudinal force acting on the vehicle using at least one of the drive and brake signal; determining a first longitudinal acceleration a.sub.acc during an acceleration process using the longitudinal acceleration signal; determining a first longitudinal force F.sub.acc at least using the drive signal; determining a second longitudinal acceleration a.sub.dec during a braking process using the longitudinal acceleration signal; determining a second longitudinal force F.sub.dec at least using the brake signal; and determining an additive displacement a.sub.x.sup.drift of the longitudinal acceleration according to: $a_x^{\mathrm{drift}}=\frac{a_{\mathrm{acc}}{F}_{\mathrm{dec}}-a_{\mathrm{dec}}{F}_{\mathrm{acc}}}{F_{\mathrm{dec}}-F_{\mathrm{acc}}}.$

2. The method of claim 1, wherein analysis of the balance of forces of the longitudinal dynamics is only carried out if a stable driving state.

3. The method of claim 2, wherein the stable driving state is traveling straight ahead.

4. The method of claim 2, wherein the stable driving state is detected if at least one of the following conditions are met: a magnitude of an acceleration demand by a driver lies within a predetermined acceleration interval; a gas pedal operation by the driver exceeds a predetermined threshold value; a brake pedal operation by the driver exceeds a predetermined threshold value; a speed of travel of the traveling motor vehicle lies within a predetermined speed interval; a magnitude of a steering angle set by the driver; a measured yaw rate below a predetermined steering angle threshold value; a predetermined yaw rate threshold value; a magnitude of a change with time of the drive signal lies below a predetermined fluctuation threshold value; a magnitude of a change with time of the brake signal lies below a predetermined fluctuation threshold value; the magnitude of the longitudinal acceleration signal exceeds a predetermined minimum threshold value; magnitude of a measured lateral acceleration lies below a predetermined turn threshold value; and none of drive dynamics control, brake slip control nor drive slip control is active.

5. The method of claim 1, wherein a wind resistance force is determined and taken into account using a speed of travel of the vehicle.

6. The method of claim 1, wherein detecting of the sensor signals is carried out continuously at fixed time intervals.

7. The method of claim 6, further comprising determining a first mass using the signals measured during acceleration processes; determining a second mass using the signals measured during braking processes; and wherein determining the additive displacement of the longitudinal acceleration signal is carried out if a difference between the first mass and the second mass exceeds a predetermined tolerance threshold value.

8. The method of claim 7, wherein determining the additive displacement is carried out recursively further comprising: determining a plurality of the first mass values and the second mass values; calculating an expected value for the first mass and an expected value for the second mass; minimizing a difference between the first expected value for the first mass and the expected value for the second mass; and maintaining the determined additive displacement when a difference between the expected values lies below a termination threshold.

9. The method of claim 8, wherein the additive displacement is predetermined when the expected value at least one of meets and exceeds the termination threshold.

10. The method of claim 1, further comprising determining a speed signal using at least one wheel revolution rate sensor.

11. The method of claim 1, further comprising determining the brake signal is determined using one of a brake pressure sensor and a pedal travel sensor on a brake pedal.

12. The method of claim 1, further comprising determining the drive signal from at least one of: a drive torque signaled by an engine control unit of an internal combustion engine, by a motor control unit of an electrical drive, and a revolution rate measured on a motor shaft.

13. An electronic control unit for a brake system of a motor vehicle, comprising: interfaces for connection of at least one wheel revolution rate sensor and at least one brake operation sensor; at least one inertial sensor disposed in a longitudinal direction and an interface to a vehicle data bus; and a computing unit with instructions for: detecting at least a longitudinal acceleration signal, a brake signal and a drive signal; analyzing a balance of forces of longitudinal dynamics of the motor vehicle; detecting the signals for at least one acceleration process and for at least one braking process, wherein the signals for acceleration processes are detected separately from the signals for braking processes; determining an additive displacement from a comparison of the signals detected during acceleration processes or values calculated from the signals detected during acceleration process with the signals detected during braking processes or values calculated from the signals detected during braking processes, wherein analysis of the balance of forces of the longitudinal dynamics is only carried out if a stable driving state, wherein the stable driving state is traveling straight ahead; determining a first longitudinal acceleration a.sub.acc during an acceleration process using the longitudinal acceleration signal, determining a first longitudinal force F.sub.acc at least using the drive signal; determining a second longitudinal acceleration a.sub.dec during a braking process using the longitudinal acceleration signal; determining a second longitudinal force F.sub.dec at least using the brake signal; and determining additive displacement a.sub.x.sup.drift of the longitudinal acceleration according to: $a_x^{\mathrm{drift}}=\frac{a_{\mathrm{acc}}{F}_{\mathrm{dec}}-a_{\mathrm{dec}}{F}_{\mathrm{acc}}}{F_{\mathrm{dec}}-F_{\mathrm{acc}}}.$

14. The electronic control unit of claim 13, the stable driving state is detected if at least one of the following conditions are met: a magnitude of an acceleration demand by a driver lies within a predetermined acceleration interval; a gas pedal operation by the driver exceeds a predetermined threshold value; a brake pedal operation by the driver exceeds a predetermined threshold value; a speed of travel of the traveling motor vehicle lies within a predetermined speed interval; a magnitude of a steering angle set by the driver; a measured yaw rate below a predetermined steering angle threshold value; a predetermined yaw rate threshold value; a magnitude of a change with time of the drive signal lies below a predetermined fluctuation threshold value; a magnitude of a change with time of the brake signal lies below a predetermined fluctuation threshold value; a magnitude of the longitudinal acceleration signal exceeds a predetermined minimum threshold value; magnitude of a measured lateral acceleration lies below a predetermined turn threshold value; and none of drive dynamics control, brake slip control nor drive slip control is active.

15. The electronic control unit of claim 13, wherein an analysis of the balance of forces of the longitudinal dynamics includes a determination of a longitudinal force acting on the vehicle using at least one of the drive and brake signal.

16. The electronic control unit of claim 13, wherein a wind resistance force is determined and taken into account using a speed of travel of the vehicle.

17. The electronic control unit of claim 13, wherein detection of the sensor signals is carried out continuously at fixed time intervals.

18. The electronic control unit of claim 17 comprising further instructions for: determining a first mass using the signals measured during acceleration processes determining a second mass using the signals measured during braking processes, and wherein determining additive displacement of the longitudinal acceleration signal is carried out if a difference between the first mass and the second mass exceeds a predetermined tolerance threshold value.

19. The electronic control unit of claim 18 wherein determining the additive displacement is carried out recursively, and electronic control unit comprises further instructions for: determining a plurality of the first mass values and the second mass values; calculating an expected value for the first mass and an expected value for the second mass; minimizing the difference between the first expected value for the first mass and the expected value for the second mass; and maintaining the determined additive displacement when the difference between the expected values lies below a termination threshold.

20. The electronic control unit of claim 19, wherein the additive displacement is predetermined when the expected value at least one of meets and exceeds the termination threshold.

21. The electronic control unit of claim 13, further comprising determining the drive signal from at least one of: a drive torque signaled by an engine control unit of an internal combustion engine, by a motor control unit of an electrical drive, and a revolution rate measured on a motor shaft.

22. The electronic control unit of claim 13, further comprising an actuator for building up a brake force on one or more vehicle wheels independently of a driver, wherein the actuator is an electrically operated hydraulic pump and at least one solenoid valve; and a computing unit that implements drive dynamics control, wherein a longitudinal acceleration signal corrected by the determined additive displacement is fed into the drive dynamics control.

Description

BRIEF DESCRIPTION OF THE DRAWINGS

(1) The present disclosure will become more fully understood from the detailed description and the accompanying drawings, wherein:

(2) FIG. 1 shows a schematic representation of forces acting on the vehicle;

(3) FIG. 2 shows a schematic structure of a recursive estimation method for the determination of a longitudinal acceleration sensor offset; and

(4) FIG. 3 shows a schematic representation of mass distributions during a successive offset correction of the longitudinal acceleration sensor signal.

DETAILED DESCRIPTION OF THE DRAWINGS

(5) FIG. 1 shows a side view representation of an exemplary vehicle, wherein the acting forces and the influence of asymmetrical loading are represented schematically.

(6) The engine torque T.sub.eng of the drive engine, for example an internal combustion engine, is transferred in said example by means of the wheels of the front axle and causes a driving force F.sub.eng on the vehicle. It is indicated by the moments of inertia J.sub.wh of the wheels that part of the power applied by the drive engine is also converted into rotational energy of the wheels and drive train. If the acting decelerating forces or drive resistances, such as the rolling resistance and wind resistance, are taken into account then an accelerating force F.sub.acc that causes an increase v{dot over ()}>0 of the longitudinal speed v.sub.x of the vehicle can be calculated from the difference of the driving force and the decelerating forces:
F.sub.acc=m.sub.Fzg.Math.{dot over (v)}.sub.x

(7) For this initially traveling on a flat section and a knowledge of the mass of the vehicle m.sub.Fzg are assumed for simplicity.

(8) If the driver operates the brake pedal with a certain force F.sub.b, then a brake force F.sub.brk=F.sub.b1+F.sub.b2 is exerted on the vehicle by the brake system (typically boosted compared to the actuating force) that is distributed according to the installed brake force distribution into the force F.sub.b1 on the front axle and F.sub.b2 on the rear axle. Accordingly, taking into account any acting force of the drive engine and the drive resistances additionally decelerating the vehicle, a decelerating force F.sub.dec can be calculated that causes a decrease v.sub.x<0 in the longitudinal speed v.sub.x of the vehicle:
F.sub.dec=n.sub.Fzg.Math.{dot over (v)}.sub.x|

(9) In the example shown, a load with the mass m has been stowed in the rear of the vehicle, therefore the contact force F.sub.n2 of the rear axle has increased more than the contact force F.sub.n1 on the front axle. The center of gravity G is shifted relative to the unladen vehicle by dx. This results in the longitudinal axis of the vehicle being at an angle to the longitudinal axis of the road. As a result, a longitudinal acceleration measured by a fixedly mounted sensor no longer corresponds to the acceleration force, but the sensor signal has an offset or an additive displacement. Furthermore, it is indicated that the center of gravity is shifted upwards by dz relative to the unladen vehicle. Said higher center of gravity results in rolling about the longitudinal axis being able to occur at a lower lateral acceleration compared to the unladen vehicle.

(10) The additive displacement in the measured longitudinal acceleration signal caused by the center of gravity displacement dx can for example cause a negative offset in the case of loading in the front region of the cargo space, which could result in an overestimate of the mass during an acceleration process.

(11) If the load is placed in the rear region of the cargo space, there is a positive offset of the acceleration signal, which could result in an underestimate of the mass during an acceleration process. Besides the influence of non-uniform loading, a large temperature change of the sensor can also cause an offset of the acceleration sensor signal. The measured acceleration value is therefore composed as follows:

(12) $\begin{array}{cc}{a}_x^{\mathrm{Sensor}}\left(k\right)=\underset{\underset{a_x^{\mathrm{true}}\left(k\right)}{}}{{\stackrel{.}{v}}_x\left(k\right)+\sin \left(\left(k\right)\right).Math.g}+a_x^{\mathrm{drift}}& \left(1\right)\end{array}$
wherein a.sub.x.sup.Sensor(k)
denotes the measured sensor value, v{dot over ()}(k) denotes the derivative of the longitudinal speed of the vehicle, (k) denotes the road gradient angle, g denotes the constant acceleration due to gravity, a.sub.x.sup.drift
denotes the additive displacement or the offset of the acceleration sensor signal and a.sub.x.sup.true(k)
denotes the offset-free acceleration sensor value.

(13) Time-varying variables or successive measurement values are denoted here by the index k, which thus indicates the respective time.

(14) If the two conditions are met, that 1.) the vehicle is at a standstill v{dot over ()}.sub.x=0 and 2.) that the road has zero gradient (y=0), the sensor offset value a.sub.x.sup.drift could in principle be determined by measuring a.sub.x.sup.sensor. For checking whether the second condition of a zero gradient road is met, either an independent sensor for measuring a road gradient angle or a check of the following conditions would be necessary, that i.) no brake is operated, ii.) that the vehicle is not being driven and iii.) that the vehicle is not rolling. Because for reasons of cost most vehicles do not comprise an independent gradient angle sensor, and with many vehicles, in particular with vehicles with automatic gearboxes, the conditions i.) and ii.), that no brake is operated and that the vehicle is not being driven, are not met continuously, the determination of the additive displacement is advantageously carried out while traveling.

(15) An electronic control unit for a brake system of the motor vehicle, provides drive dynamics control and/or brake slip control and thereby comprises one or more computing units, in particular a redundant core microcontroller, analyzing circuits for signals of connected sensors as well as one or more interfaces connected to a vehicle data bus such as a CAN bus or a FlexRay bus. In this case, advantageously only signals of sensors already present in the vehicle are detected and/or information such as a drive signal of the engine controller are read out by means of the CAN bus. For example, the revolution rate of the engine can be determined by means of a sensor on the crankshaft connected to the engine controller, and the transmission ratio or the engaged gear can be determined by means of a comparison with the wheel revolution rate.

(16) The electronic control unit is preferably connected to wheel revolution rate sensors on all wheels that are each associated with a wheel and that provide a speed signal. A vehicle speed and a (in particular averaged) wheel acceleration can be determined from the speed signals of the for example 4 wheel revolution rate sensors. A brake operation by the driver in a hydraulic brake system can be detected using the pressure in the master brake cylinder; alternatively or additionally, the signal of a brake pedal angle sensor or of an actuation travel sensor can also be considered. Furthermore, the electronic control unit advantageously comprises a lateral acceleration sensor and/or a yaw rate sensor or is connected to the same. The longitudinal acceleration sensor is implemented as an inertial sensor that comprises a displaceably supported test mass and in particular is implemented as a microelectromechanical system.

(17) In principle, the additive displacement can be determined from the comparison of a braking process with an acceleration process, as is briefly explained below. The following equation for the determined mass m.sub.acc during an acceleration process applies following compensation of the additive displacement:

(18) $m_{\mathrm{acc}}=\frac{F_{\mathrm{acc}}}{a_{\mathrm{acc}}-a_x^{\mathrm{drift}}}$

(19) Accordingly, the following equation for the determined mass m.sub.dec during a braking process applies after compensation:

(20) $m_{\mathrm{dec}}=\frac{F_{\mathrm{dec}}}{a_{\mathrm{dec}}-a_x^{\mathrm{drift}}}$

(21) Because the mass of a vehicle while traveling remains the same (except for the decrease in the tank contents that is neglected here), it must further be true that: m.sub.acc=m.sub.dec.

(22) The offset of the acceleration sensor can thus be calculated according to the following equation:

(23) $\begin{array}{cc}{a}_x^{\mathrm{drift}}=\frac{a_{\mathrm{acc}}.Math.F_{\mathrm{dec}}-a_{\mathrm{dec}}.Math.F_{\mathrm{acc}}}{F_{\mathrm{dec}}-F_{\mathrm{acc}}}.& \left(2\right)\end{array}$

(24) If the calculation of the additive displacement is carried out with consideration of a plurality of braking and drive processes. For an accurate and reliable result, it is necessary for this that there is an equivalent representation of samples from the two groups (acceleration and braking), for example the number of braking processes corresponds to the number of acceleration processes, wherein in the sense of adequate statistical quality, it can be checked for example that the number of braking or acceleration processes considered exceeds a predetermined threshold value. So that a useful additive displacement can be determined, a.sub.x.sup.drift may not change during the estimation process, therefore it should in particular be ensured that the load in the cargo space is disposed in such a way and/or an active fixing of the load in the cargo space is carried out so that the load remains in the same position and does not slide back and forth during braking and acceleration processes.

(25) Alternatively or in combination with the calculation according to equation (2), which for a correct result should be based on using a number of raw measurement values of vehicle masses m.sub.dec and m.sub.acc determined during braking and acceleration processes that are only available after a certain traveling time, a recursive determination or a successive approximation of the additive displacement can also be carried out.

(26) FIG. 2 shows a schematic structure of a recursive estimation method for the determination of a longitudinal acceleration sensor offset. Individual modules can be interpreted here as individual steps of the method. The signals from sensors present in the vehicle and/or information from electronic control units such as engine torque signals and/or acceleration sensor signals and/or a vehicle speed and/or a pre-pressure or a pressure in the brake master cylinder and/or a pedal travel and/or a yaw rate as well as the measured longitudinal acceleration are fed to the estimation device.

(27) Module 401 is used for learning coordination and is advantageously implemented as a state machine. Said module can predetermine the start and the end of the determination of raw measurement values and the start of a statistical analysis of the raw measurement values and can also carry out re-initialization. For this purpose, it can send signals to the other modules that are indicated in the figure by arrows. Advantageously, said module also comprises an a-priori analysis of the sensor signals, by which a check can be made of whether there is a suitable acceleration or braking process using various criteria.

(28) In particular, no engagement of drive dynamics control or brake slip control should take place, the speed of travel should lie within a predetermined interval of average speeds, the drive torque also should be constant to within a predetermined accuracy even during braking processes, advantageously the transmission ratio of the gearbox (or the engaged gear) should not be too high and should be constant, and travel should be straight-ahead on a road with not too great a gradient. It can be provided that an assessment is additionally or alternatively carried out by means of fuzzy classifiers. On meeting one or more of, preferably all, conditions and if in addition the magnitude of the braking or drive acceleration lies within a predetermined interval, a learning phase is detected or started.

(29) Module 402 is used for raw mass estimation, wherein at each sampling time point at which there are new sensor signals, the balance of forces in the longitudinal direction is analyzed and a raw mass value {circumflex over (m)}(k) is calculated according to the following equation:

(30) $\begin{array}{cc}\hat{m}\left(k\right)=\frac{F\left(k\right)}{a_x^{\mathrm{Sensor}}\left(k\right)-{\hat{a}}_x^{\mathrm{drift}}\left(k\right)}& \left(3\right)\end{array}$

(31) Here {circumflex over (m)}(k) denotes the calculated raw mass value at a point in time k, F(k) denotes the sum of the forces determined in the longitudinal direction at the point in time k, which is calculated inter alia using engine torque and brake pressures, and .sub.x.sup.drift(k) denotes the value of the true acceleration sensor offset a.sub.x.sup.drift at the point in time k estimated by the method.

(32) Using the insertion of equation (1) in equation (3),

(33) $\hat{m}\left(k\right)=\frac{F\left(k\right)}{a_x^{\mathrm{true}}\left(k\right)+a_x^{\mathrm{drift}}-{\hat{a}}_x^{\mathrm{drift}}\left(k\right)}$

(34) It can be seen that the aim of the estimation is to compensate the effect of the additive displacement a.sub.x.sup.drift with a correction constant .sub.x.sup.drift(k).

(35) Uncompensated, a.sub.x.sup.drift causes a deviation m from the mass m(k) sought: {circumflex over (m)}(k)=m(k)+m(.sub.x.sup.drift(k))

(36) The sign of the averaged deviation B{m(.sub.x.sup.drift(k))} without compensation, i.e. with .sub.x.sup.drift(k)=0, is dependent on the driving situation for a constant offset, as can be seen from the following table:

(37) TABLE-US-00001 Acceleration: Braking: a.sub.x.sup.Sensor (k) > 0 a.sub.x.sup.Sensor (k) < 0 a.sub.x.sup.drift > 0 E{m(.sub.x.sup.drift (k))} > 0 E{m (.sub.x.sup.drift (k))} < 0 (for example rear load) a.sub.x.sup.drift < 0 E{m(.sub.x.sup.drift (k))} < 0 E{m (.sub.x.sup.drift (k))} > 0 (for example front load)

(38) If compensation of the additive displacement takes place, then the averaged deviation E{m} in each row of the table tends to zero regardless of the acceleration and braking.

(39) The additive displacement can thus be determined in principle using the following optimization problem:

(40) $\underset{{\hat{a}}_x^{\mathrm{drift}}}{\min }.Math.E\left\{m\left({\hat{a}}_x^{\mathrm{drift}}\right)\right\}.Math.$

(41) Because m(k) is unknown, said optimization problem cannot be directly processed. Owing to the ability to calculate {circumflex over (m)}(k), the following analogue optimization problem is considered instead:
min.sub..sub.x.sub.driftE{{circumflex over (m)}|dec,.sub.x.sup.drift}E{{circumflex over (m)}|acc,.sub.x.sup.drift}(4)

(42) To solve the optimization problem of (4), the present method divides the offset-affected raw measurement values calculated in module 402 into two groupsa group for raw measurement values {circumflex over (m)}/acc from acceleration phases and a group for raw measurement values {circumflex over (m)}/dec from braking phases.

(43) In module 403 the statistical properties of the raw measurement values {circumflex over (m)}.sub.dec determined during braking processes are considered, wherein in particular the weighted average value , the weighted variance .sup.2 or the weighted standard deviation are output.

(44) Accordingly, in module 404 the statistical properties of the raw measurement values {circumflex over (m)}.sub.acc determined during the acceleration processes are considered, wherein a (in particular recursive) calculation of the statistical moments such as the weighted average value , the weighted variance .sup.2 or the weighted standard deviation is carried out.

(45) Using a statistical test, preferably a hypothesis test, on the one hand a check is made in module 405 of whether the representativeness of the random samples is guaranteed, and on the other hand by means of the null hypothesis it is determined whether the expected values of the two groups are identical to a certain significance. If the null hypothesis is discarded at a specified significance level, this means that at the significance level part of the offset a.sub.x.sup.drift is still not compensated. Poor candidates for offset values are discarded in this way.

(46) Module 406 carries out an optimization method using the raw measurement values and the statistical test, wherein both grid-based methods and also gradient methods can be used for methodical trial-and-error purposes.

(47) Grid-based methods aim to achieve a global optimum of .sub.x.sup.drift. For this purpose, possible values of (.sub.x.sup.drift).sub.i and a.sub.x.sup.drift are distributed at regular intervals. For each of said values,
E.sub.i{{circumflex over (m)}|braked,(.sub.x.sup.drift).sub.i}E.sub.i{{circumflex over (m)}|accelerated,(.sub.x.sup.drift).sub.i}

(48) is calculated, wherein the value of .sub.x.sup.drift that has the deviation with the smallest magnitude is determined.

(49) Gradient methods aim to achieve a global optimum of .sub.x.sup.drift. In this case an offset estimate .sub.x.sup.drift(k) is iteratively determined until the improvement of a step lies below a threshold.

(50) Both grid-based methods and also gradient methods can be configured recursively in order to save hardware resources.

(51) FIG. 3 shows a schematic representation of mass distributions during a successive offset correction of the longitudinal acceleration sensor signal. In said example, a load is placed in the rear load space region. Diagrams with estimated vehicle masses at successive points in time t are shown here, wherein m.sub.dec denotes a mass estimated during a braking process and m.sub.acc a mass estimated during an acceleration process. A finite number of calculations or raw measurement values would only enable a coarse approximation to the continuous mass distributions shown.

(52) From a number of mass value calculations, the mass distributions shown at a point in time t.sub.0 would thus be obtained, which yields an average mass or an expected value E{m.sub.dec} for the estimated mass from a consideration of braking processes and an average mass or an expected value E{m.sub.acc} for the estimated mass from a consideration of acceleration processes. Because the offset caused by the load and/or a temperature-related offset has/have not yet been determined or compensated, a significant deviation between the expected values or average masses for braking and acceleration processes is apparent.

(53) Based on equation (2) or an optimization method or a consideration of the deviation, an additive displacement can be determined and thus compensated, wherein the expected values or mass distributions show a reduced deviation at a later point in time t.sub.1>t.sub.0.

(54) Accordingly, in a next step a newer or more accurate estimate of the additive displacement is determined, whereupon the expected values or mass distributions show a further reduced deviation at a later point in time t.sub.2>t.sub.1.

(55) Following a further step in the determination of the additive displacement, at a later point in time t.sub.3>t.sub.2 the mass distributions deviate from each other by less than the half-value width thereof (or the variance or a differently predetermined measure of the width of the mass distribution). The additive displacement has been determined sufficiently accurately and can be compensated for the duration of the journey.

(56) The foregoing preferred embodiments have been shown and described for the purposes of illustrating the structural and functional principles of the present invention, as well as illustrating the methods of employing the preferred embodiments and are subject to change without departing from such principles. Therefore, this invention includes all modifications encompassed within the scope of the following claims.