Method for detecting an imbalance of a vehicle wheel while the vehicle is rolling

09903780 · 2018-02-27

Assignee

RENAULT s.a.s. (Boulogne Billancourt, FR)

Inventors

Cpc classification

International classification

Abstract

A method and device detect the imbalance of a vehicle wheel. The method includes measuring a rotation speed of the wheel while the vehicle is moving, and calculating a filtered value by applying at least one step of band-pass filtering to the measured rotation speed value of the wheel. A position of a pass-band of the band-pass filtering step is offset while moving as a function of the rotation speed of the wheel.

Claims

1. A method of detecting the imbalance of a vehicle wheel, comprising: measuring a rotation speed of the wheel while the vehicle is moving; calculating a filtered value by applying at least one step of band-pass filtering to the measured rotation speed value of the wheel, wherein a position of a pass-band of the band-pass filtering step is offset while moving as a function of the rotation speed of the wheel.

2. The detection method as claimed in claim 1, further comprising: calculating a variance value representative of a statistical variance of the filtered value; and comparing the variance value to a variance threshold.

3. The detection method as claimed in claim 2, wherein the variance value is calculated by applying a step of low-pass filtering to the absolute value of a difference between the filtered value and a mean filtered value, the mean filtered value being a result of low-pass filtering of the filtered value.

4. The detection method as claimed in claim 2, wherein a result of the comparing the variance value to the variance threshold is taken into account only during phases of movement in which a speed of the vehicle is considered as stabilized.

5. The detection method as claimed in claim 2, further comprising: applying a first step of low-pass filtering to a value delivered by a wheel speed sensor to obtain a wheel rotation speed to which the band-pass filtering step is then applied.

6. The detection method as claimed in claim 2, further comprising: applying a final step of low-pass filtering to the variance value to obtain a filtered variance value that is compared to the variance threshold.

7. The detection method as claimed in claim 2, wherein the detection method is applied to monitoring of a level of inflation of the wheels of the vehicle.

8. The detection method as claimed in claim 1, wherein the band-pass filtering step is effected by means of a discrete band-pass filter characterized by five variable coefficients, the five coefficients being calculated as a function of three constant coefficients and a rotation speed of the wheel.

9. A device for detecting the imbalance of a motor vehicle wheel, comprising: a sensor of a rotation speed of a wheel; and a calculation unit to calculate an imbalance criterion for the wheel, wherein the calculation unit is configured to effect at least one step of band-pass filtering of a speed sensor signal of the sensor, a position of a pass-band being offset while moving as a function of the rotation speed of the wheel.

10. The device as claimed in claim 9, wherein the calculation unit is configured, during at least some steps of movement of the vehicle, to compare to a threshold a criterion value produced by filtering including the band-pass filtering step.

11. A vehicle, comprising: a detection device as claimed in claim 9.

Description

(1) Other objects, features and advantages of the invention will appear on reading the following description, given by way of nonlimiting example only and with reference to the appended drawings, in which:

(2) FIG. 1 shows a vehicle equipped with a detection device in accordance with the invention,

(3) FIG. 2 is a simplified diagram of the operation of a detection device in accordance with the invention,

(4) FIG. 3 shows wheel speed curves obtained in a calculation step of a detection method in accordance with the invention,

(5) FIG. 4 shows wheel speed variance curves obtained in another calculation step of a detection method in accordance with the invention.

(6) As shown in FIG. 1, a vehicle 5 moves on four wheels 1, 2, 3 and 4. Each of the wheels is equipped with a speed sensor 6 dedicated to the wheel. The speed sensors 6 dedicated to the respective wheels 1, 2, 3 and 4 deliver a respective angular speed value or a respective value proportional to an angular speed .sub.1, (.sub.2, .sub.3, .sub.4 to an electronic control unit 7 of the vehicle. To be more precise, the speed sensors actually associated with the wheels 1, 2, 3 and 4 deliver the speed they register to a first, second, third and fourth module, respectively, for calculating a balancing criterion B.sub.1, B.sub.2, B.sub.3, B.sub.4, respectively, associated with the wheel.

(7) Each module for calculating a balancing criterion sends a respective Boolean value BAL.sub.1, BAL.sub.2, BAL.sub.3, BAL.sub.4 to a first, second, third and fourth module, respectively, for calculating a respective pressure criterion P.sub.1, P.sub.2, P.sub.3, P.sub.4 for the tire associated with the wheel.

(8) Each module for calculating the pressure criterion also receives the respective rotation speed value .sub.1, .sub.2, .sub.3, .sub.4 coming from the speed sensor 6 of the wheel concerned.

(9) The tire pressure calculation modules can use various methods for calculating the pressure in the tire, notably a method designated Method1 and a method designated Method2 which is less sensitive than Method1 to the effects of the imbalances of the wheel, whilst being a priori less accurate than Method1 for a correctly balanced wheel.

(10) As long as the indicators BAL.sub.1, BAL.sub.2, BAL.sub.3, BAL.sub.4 indicate that the wheel concerned is correctly balanced, the modules for calculating the pressure criterion P.sub.1, P.sub.2, P.sub.3, P.sub.4 use the first method Method1 to calculate a respective pressure .sub.1, .sub.2, .sub.3, .sub.4 of the pressure in the tire. If the Boolean values corresponding to the balancing criteria BAL.sub.1, BAL.sub.2, BAL.sub.3, BAL.sub.4 indicate that the balancing of the wheel is not sufficient, as illustrated here for wheel 4, for example, the balancing criteria calculation module, for example P.sub.4 here, sends a negative Boolean signal to the corresponding tire pressure calculation module, which then selects a method, here Method2, that is less sensitive to the effects of the imbalance of the wheel.

(11) In the example shown in FIG. 1, the pressures of the tires of wheels 1, 2 and 3 are therefore calculated in accordance with Method1 because the wheels are correctly balanced and the pressure .sub.4 of the fourth wheel is calculated by another method, Method2, that is less accurate overall but is also less sensitive to the effects of the wheel imbalance.

(12) FIG. 2 illustrates a portion of the mode of operation of the unit 8 for evaluating the imbalance of the wheels combining the modules B.sub.1, B.sub.2, B.sub.3, B.sub.4 from FIG. 1.

(13) In FIG. 1, the modules for calculating the tire pressure criteria of the wheel can be grouped together within the same unit 9 for calculating the pressure of the wheels of the vehicle. FIG. 2 illustrates a portion of the mode of operation of one of the calculation modules B.sub.1, B.sub.2, B.sub.3, B.sub.4 of the electronic control unit 7, for example the module B.sub.1.

(14) As shown in FIG. 2, the rotation speed .sub.1 of the wheel 1, measured by one of the speed sensors 6, is transmitted to a first low-pass filter 12 that delivers a filtered wheel rotation speed value w.sub.0. This filtered speed value is sent on the one hand to a module 13 for defining parameters of a band-pass filter 14 and on the other hand to the corresponding band-pass filter 14. The filtered speed w.sub.0 is also sent to a comparator unit that calculates a filtered differentiated value {dot over (w)}.sub.0 of the speed and compares the absolute value |{dot over (w)}.sub.0| of the acceleration obtained in this way to a threshold to decide if the speed of the vehicle can be considered as stabilized or not.

(15) The definition module 13 of the filter 14 uses three values .sub.1, .sub.2, T.sub.e stored in memory units 11. These values .sub.1, .sub.2, T.sub.e are constant values and are used by the module 13 conjointly with the speed w.sub.0 of the wheel to define five filter parameters a.sub.0, a.sub.1, a.sub.2, b.sub.1, b.sub.2 that are sent to the band-pass filter 14. The band-pass filter 14 applies band-pass type filtering to the speed w.sub.0 and delivers a value .sub.band to a module 15 for calculating the variance of the speed.

(16) The speed variance module 15 then delivers a value .sub.var to a low-pass filter 16 that in turn delivers a value .sub.varfilt corresponding to a filtered variance of the speed of the wheel 1.

(17) A comparator 17 compares this value to a variance threshold threshold. If the filtered variance value .sub.varfilt is greater than the threshold, the comparator 17 delivers a Boolean value BAL, for example a negative value, to the unit 9 for calculating the inflation pressure of the tires.

(18) If the filtered variance .sub.varfilt is less than the threshold, it is an opposite Boolean value BAL, for example a positive value here, that the comparator delivers to the tire pressure monitoring module 9. Variant embodiments may be envisaged in which the meaning of the Boolean variant BAL would be inverted and the overall reasoning would remain the same.

(19) However, the Boolean value BAL is delivered only if the test 18 concerning the stabilization of the speed w.sub.0 indicates that the speed of the vehicle is actually stabilized. If the speed of the vehicle is not stabilized, the Boolean value BAL is not sent and monitoring of the speed of the wheel continues, for example by returning to the method on the upstream side of the first low-pass filter 12.

(20) In FIG. 3 two curves of a wheel speed of a vehicle as a function of time are grouped together. There are seen a darker gray curve corresponding to an unbalanced wheel speed and a lighter gray speed corresponding to a balanced wheel speed. In the example shown, the vehicle first travels at a speed close to 30 km/h before accelerating and stabilizing at a speed close to 140 km/h, and then falls again to 30 km/h. The time intervals 21 and 24 correspond to stabilized travel at 30 km/h, the time interval 22 corresponds to travel at 140 km/h and two time intervals 23 correspond to a non-stabilized travel regime for changing the speed from 30 to 140 km/h and vice versa.

(21) Note that during the plateau phases 21, 22 and 24 the speed of the unbalanced wheel features oscillations of greater amplitude than the speed of the correctly balanced wheel. The curves shown in FIG. 3 correspond to curves resulting from a first step of low-pass filtering, for example of the same type as the filtering step 12 shown in FIG. 2.

(22) FIG. 4 shows the curves from FIG. 3 after they are processed by the band-pass filter 14 from FIG. 2 and then after they are filtered by a final low-pass filter 16. The dark gray continuous curve 25 therefore represents the speed value of an unbalanced wheel after the speed sent by the sensors has passed through the low-pass filter 12 and then through the band-pass filter 14 and then through the variance calculation module 15.

(23) The light gray continuous curve 27 represents the rotation speed variance of a correctly balanced wheel after the same process of filtering the speed value delivered by the sensor associated with that wheel. The dark dashed line curve 26 is obtained by applying a low-pass filter to the curve 25 during the phases 21, 22 and 24 during which the speed 1 of the vehicle is stabilized. During the transient phases 23, the value of the curve 26 remains constant, equal to the latest preceding value calculated. The light dashed line curve is the result of analogous processing of the variance curve 27 of the balanced wheel. The curves 26 and 28 can be compared to a threshold, here chosen as equal to 0.04 km/h. This comparison corresponds to the test effected by the comparator 17 from FIG. 2 and can be applied only during the phases 21, 22 and 24 in which the speed of the vehicle is stabilized. During the transient phases 23, either the test 17 may be omitted, as in FIG. 4, or an arbitrary value may be assigned to the filtered variance function. As can be seen in FIG. 4, the curves resulting from the processing in accordance with the invention of the speed delivered by a speed sensor for each wheel may each be compared to a threshold and deliver relatively constant information making it possible to tell if the wheel is correctly balanced or not.

(24) The band-pass filter of the invention is chosen to have a very narrow pass-band. It is a discrete filter because its object is to be implemented in a vehicle computer. This pass-band is centered on the speed of the wheel, estimated on the basis of the output of the sensor associated with the wheel, to the exclusion of the sensors associated with the other wheels of the vehicle. The estimated speed of the wheel on which the pass-band is centered may be a value that has already undergone digital filtering to eliminate the noise from the signal from the sensor.

(25) The band-pass filter is an adaptive filter, i.e. the passing frequencies are offset as a function of the instantaneous rotation speed of the wheel, to remain centered on this rotation speed.

(26) The preferred band-pass filter in the context of the invention is a band-pass filter of simple structure but one incorporating parameters that may be linked to physical values measurable on the vehicle. The proposed structure is described by the following transfer function:

(27) $PB (s) = \frac{s^{2} + 2_{1} w_{0} s + w_{0}^{2}}{s^{2} + 2_{2} w_{0} s + w_{0}^{2}}$
where w.sub.0 is the passing frequency and .sub.1 and .sub.2 are the damping coefficients of the two complex conjugate poles (respectively zero).

(28) The parameters .sub.1 and .sub.2 are chosen by trial and error, for example, so that the width of the pass-band encompasses the components of the signal linked to the imbalance, at the same time as eliminating the peaks linked to phenomena other than the imbalance.

(29) The invention proposes to effect a change of variables making it possible to replace this filter with an equivalent discrete filter.

(30) The roots of the numerator/denominator are:

(31) $s = \frac{- 2_{i} w_{0} \sqrt{4_{i}^{2} w_{0}^{2} - 4 w_{0}^{2}}}{2} = -_{i} w_{0} \sqrt{_{i}^{2} w_{0}^{2} - w_{0}^{2}} = -_{i} w_{0} w_{0} \sqrt{_{i}^{2} - 1}$

(32) Moreover, it is known that by construction the roots are conjugate and complex and therefore: .sub.1.sup.21<0. This enables the roots of the numerator/denominator to be written in the form:
s=.sub.1w.sub.0w.sub.0{square root over (1.sub.1.sup.2i)}
where i is the imaginary number corresponding to the root of 1.

(33) In the invention, a discrete band-pass filter is chosen having a transfer function of the type:

(34) $PB (z^{- 1}) = \frac{a_{0} + a_{1} z^{- 1} + a_{2} z^{- 2}}{1 + b_{1} z^{- 1} + b_{2} z^{- 2}} = \frac{a_{0} z^{2} + a_{1} z^{1} + a_{2}}{z^{2} + b_{1} z^{1} + b_{2}}$
where z.sup.1 is the unitary delay operator.

(35) The band-pass filter is therefore characterized by five coefficients a0, a1, a2, b1, b2.

(36) The invention proposes to use a particular discretization method, matched discretization. This discretization method makes it possible to ensure that the passing frequencies of the discrete filter remain correctly centered on the same passing frequencies as those of the chosen continuous filter.

(37) Applying this discretization method yields:

(38) $a_{0} z^{2} + a_{1} z^{1} + a_{2} = 0 .Math. a_{0} z^{2} + \frac{a_{1}}{a_{0}} z + \frac{a_{2}}{a_{0}} = 0$
where:

(39) $z = \frac{- \frac{a_{1}}{a_{0}} \sqrt{\frac{a_{1}^{2}}{a_{0}^{2}} - \frac{4 a_{2}}{a_{0}}}}{2} = - \frac{a_{1}}{2 a_{0}} \sqrt{{(\frac{a_{1}}{2 a_{0}})}^{2} - \frac{a_{2}}{a_{0}}}$

(40) It is also known that the roots are conjugate and complex and therefore:

(41) ${(\frac{a_{1}}{2 a_{0}})}^{2} - \frac{a_{2}}{a_{0}} < 0$

(42) That is to say:

(43) $z = - \frac{a_{1}}{2 a_{0}} \sqrt{\frac{a_{2}}{a_{0}} - {(\frac{a_{1}}{2 a_{0}})}^{2} i}$

(44) And for the denominator:

(45) $z = - \frac{b_{1}}{2} \sqrt{b_{2} - {(\frac{b_{1}}{2})}^{2} i}$

(46) The poles and the zeroes of the filters are then matched:
z=e.sup.sT.sup.e
where T.sub.e is the sampling period.

(47) $- \frac{a_{1}}{2 a_{0}} \sqrt{\frac{a_{2}}{a_{0}} - {(\frac{a_{1}}{2 a_{0}})}^{2}} i = e^{T_{e} [-_{1} w_{0} w_{0} \sqrt{1 -_{1}^{2}}]} - \frac{a_{1}}{2 a_{0}} \sqrt{\frac{a_{2}}{a_{0}} - {(\frac{a_{1}}{2 a_{0}})}^{2}} i = e^{-_{1} w_{0} T_{e}} e^{w_{0} T_{e} \sqrt{1 -_{1}^{2}} i}$

(48) Four intermediate calculation variables C.sub.1, C.sub.2, C.sub.3, C.sub.4 are defined.

(49) 0 ${\begin{matrix} - \frac{a_{1}}{2 a_{0}} = e^{-_{1} w_{0} T_{e}} \cos (w_{0} T_{e} \sqrt{1 -_{1}^{2}}) = C_{1} \\ \sqrt{\frac{a_{2}}{a_{0}} - {(\frac{a_{1}}{2 a_{0}})}^{2}} = e^{-_{1} w_{0} T_{e}} \sin (w_{0} T_{e} \sqrt{1 -_{1}^{2}}) = C_{2} \end{matrix} {\begin{matrix} - \frac{a_{1}}{2} = e^{-_{2} w_{0} T_{e}} \cos (w_{0} T_{e} \sqrt{1 -_{1}^{2}}) = C_{3} \\ \sqrt{\frac{a_{2}}{1} - {(\frac{a_{1}}{2})}^{2}} = e^{-_{2} w_{0} T_{e}} \sin (w_{0} T_{e} \sqrt{1 -_{2}^{2}}) = C_{4} \end{matrix}$

(50) A supplementary condition to be respected for the discretization is that the static gain must be unitary, that is to say:
a.sub.0+a.sub.1+a.sub.2=1+b.sub.1+b.sub.2

(51) This yields a system of five equations with five unknowns the solution of which is:

(52) ${\begin{matrix} b_{1} = - 2 C_{3} \\ b_{2} = C_{4}^{2} + C_{3}^{2} \\ a_{0} = \frac{1 + C_{4}^{2} + C_{3}^{2} - 2 C_{3}}{1 + C_{2}^{2} - 2 C_{1} + C_{1}^{2}} \\ a_{1} = \frac{- 2 C_{1} (1 + C_{4}^{2} + C_{3}^{2} - 2 C_{3})}{1 + C_{2}^{2} - 2 C_{1} + C_{1}^{2}} \\ a_{2} = 1 + b_{1} + b_{2} - a_{0} - a_{1} \end{matrix}$

(53) The filter definition module 13 is configured to calculate the five coefficients a.sub.0, a.sub.1, a.sub.2, b.sub.1, b.sub.2 from the filtered speed w.sub.0 and from three constant parameters .sub.1, .sub.2, T.sub.e the last of which is correlated with a sampling period of the computer applying the band-pass filter 14.

(54) If the signals from two different wheels of the vehicle are observed at the output of the band-pass filter (these signals are shown in FIG. 3) there can very clearly be seen a difference between the signals coming from a balanced wheel (light gray) and from an unbalanced wheel (dark gray). Whereas if a signal equal to the difference between the speed signals of two wheels is looked at it is not possible to tell the status of each wheel.

(55) FIG. 3 shows a difference in terms of the variance of the signal between the balanced wheel and the unbalanced wheel. In other words, the two signals have the same mean values but the signal coming from an unbalanced wheel varies more around this common mean value.

(56) This is why the invention proposes to apply a transformation of the signal at the level of the variance calculator 15 the result of which is indicative of the variance of the signal. Multiple calculation methods may be envisaged. To make the calculation process simple and robust, the invention proposes the following transformation:
.sub.var= custom character |(.sub.band.sub.band)|
where is an averaging operator, for example a low-pass filter.

(57) In other words, a filtered mean speed is calculated by applying low-pass filtering to the filtered speed .sub.band after which the absolute value of the difference between the output .sub.band of the band-pass filter and the filtered mean speed is calculated. Numerous other formulas for quantifying a variance may be envisaged.

(58) The variable .sub.var may itself be subjected to filtering, for example low-pass type filtering 16, to obtain a filtered variance value .sub.varfilt with less noise.

(59) The value .sub.var can then be compared to a variance threshold threshold. In the variant embodiments, the value .sub.var may be further filtered, for example by a low-pass filter 16, before it is compared to the threshold. As shown in FIG. 2, the result of the comparison with the threshold is taken into account only if the rotation regime of the wheel is in a stabilized phase, for example if the absolute value of an angular acceleration of the wheel, calculated from the filtered speed w.sub.0, for example, has remained less than an acceleration threshold for a minimum time t.sub.0.

(60) The system can learn the variance threshold threshold when the vehicle moves the first few times or it may be a value stored in the computer of the vehicle during the process of manufacture of the car. In accordance with a variant embodiment, this value may be modified when the wheels are rebalanced.

(61) The invention is not limited to the example embodiments described and lends itself to numerous variants. Supplementary noise filtering steps may be introduced into the procedure or the low-pass filter 12 and/or 16 omitted. The wheel speed subjected to the filtering may be an angular speed or a value proportional to that angular speed with a constant factor of proportionality, for example the corresponding speed in km/h of the vehicle based on the theoretical diameter of the wheel.

(62) Thanks to the detection system in accordance with the invention, it is possible to detect at any time when the vehicle is moving that one of the wheels of the vehicle is no longer balanced and to adapt the wheel monitoring systems, for example a system monitoring the inflation of the tires, to limit the inaccuracies linked to this imbalance.