DE112019005731T5 - Acquisition system for a vehicle steering system that enables the torque and the absolute steering wheel angle to be measured at several revolutions - Google Patents

Abstract

The invention relates to a steering detection system which includes: - a first angular position sensor (11) which measures the angle of the electric motor (3) and delivers a first signal, - a second angular position sensor (12) which measures the steering angle between the reduction gear (4) and a first side of the torsion bar (7) and supplies a second signal, - a third angular position sensor (13) which measures the steering angle, which is located on a second side of the torsion bar (7) and supplies a third signal , - and a processing unit (15), on the one hand, a Vernier calculation on the basis of the first and the second signal to generate a first calculated signal (see "Signal 1") that is proportional to the absolute steering wheel angle over more than one mechanical rotation (θ̂2), and on the other hand a weighted angular sum of the third signal and one of the signals from the first signal, the second signal and the er The first calculated signal (see "Signal 1") is caused to generate a second calculated signal (see "Signal 2") that is proportional to a torque (T).

Description

The present invention relates generally to the technical field of detection systems for measuring the torque and the absolute steering wheel angle during several turns of a steering of a vehicle.

A preferred application of the invention is aimed at detection systems for measuring the torque and the absolute steering wheel angle during several revolutions of an electric power steering system of a vehicle.

Another application of the invention is aimed at detection systems for measuring the torque and the absolute steering wheel angle during several revolutions of a steering system of a vehicle in which the steering wheel is mechanically separated from the wheels.

Classically, an electric power steering includes an electric motor equipped with a reduction gear that exerts a supporting torque on the steering of the vehicle, namely the steering column or the steering rack. Operating an electric power steering requires knowing the intensity of the torque applied to the steering and the steering wheel or steering angle.

In general, the steering wheel of a vehicle is designed to turn left and right from a neutral point, approximately one and a half turns. In other words, the steering wheel can allow rotation of approximately three turns from a left end to a right end. Consequently, an electric power steering system must be equipped with a steering angle sensor which is able to detect an angular range greater than or equal to three revolutions (3 x 360 °) in order to detect this steering angle in an appropriate manner. In addition, the information on the steering wheel angle is also required for the operation of other functions of the vehicle, such as electronic lane correction, or for the new auxiliary functions when driving a car.

The patent FR 2 964 190 from the company MMT proposes a magnetic device for detecting the absolute steering wheel angle at several revolutions, which in particular converts magnets and magnetically sensitive probes. This detection device also requires motion conversion systems (gears), resulting in an expensive detection device.

The patent application EP 3 090 921 of NSK describes a detection device of the steering angle of a vehicle which includes a vernier calculation section that effects vernier calculation based on an angle of the steering shaft and an angle of the electric servomotor shaft. This device also implements a neutral period determining section including a neutral point based on a reference angle calculated by Vernier calculation and a neutral point specification section including the neutral point from the neutral period and a stored one The value of the neutral point is specified. The device enables the steering angle to be determined with several revolutions only after a learning step, so that this device does not make it possible to know the steering angle from the start.

The patent FR 2 872 896 from MMT describes a torque sensor that uses the angular deformation of a torsion bar of known stiffness. This differential angular deformation sensor includes multiple magnetic concentrators, a magnetized target, and a Hall effect detection probe. This sensor has a complex design and only measures the angle between the input shaft and the output shaft.

It does not give any information about the angle of the input and output shafts in relation to the chassis of the vehicle.

The patent application DE 102009039764 from the company BMW describes a detection system for measuring the torque and the absolute steering wheel angle during several revolutions of a steering system of a vehicle for which the steering wheel is mechanically separated from the wheels. Such a steering system, known as “steer by wire”, includes an electrical or hydraulic connection between the steering wheel and the wheels. An electric motor is equipped with a reduction gear that provides drag torque for the steering. This document proposes to measure the steering wheel angle with the angular position sensor, which measures the angle of the electric motor, but also with an additional probe with Hall effect, which detects the leakage magnetic flux of the magnetized target of the torque sensor (des Type MMT) and is located on the side of the reduction gear of the torsion bar. The use of a classic torque sensor leads to an expensive acquisition system and the robustness of a position sensor that measures a magnetic leakage flux has yet to be demonstrated.

The present invention seeks to remedy the disadvantages of the prior art by proposing a new detection system for a steering of a vehicle that enables the measurement of the torque, but also the absolute steering wheel angle, over more than one mechanical revolution, without the need to add complementary gears and without the need to implement a learning phase, this detection system incorporating position sensors which are simpler and less expensive than the prior art.

• - einen ersten Winkelpositionssensor, der N1 Polpaare besitzt, wobei N1 eine ganze Zahl größer oder gleich 1 ist, wobei dieser erste Winkelpositionssensor den Winkel des Elektromotors misst und ein erstes Signal liefert,
• - einen zweiten Winkelpositionssensor, der N2 Polpaare besitzt, wobei N2 eine ganze Zahl größer oder gleich 1 ist, wobei dieser zweite Winkelpositionssensor den Winkel der Lenkung misst, die sich zwischen dem Untersetzungsgetriebe und einer ersten Seite des Torsionsstabs befindet, wobei dieser zweite Winkelpositionssensor ein zweites Signal liefert,
• - einen dritten Winkelpositionssensor, der N3 Polpaare besitzt, wobei N3 eine ganze Zahl größer oder gleich 1 ist, wobei dieser dritte Winkelpositionssensor den Winkel der Lenkung misst, die sich auf einer zweiten Seite des Torsionsstabs befindet, die sich gegenüber der ersten Seite befindet, und ein drittes Signal liefert,
• - und eine Verarbeitungseinheit, die einerseits eine Vernier-Berechnung auf der Grundlage von zumindest dem ersten Signal und dem zweiten Signal, um ein erstes berechnetes Signal zu erzeugen, das proportional zu dem absoluten Lenkradwinkel über mehr als eine mechanische Umdrehung ist, und andererseits eine gewichtete Winkelsumme des dritten Signals und von einem der Signale von dem ersten Signal, dem zweiten Signal und dem ersten berechneten Signal bewirkt, um ein zweites berechnetes Signal zu erzeugen, das proportional zu dem Drehmoment ist.

In order to achieve such an aim, the subject matter of the invention relates to a new detection system for a steering system of a vehicle which enables the torque and the absolute steering wheel angle to be measured over more than one mechanical revolution, this steering system comprising a torsion bar and being equipped with an electric motor which is provided with a reduction gear. According to the invention, this detection system includes:

• a first angular position sensor which has N1 pole pairs, where N1 is an integer greater than or equal to 1, this first angular position sensor measuring the angle of the electric motor and delivering a first signal,
• - A second angular position sensor having N2 pole pairs, where N2 is an integer greater than or equal to 1, this second angular position sensor measuring the angle of the steering that is located between the reduction gear and a first side of the torsion bar, this second angular position sensor having a second Signal delivers,
• a third angular position sensor having N3 pole pairs, where N3 is an integer greater than or equal to 1, this third angular position sensor measuring the angle of the steering located on a second side of the torsion bar opposite the first side, and supplies a third signal,
• - and a processing unit, on the one hand a Vernier calculation on the basis of at least the first signal and the second signal in order to generate a first calculated signal which is proportional to the absolute steering wheel angle over more than one mechanical revolution, and on the other hand a weighted one Angular sum of the third signal and one of the signals from the first signal, the second signal, and the first calculated signal to produce a second calculated signal proportional to the torque.

• - die Verarbeitungseinheit geht davon aus, dass das erste berechnete Signal dem absoluten Lenkradwinkel über mehr als eine mechanische Umdrehung entspricht und dass das zweite berechnete Signal dem ausgeübten Drehmoment entspricht,
• - die Verarbeitungseinheit bewirkt die Vernier-Berechnung, um das erste berechnete Signal zu erzeugen, die lautet:

$${\widehat{\theta }}_2=N_{turns}\cdot f_s\left(\mathrm{\Theta }-{\mathrm{<figure-callout\; id="1"\; label="q"\; filenames="DE112019005731T5\_0084.png"\; state="\{\{state\}\}">q</figure-callout>}}_1\cdot f_s\left(N_1{R}_{red}{N}_{turns}\mathrm{\Theta }-{\alpha }_1\right)-{\mathrm{<figure-callout\; id="2"\; label="q"\; filenames="DE112019005731T5\_0084.png"\; state="\{\{state\}\}">q</figure-callout>}}_2\cdot f_s\left(N_2{N}_{turns}\mathrm{\Theta }-{\alpha }_2\right)\right)$$

wobei: θ = f_s(p₁α₁ +p₂α₂)
und f_s die mathematische Funktion in Sägezahnform und mit Steigung gleich 1 ist, wobei q₁ und q₂ feste ausgewählte Gewichtungskoeffizienten sind, und α₁ das Signal ist, das durch den ersten Winkelpositionssensor erzeugt wird, α₂ das Signal ist, das durch den zweiten Winkelpositionssensor erzeugt wird,
und wobei p₁, p₂, und N_turns numerische Koeffizienten sind, die ausgewählt sind und die folgenden Bedingungen erfüllen müssen: $$\{\begin{array}{c}{p}_{1}und{p}_{2}relativeganzeZahlen\\ \left(R_{red}{N}_1{\mathrm{<figure-callout\; id="1"\; label="p"\; filenames="DE112019005731T5\_0084.png"\; state="\{\{state\}\}">p</figure-callout>}}_1+N_2{p}_2\right)\cdot N_{turns}=1\\ N_{turns}\ge \frac{\mathrm{\Delta }{\theta }_2}{360}\end{array}$$

wobei Δθ₂ die Spitze-zu-Spitze-Variation des absoluten Lenkradwinkels über mehr als eine mechanische Umdrehung ist und R_red das Untersetzungsverhältnis des Untersetzungsgetriebes ist,

• - die Gewichtungskoeffizienten q₁ und q₂ vorteilhafterweise ausgewählt sein können gleich:

$$\begin{array}{l}we\mathrm{<figure-callout\; id="2"\; label="n\; n\; N"\; filenames="DE112019005731T5\_0084.png"\; state="\{\{state\}\}">n</figure-callout>}n{N}_{}{}_2<N_1{R}_{red}\\ \{\begin{array}{c}{\mathrm{<figure-callout\; id="1"\; label="q"\; filenames="DE112019005731T5\_0084.png"\; state="\{\{state\}\}">q</figure-callout>}}_1=0\\ {\mathrm{<figure-callout\; id="2"\; label="q"\; filenames="DE112019005731T5\_0084.png"\; state="\{\{state\}\}">q</figure-callout>}}_2=\frac{1}{p_1{N}_{turns}}\cdot \frac{{\mathrm{<figure-callout\; id="1"\; label="\sigma "\; filenames="DE112019005731T5\_0084.png"\; state="\{\{state\}\}">\sigma </figure-callout>}}_1^2{p}_1{\mathrm{<figure-callout\; id="2"\; label="N"\; filenames="DE112019005731T5\_0084.png"\; state="\{\{state\}\}">N</figure-callout>}}_2-{\mathrm{<figure-callout\; id="2"\; label="\sigma "\; filenames="DE112019005731T5\_0084.png"\; state="\{\{state\}\}">\sigma </figure-callout>}}_2^2{p}_2{N}_1{R}_{red}}{{\left({\sigma }_1{N}_2\right)}^2+{\left({\sigma }_2{N}_1{R}_{red}\right)}^2}\end{array}\end{array}$$

$$\begin{array}{l}we\mathrm{<figure-callout\; id="2"\; label="n\; n\; N"\; filenames="DE112019005731T5\_0084.png"\; state="\{\{state\}\}">n</figure-callout>}n{N}_{}{}_2>N_1{R}_{red}\\ \{\begin{array}{c}{\mathrm{<figure-callout\; id="1"\; label="q"\; filenames="DE112019005731T5\_0084.png"\; state="\{\{state\}\}">q</figure-callout>}}_1=\frac{1}{p_2{N}_{turns}}\cdot \frac{{\mathrm{<figure-callout\; id="2"\; label="\sigma "\; filenames="DE112019005731T5\_0084.png"\; state="\{\{state\}\}">\sigma </figure-callout>}}_2^2{p}_2{N}_1{R}_{red}-{\mathrm{<figure-callout\; id="1"\; label="\sigma "\; filenames="DE112019005731T5\_0084.png"\; state="\{\{state\}\}">\sigma </figure-callout>}}_1^2{p}_1{N}_2}{{\left({\sigma }_1{N}_2\right)}^2+{\left({\sigma }_2{N}_1{R}_{red}\right)}^2}\\ {\mathrm{<figure-callout\; id="2"\; label="q"\; filenames="DE112019005731T5\_0084.png"\; state="\{\{state\}\}">q</figure-callout>}}_2=0\end{array}\end{array}$$

wobei σ₁ die typische Fehleramplitude des ersten Winkelpositionssensors ist und wobei σ2 die typische Fehleramplitude des zweiten Winkelpositionssensors ist.The device according to the invention also includes in combination one and / or more of the following additional properties:

• the processing unit assumes that the first calculated signal corresponds to the absolute steering wheel angle over more than one mechanical revolution and that the second calculated signal corresponds to the torque exerted,
• the processing unit carries out the vernier calculation in order to generate the first calculated signal which reads:

$${\widehat{\theta }}_2=N_{turns}\cdot f_s\left(\mathrm{\Theta }-{\mathrm{<figure-callout\; id="1"\; label="q"\; filenames="DE112019005731T5\_0084.png"\; state="\{\{state\}\}">q</figure-callout>}}_1\cdot f_s\left({\mathrm{<figure-callout\; id="1"\; label="N"\; filenames="DE112019005731T5\_0084.png"\; state="\{\{state\}\}">N</figure-callout>}}_1{\mathrm{R.}}_{red}{N}_{turns}\mathrm{\Theta }-{\alpha }_1\right)-{\mathrm{<figure-callout\; id="2"\; label="q"\; filenames="DE112019005731T5\_0084.png"\; state="\{\{state\}\}">q</figure-callout>}}_2\cdot f_s\left(N_2{N}_{turns}\mathrm{\Theta }-{\alpha }_2\right)\right)$$

where: θ = f _s (p ₁ α ₁ + p ₂ α ₂ )
and f _{s is} the mathematical function in sawtooth form and with a slope equal to 1, where q ₁ and q _{2 are} fixed selected weighting coefficients, and α _{1 is} the signal generated by the first angular position sensor, α _{2 is} the signal generated by the second angular position sensor is generated,
and where p ₁ , p ₂ , and N _{turns are} numeric coefficients that are selected and must meet the following conditions: $$\{\begin{array}{c}{p}_{1}und{p}_{2}relativeGanzeZaHlen\\ \left({\mathrm{R.}}_{red}{N}_1{\mathrm{<figure-callout\; id="1"\; label="p"\; filenames="DE112019005731T5\_0084.png"\; state="\{\{state\}\}">p</figure-callout>}}_1+N_2{p}_2\right)\cdot N_{turns}=1\\ N_{turns}\ge \frac{\mathrm{<figure-callout\; id="2"\; label="\Delta \; \theta "\; filenames="DE112019005731T5\_0084.png"\; state="\{\{state\}\}">\Delta </figure-callout>}{\theta }_{}{}_2}{360}\end{array}$$

where Δθ _{2 is} the peak-to-peak variation of the absolute steering wheel angle over more than one mechanical revolution and R _{red is} the reduction ratio of the reduction gear,

• the weighting coefficients q ₁ and q _{2 can} advantageously be selected equal to:

$$\begin{array}{l}we\mathrm{<figure-callout\; id="2"\; label="n\; n\; N"\; filenames="DE112019005731T5\_0084.png"\; state="\{\{state\}\}">n</figure-callout>}n{N}_{}{}_2<{\mathrm{<figure-callout\; id="1"\; label="N"\; filenames="DE112019005731T5\_0084.png"\; state="\{\{state\}\}">N</figure-callout>}}_1{\mathrm{R.}}_{red}\\ \{\begin{array}{c}{\mathrm{<figure-callout\; id="1"\; label="q"\; filenames="DE112019005731T5\_0084.png"\; state="\{\{state\}\}">q</figure-callout>}}_1=0\\ {\mathrm{<figure-callout\; id="2"\; label="q"\; filenames="DE112019005731T5\_0084.png"\; state="\{\{state\}\}">q</figure-callout>}}_2=\frac{1}{p_1{N}_{turns}}\cdot \frac{{\mathrm{<figure-callout\; id="1"\; label="\sigma "\; filenames="DE112019005731T5\_0084.png"\; state="\{\{state\}\}">\sigma </figure-callout>}}_1^2{p}_1{N}_2-{\mathrm{<figure-callout\; id="2"\; label="\sigma "\; filenames="DE112019005731T5\_0084.png"\; state="\{\{state\}\}">\sigma </figure-callout>}}_2^2{p}_2{\mathrm{<figure-callout\; id="1"\; label="N"\; filenames="DE112019005731T5\_0084.png"\; state="\{\{state\}\}">N</figure-callout>}}_1{\mathrm{R.}}_{red}}{{\left({\sigma }_1{N}_2\right)}^2+{\left({\sigma }_2{\mathrm{<figure-callout\; id="1"\; label="N"\; filenames="DE112019005731T5\_0084.png"\; state="\{\{state\}\}">N</figure-callout>}}_1{\mathrm{R.}}_{red}\right)}^2}\end{array}\end{array}$$

$$\begin{array}{l}we\mathrm{<figure-callout\; id="2"\; label="n\; n\; N"\; filenames="DE112019005731T5\_0084.png"\; state="\{\{state\}\}">n</figure-callout>}n{N}_{}{}_2>{\mathrm{<figure-callout\; id="1"\; label="N"\; filenames="DE112019005731T5\_0084.png"\; state="\{\{state\}\}">N</figure-callout>}}_1{\mathrm{R.}}_{red}\\ \{\begin{array}{c}{\mathrm{<figure-callout\; id="1"\; label="q"\; filenames="DE112019005731T5\_0084.png"\; state="\{\{state\}\}">q</figure-callout>}}_1=\frac{1}{p_2{N}_{turns}}\cdot \frac{{\mathrm{<figure-callout\; id="2"\; label="\sigma "\; filenames="DE112019005731T5\_0084.png"\; state="\{\{state\}\}">\sigma </figure-callout>}}_2^2{p}_2{\mathrm{<figure-callout\; id="1"\; label="N"\; filenames="DE112019005731T5\_0084.png"\; state="\{\{state\}\}">N</figure-callout>}}_1{\mathrm{R.}}_{red}-{\mathrm{<figure-callout\; id="1"\; label="\sigma "\; filenames="DE112019005731T5\_0084.png"\; state="\{\{state\}\}">\sigma </figure-callout>}}_1^2{p}_1{N}_2}{{\left({\sigma }_1{N}_2\right)}^2+{\left({\sigma }_2{\mathrm{<figure-callout\; id="1"\; label="N"\; filenames="DE112019005731T5\_0084.png"\; state="\{\{state\}\}">N</figure-callout>}}_1{\mathrm{R.}}_{red}\right)}^2}\\ {\mathrm{<figure-callout\; id="2"\; label="q"\; filenames="DE112019005731T5\_0084.png"\; state="\{\{state\}\}">q</figure-callout>}}_2=0\end{array}\end{array}$$

where σ _{1 is} the typical error amplitude of the first angular position sensor and where σ2 is the typical error amplitude of the second angular position sensor.

wobei:

• - $${\mathrm{\Theta }}^{*}=f_s\left({\mathrm{<figure-callout\; id="1"\; label="p"\; filenames="DE112019005731T5\_0084.png"\; state="\{\{state\}\}">p</figure-callout>}}_1^{*}{\mathrm{<figure-callout\; id="1"\; label="\alpha "\; filenames="DE112019005731T5\_0084.png"\; state="\{\{state\}\}">\alpha </figure-callout>}}_1+{\mathrm{<figure-callout\; id="2"\; label="p"\; filenames="DE112019005731T5\_0084.png"\; state="\{\{state\}\}">p</figure-callout>}}_2^{*}{\mathrm{<figure-callout\; id="2"\; label="\alpha "\; filenames="DE112019005731T5\_0084.png"\; state="\{\{state\}\}">\alpha </figure-callout>}}_2^{*}\right)$$
  
• - $${\mathrm{<figure-callout\; id="2"\; label="\alpha "\; filenames="DE112019005731T5\_0084.png"\; state="\{\{state\}\}">\alpha </figure-callout>}}_2^{*}=f_s\left(f_s\left({\mathrm{<figure-callout\; id="2"\; label="c"\; filenames="DE112019005731T5\_0084.png"\; state="\{\{state\}\}">c</figure-callout>}}_2\cdot f_s\left({\mathrm{<figure-callout\; id="2"\; label="\alpha "\; filenames="DE112019005731T5\_0084.png"\; state="\{\{state\}\}">\alpha </figure-callout>}}_2+{\mathrm{<figure-callout\; id="3"\; label="n"\; filenames="DE112019005731T5\_0084.png"\; state="\{\{state\}\}">n</figure-callout>}}_3\delta \right)+{\mathrm{<figure-callout\; id="3"\; label="c"\; filenames="DE112019005731T5\_0084.png"\; state="\{\{state\}\}">c</figure-callout>}}_3\cdot f_s\left({\mathrm{<figure-callout\; id="3"\; label="\alpha "\; filenames="DE112019005731T5\_0084.png"\; state="\{\{state\}\}">\alpha </figure-callout>}}_3-{\mathrm{<figure-callout\; id="2"\; label="n"\; filenames="DE112019005731T5\_0084.png"\; state="\{\{state\}\}">n</figure-callout>}}_2\delta \right)\right)-\frac{n_3}{{\mathrm{<figure-callout\; id="2"\; label="n"\; filenames="DE112019005731T5\_0084.png"\; state="\{\{state\}\}">n</figure-callout>}}_2}\delta \right)$$
  
• - $$\delta =\frac{f_s\left(-{\mathrm{<figure-callout\; id="3"\; label="n"\; filenames="DE112019005731T5\_0084.png"\; state="\{\{state\}\}">n</figure-callout>}}_3{\mathrm{<figure-callout\; id="2"\; label="\alpha "\; filenames="DE112019005731T5\_0084.png"\; state="\{\{state\}\}">\alpha </figure-callout>}}_2+{\mathrm{<figure-callout\; id="2"\; label="n"\; filenames="DE112019005731T5\_0084.png"\; state="\{\{state\}\}">n</figure-callout>}}_2{\alpha }_3\right)}{{\mathrm{<figure-callout\; id="2"\; label="n"\; filenames="DE112019005731T5\_0084.png"\; state="\{\{state\}\}">n</figure-callout>}}_2^2+{\mathrm{<figure-callout\; id="3"\; label="n"\; filenames="DE112019005731T5\_0084.png"\; state="\{\{state\}\}">n</figure-callout>}}_3^2}$$
  
• - $$\{\begin{array}{c}{\mathrm{<figure-callout\; id="23"\; label="N"\; filenames="DE112019005731T5\_0088.png"\; state="\{\{state\}\}">N</figure-callout>}}_{23}=pgcd\left({\mathrm{<figure-callout\; id="2"\; label="N"\; filenames="DE112019005731T5\_0084.png"\; state="\{\{state\}\}">N</figure-callout>}}_2,N_3\right)\\ {\mathrm{<figure-callout\; id="2"\; label="n"\; filenames="DE112019005731T5\_0084.png"\; state="\{\{state\}\}">n</figure-callout>}}_2=\frac{N_2}{N_{23}}\\ {\mathrm{<figure-callout\; id="3"\; label="n"\; filenames="DE112019005731T5\_0084.png"\; state="\{\{state\}\}">n</figure-callout>}}_3=\frac{N_3}{{\mathrm{<figure-callout\; id="23"\; label="N"\; filenames="DE112019005731T5\_0088.png"\; state="\{\{state\}\}">N</figure-callout>}}_{23}}\end{array}$$
  

und f_s die mathematische Funktion in Sägezahnform und mit Steigung gleich 1 ist, wobei pgcd der Operator „größter gemeinsamer Teiler“ ist, wobei $q_{}^{}$${}_1^{*}$

und $q_{}^{}$${}_2^{*}$

feste ausgewählte Gewichtungskoeffizienten sind, und a₁ das Signal ist, das durch den ersten Winkelpositionssensor erzeugt wird, α₂ das Signal ist, das durch den zweiten Winkelpositionssensor erzeugt wird, α₃ das Signal ist, das durch den dritten Winkelpositionssensor erzeugt wird.According to a further variant embodiment, the processing unit carries out the vernier calculation in order _{to generate the first calculated signal (θ̂ 2} ), which reads: $${\widehat{\theta }}_2=N_{{}_{turns}}^{*}\cdot f_s\left({\mathrm{\Theta }}^{\mathrm{\ast }}-{\mathrm{<figure-callout\; id="1"\; label="q"\; filenames="DE112019005731T5\_0084.png"\; state="\{\{state\}\}">q</figure-callout>}}_1^{*}\cdot f_s\left({\mathrm{<figure-callout\; id="1"\; label="N"\; filenames="DE112019005731T5\_0084.png"\; state="\{\{state\}\}">N</figure-callout>}}_1{\mathrm{R.}}_{red}{N}_{turns}^{*}{\mathrm{\Theta }}^{\mathrm{\ast }}-{\alpha }_1\right)-{\mathrm{<figure-callout\; id="2"\; label="q"\; filenames="DE112019005731T5\_0084.png"\; state="\{\{state\}\}">q</figure-callout>}}_2^{*}\cdot f_s\left(N_{23}{N}_{turn}^{*}{\mathrm{\Theta }}^{\mathrm{\ast }}-{\mathrm{<figure-callout\; id="2"\; label="\alpha "\; filenames="DE112019005731T5\_0084.png"\; state="\{\{state\}\}">\alpha </figure-callout>}}_2^{*}\right)\right)$$

whereby:

• - $${\mathrm{\Theta }}^{*}=f_s\left(p_1^{*}{\alpha }_1+p_2^{*}{\alpha }_2^{*}\right)$$
  
• - $${\alpha }_2^{*}=f_s\left(f_s\left(c_2\cdot f_s\left({\alpha }_2+n_3\delta \right)+c_3\cdot f_s\left({\alpha }_3-n_2\delta \right)\right)-\frac{n_3}{n_2}\delta \right)$$
  
• - $$\delta =\frac{f_s\left(-n_3{\alpha }_2+n_2{\alpha }_3\right)}{n_2^2+n_3^2}$$
  
• - $$\{\begin{array}{c}{N}_{23}=pGcd\left(N_2,N_3\right)\\ n_2=\frac{N_2}{N_{23}}\\ n_3=\frac{N_3}{N_{23}}\end{array}$$
  

and f _{s is} the mathematical function in sawtooth form and with a slope equal to 1, where pgcd is the operator "greatest common divisor", where ${\mathrm{<figure-callout\; id="1"\; label="q"\; filenames="DE112019005731T5\_0084.png"\; state="\{\{state\}\}">q</figure-callout>}}_1^{*}$

and ${\mathrm{<figure-callout\; id="2"\; label="q"\; filenames="DE112019005731T5\_0084.png"\; state="\{\{state\}\}">q</figure-callout>}}_2^{*}$

are fixed selected weighting coefficients, and a _{1 is} the signal generated by the first angular position sensor, α _{2 is} the signal generated by the second angular position sensor, α _{3 is} the signal generated by the third angular position sensor.

$p_2^{*}$

und $N_{turns}^{*}$

numerische Koeffizienten sind, die ausgewählt sind und die folgenden Bedingungen erfüllen müssen: $$\{\begin{array}{c}\begin{array}{l}{c}_{2}und{c}_{3}relativeganzeZahlen\\ p{*}_{1}undp{*}_{2}relativeganzeZahlen\end{array}\\ c_2{n}_2+c_3{n}_3=1\\ \left(R_{red}{N}_1{p}_1^{*}+N_{23}{p}_2^{*}\right)\cdot N_{turns}^{*}=1\\ N_{turns}^{*}\ge \frac{\mathrm{\Delta }{\theta }_2}{360}\end{array}$$

und wobei Δθ₂ die Spitze-zu-Spitze-Variation des absoluten Lenkradwinkels über mehr als eine mechanische Umdrehung ist und R_red das Untersetzungsverhältnis des Untersetzungsgetriebes ist,

• - gemäß einer ersten Ausführungsvariante lautet das zweite berechnete Signal:

$$\widehat{T}=\frac{G}{\left|k_3{\mathrm{<figure-callout\; id="3"\; label="N"\; filenames="DE112019005731T5\_0084.png"\; state="\{\{state\}\}">N</figure-callout>}}_3\right|}\cdot f_s\left({\mathrm{<figure-callout\; id="2"\; label="k"\; filenames="DE112019005731T5\_0084.png"\; state="\{\{state\}\}">k</figure-callout>}}_2{\mathrm{<figure-callout\; id="2"\; label="\alpha "\; filenames="DE112019005731T5\_0084.png"\; state="\{\{state\}\}">\alpha </figure-callout>}}_2+{\mathrm{<figure-callout\; id="3"\; label="k"\; filenames="DE112019005731T5\_0084.png"\; state="\{\{state\}\}">k</figure-callout>}}_3{\alpha }_3\right)$$

wobei f_s die mathematische Funktion in Sägezahnform und mit Steigung gleich 1 ist, α₂ das Signal ist, das durch den zweiten Winkelpositionssensor erzeugt wird, α₃ das Signal ist, das durch den dritten Winkelpositionssensor erzeugt wird, und G die Steifigkeit des Torsionsstabs ist, und wobei k₂, k₃ numerische Koeffizienten sind, die ausgewählt sind und die folgenden Bedingungen erfüllen müssen: $$\{\begin{array}{l}{k}_{2}und{k}_{3}relativeganzeZahlen\hfill \\ N_2{\mathrm{<figure-callout\; id="2"\; label="k"\; filenames="DE112019005731T5\_0084.png"\; state="\{\{state\}\}">k</figure-callout>}}_2+N_3{\mathrm{<figure-callout\; id="3"\; label="k"\; filenames="DE112019005731T5\_0084.png"\; state="\{\{state\}\}">k</figure-callout>}}_3=0\hfill \\ \frac{360}{\left|k_3{\mathrm{<figure-callout\; id="3"\; label="N"\; filenames="DE112019005731T5\_0084.png"\; state="\{\{state\}\}">N</figure-callout>}}_3\right|}>\mathrm{\Delta }{\theta }_{shift}\hfill \end{array}$$

wobei Δθ_shift die Spitze-zu-Spitze-Variation der Winkelverformung des Torsionsstabs ist,

• - gemäß einer zweiten Ausführungsvariante lautet das zweite berechnete Signal:

$$\widehat{T}=\frac{G}{\left|k_3{\mathrm{<figure-callout\; id="3"\; label="N"\; filenames="DE112019005731T5\_0084.png"\; state="\{\{state\}\}">N</figure-callout>}}_3\right|}\cdot f_s\left({\mathrm{<figure-callout\; id="1"\; label="k"\; filenames="DE112019005731T5\_0084.png"\; state="\{\{state\}\}">k</figure-callout>}}_1{\mathrm{<figure-callout\; id="1"\; label="\alpha "\; filenames="DE112019005731T5\_0084.png"\; state="\{\{state\}\}">\alpha </figure-callout>}}_1+{\mathrm{<figure-callout\; id="3"\; label="k"\; filenames="DE112019005731T5\_0084.png"\; state="\{\{state\}\}">k</figure-callout>}}_3{\alpha }_3\right)$$

wobei f_s die mathematische Funktion in Sägezahnform und mit Steigung gleich 1 ist, α₁ das Signal ist, das durch den ersten Winkelpositionssensor erzeugt wird, α₃ das Signal ist, das durch den dritten Winkelpositionssensor erzeugt wird, und G die Steifigkeit des Torsionsstabs ist, und wobei k₁, k₃ numerische Koeffizienten sind, die ausgewählt sind und die folgenden Bedingungen erfüllen müssen: $$\{\begin{array}{l}{k}_{2}und{k}_{3}relativeganzeZahlen\hfill \\ R_{red}{N}_1{\mathrm{<figure-callout\; id="1"\; label="k"\; filenames="DE112019005731T5\_0084.png"\; state="\{\{state\}\}">k</figure-callout>}}_1+N_3{\mathrm{<figure-callout\; id="3"\; label="k"\; filenames="DE112019005731T5\_0084.png"\; state="\{\{state\}\}">k</figure-callout>}}_3=0\hfill \\ \frac{360}{\left|k_3{\mathrm{<figure-callout\; id="3"\; label="N"\; filenames="DE112019005731T5\_0084.png"\; state="\{\{state\}\}">N</figure-callout>}}_3\right|}>\mathrm{\Delta }{\theta }_{shift}\hfill \end{array}$$

wobei Δθ_Shift die Spitze-zu-Spitze-Variation der Winkelverformung des Torsionsstabs ist,

• - gemäß einer dritten Ausführungsvariante lautet das zweite berechnete Signal:

$$\widehat{T}=\frac{G}{\left|k_3{\mathrm{<figure-callout\; id="3"\; label="N"\; filenames="DE112019005731T5\_0084.png"\; state="\{\{state\}\}">N</figure-callout>}}_3\right|}\cdot f_s\left({\mathrm{<figure-callout\; id="3"\; label="k"\; filenames="DE112019005731T5\_0084.png"\; state="\{\{state\}\}">k</figure-callout>}}_3{\mathrm{<figure-callout\; id="3"\; label="\alpha "\; filenames="DE112019005731T5\_0084.png"\; state="\{\{state\}\}">\alpha </figure-callout>}}_3{\widehat{\theta }}_2+{\mathrm{<figure-callout\; id="3"\; label="k"\; filenames="DE112019005731T5\_0084.png"\; state="\{\{state\}\}">k</figure-callout>}}_3{\alpha }_3\right)$$

wobei f_s die mathematische Funktion in Sägezahnform und mit Steigung gleich 1 ist, θ̂₂ das erste berechnete Signal ist, a₃ das Signal ist, das durch den dritten Winkelpositionssensor erzeugt wird, und G die Steifigkeit des Torsionsstabs ist, und wobei k₃ ein numerischer Koeffizient ist, der ausgewählt ist und die folgenden Bedingungen erfüllen muss: $$\{\begin{array}{l}{k}_{3}relativeganzeZahl\hfill \\ \frac{360}{\left|k_3{\mathrm{<figure-callout\; id="3"\; label="N"\; filenames="DE112019005731T5\_0084.png"\; state="\{\{state\}\}">N</figure-callout>}}_3\right|}>\mathrm{\Delta }{\theta }_{shift}\hfill \end{array}$$

wobei Δθ_shift die Spitze-zu-Spitze-Variation der Winkelverformung des Torsionsstabs ist,

• - die Verarbeitungseinheit verifiziert, dass die Gesamtheit der gemessenen und berechneten Signale einer Gesamtheit an zulässigen Werten angehört, wobei die Verarbeitungseinheit ein Warnsignal liefert, wenn die Gesamtheit an Werten keiner Gesamtheit an zulässigen Werten angehört,
• - der erste Winkelpositionssensor, der zweite Winkelpositionssensor und/oder der dritte Winkelpositionssensor sind Sensoren oder Kombinationen von Positionssensoren mit Hall-Effekt, Magnetwiderstand, Fluxgates, Induktion, mit Wirbelstrom oder mit variabler Reluktanz,
• - der zweite Winkelpositionssensor und der dritte Winkelpositionssensor sind Positionssensoren mit Wirbelstrom, die eine gemeinsame Erfassungssonde beinhalten, die eine gemeinsame Stützplatte für die Wicklungen des zweiten Winkelpositionssensors und des dritten Winkelpositionssensors umfasst.

And where c ₂ , c ₃ , ${\mathrm{<figure-callout\; id="1"\; label="p"\; filenames="DE112019005731T5\_0084.png"\; state="\{\{state\}\}">p</figure-callout>}}_1^{*},$

${\mathrm{<figure-callout\; id="2"\; label="p"\; filenames="DE112019005731T5\_0084.png"\; state="\{\{state\}\}">p</figure-callout>}}_2^{*}$

and $N_{turns}^{*}$

are numerical coefficients that are selected and must meet the following conditions: $$\{\begin{array}{c}\begin{array}{l}{c}_{2}und{c}_{3}relativeGanzeZaHlen\\ p{*}_{1}undp{*}_{2}relativeGanzeZaHlen\end{array}\\ c_2{\mathrm{<figure-callout\; id="2"\; label="n"\; filenames="DE112019005731T5\_0084.png"\; state="\{\{state\}\}">n</figure-callout>}}_2+c_3{\mathrm{<figure-callout\; id="3"\; label="n"\; filenames="DE112019005731T5\_0084.png"\; state="\{\{state\}\}">n</figure-callout>}}_3=1\\ \left({\mathrm{R.}}_{red}{N}_1{\mathrm{<figure-callout\; id="1"\; label="p"\; filenames="DE112019005731T5\_0084.png"\; state="\{\{state\}\}">p</figure-callout>}}_1^{*}+N_{23}{\mathrm{<figure-callout\; id="2"\; label="p"\; filenames="DE112019005731T5\_0084.png"\; state="\{\{state\}\}">p</figure-callout>}}_2^{*}\right)\cdot N_{turns}^{*}=1\\ N_{turns}^{*}\ge \frac{\mathrm{<figure-callout\; id="2"\; label="\Delta \; \theta "\; filenames="DE112019005731T5\_0084.png"\; state="\{\{state\}\}">\Delta </figure-callout>}{\theta }_{}{}_2}{360}\end{array}$$

and where Δθ _{2 is} the peak-to-peak variation of the absolute steering wheel angle over more than one mechanical revolution and R _{red is} the reduction gear ratio,

• - According to a first variant, the second calculated signal reads:

$$\widehat{T}=\frac{G}{\left|k_3{\mathrm{<figure-callout\; id="3"\; label="N"\; filenames="DE112019005731T5\_0084.png"\; state="\{\{state\}\}">N</figure-callout>}}_3\right|}\cdot f_s\left({\mathrm{<figure-callout\; id="2"\; label="k"\; filenames="DE112019005731T5\_0084.png"\; state="\{\{state\}\}">k</figure-callout>}}_2{\mathrm{<figure-callout\; id="2"\; label="\alpha "\; filenames="DE112019005731T5\_0084.png"\; state="\{\{state\}\}">\alpha </figure-callout>}}_2+{\mathrm{<figure-callout\; id="3"\; label="k"\; filenames="DE112019005731T5\_0084.png"\; state="\{\{state\}\}">k</figure-callout>}}_3{\alpha }_3\right)$$

where f _{s is} the mathematical function in sawtooth form and with a slope equal to 1, α _{2 is} the signal generated by the second angular position sensor, α _{3 is} the signal generated by the third angular position sensor, and G is the stiffness of the torsion bar , and where k ₂ , k _{3 are} numeric coefficients that are selected and must meet the following conditions: $$\{\begin{array}{l}{k}_{2}und{k}_{3}relativeGanzeZaHlen\hfill \\ N_2{\mathrm{<figure-callout\; id="2"\; label="k"\; filenames="DE112019005731T5\_0084.png"\; state="\{\{state\}\}">k</figure-callout>}}_2+N_3{\mathrm{<figure-callout\; id="3"\; label="k"\; filenames="DE112019005731T5\_0084.png"\; state="\{\{state\}\}">k</figure-callout>}}_3=0\hfill \\ \frac{360}{\left|k_3{\mathrm{<figure-callout\; id="3"\; label="N"\; filenames="DE112019005731T5\_0084.png"\; state="\{\{state\}\}">N</figure-callout>}}_3\right|}>\mathrm{\Delta }{\theta }_{sHift}\hfill \end{array}$$

where Δθ _{shift is} the peak-to-peak variation in the angular deformation of the torsion bar,

• - According to a second variant, the second calculated signal reads:

$$\widehat{T}=\frac{G}{\left|k_3{\mathrm{<figure-callout\; id="3"\; label="N"\; filenames="DE112019005731T5\_0084.png"\; state="\{\{state\}\}">N</figure-callout>}}_3\right|}\cdot f_s\left({\mathrm{<figure-callout\; id="1"\; label="k"\; filenames="DE112019005731T5\_0084.png"\; state="\{\{state\}\}">k</figure-callout>}}_1{\mathrm{<figure-callout\; id="1"\; label="\alpha "\; filenames="DE112019005731T5\_0084.png"\; state="\{\{state\}\}">\alpha </figure-callout>}}_1+{\mathrm{<figure-callout\; id="3"\; label="k"\; filenames="DE112019005731T5\_0084.png"\; state="\{\{state\}\}">k</figure-callout>}}_3{\alpha }_3\right)$$

where f _{s is} the mathematical function in sawtooth form and with a slope equal to 1, α _{1 is} the signal generated by the first angular position sensor, α _{3 is} the signal generated by the third angular position sensor, and G is the stiffness of the torsion bar , and where k ₁ , k _{3 are} numeric coefficients that are selected and must meet the following conditions: $$\{\begin{array}{l}{k}_{2}und{k}_{3}relativeGanzeZaHlen\hfill \\ {\mathrm{R.}}_{red}{N}_1{\mathrm{<figure-callout\; id="1"\; label="k"\; filenames="DE112019005731T5\_0084.png"\; state="\{\{state\}\}">k</figure-callout>}}_1+N_3{\mathrm{<figure-callout\; id="3"\; label="k"\; filenames="DE112019005731T5\_0084.png"\; state="\{\{state\}\}">k</figure-callout>}}_3=0\hfill \\ \frac{360}{\left|k_3{\mathrm{<figure-callout\; id="3"\; label="N"\; filenames="DE112019005731T5\_0084.png"\; state="\{\{state\}\}">N</figure-callout>}}_3\right|}>\mathrm{\Delta }{\theta }_{sHift}\hfill \end{array}$$

where Δθ _{Shift is} the peak-to-peak variation in the angular deformation of the torsion bar,

• - According to a third embodiment, the second calculated signal reads:

$$\widehat{T}=\frac{G}{\left|k_3{\mathrm{<figure-callout\; id="3"\; label="N"\; filenames="DE112019005731T5\_0084.png"\; state="\{\{state\}\}">N</figure-callout>}}_3\right|}\cdot f_s\left({\mathrm{<figure-callout\; id="3"\; label="k"\; filenames="DE112019005731T5\_0084.png"\; state="\{\{state\}\}">k</figure-callout>}}_3{\mathrm{<figure-callout\; id="3"\; label="\alpha "\; filenames="DE112019005731T5\_0084.png"\; state="\{\{state\}\}">\alpha </figure-callout>}}_3{\widehat{\theta }}_2+{\mathrm{<figure-callout\; id="3"\; label="k"\; filenames="DE112019005731T5\_0084.png"\; state="\{\{state\}\}">k</figure-callout>}}_3{\alpha }_3\right)$$

where f _{s is} the mathematical function in sawtooth form and with a slope equal to 1, θ̂ _{2 is} the first calculated signal, a _{3 is} the signal generated by the third angular position sensor, and G is the stiffness of the torsion bar, and where k _{3 is} a is a numerical coefficient that is selected and must meet the following conditions: $$\{\begin{array}{l}{k}_{3}relativeGanzeZaHl\hfill \\ \frac{360}{\left|k_3{\mathrm{<figure-callout\; id="3"\; label="N"\; filenames="DE112019005731T5\_0084.png"\; state="\{\{state\}\}">N</figure-callout>}}_3\right|}>\mathrm{\Delta }{\theta }_{sHift}\hfill \end{array}$$

where Δθ _{shift is} the peak-to-peak variation in the angular deformation of the torsion bar,

• the processing unit verifies that the totality of the measured and calculated signals belongs to a totality of admissible values, the processing unit delivering a warning signal if the totality of values does not belong to any set of admissible values,
• - the first angular position sensor, the second angular position sensor and / or the third angular position sensor are sensors or combinations of position sensors with Hall effect, magnetic resistance, fluxgates, induction, with eddy current or with variable reluctance,
• the second angular position sensor and the third angular position sensor are eddy current position sensors that include a common detection probe comprising a common support plate for the windings of the second angular position sensor and the third angular position sensor.

Another object of the invention relates to a steering system equipped with the detection system according to the invention which executes a steering command as a function of the absolute steering wheel angle over more than one mechanical revolution and the torque exerted.

• 1 ist das Schema, welches das Erfassungssystem gemäß der Erfindung veranschaulicht, das die Messung des Drehmoments und des absoluten Lenkradwinkels über mehr als eine mechanische Umdrehung für eine Lenkung eines Fahrzeugs ermöglicht.
• 2 bis 4 sind synoptische Schemata, die drei Ausführungsvarianten für die Berechnung des ausgeübten Drehmoments veranschaulichen.
• 5 veranschaulicht die Form der durch die drei Winkelpositionssensoren gelieferten Signale umgesetzt durch das Erfassungssystem gemäß der Erfindung in Abhängigkeit von dem Lenkradwinkel in Grad und für ein ausgeübtes maximales Drehmoment.
• 6 veranschaulicht das erste berechnete Signal θ̂₂ proportional zu dem absoluten Lenkradwinkel über mehr als eine mechanische Umdrehung für unterschiedliche Werte des Drehmoments.
• 7 veranschaulicht das zweite berechnete Signal T̂ proportional zu dem Drehmoment für unterschiedliche Werte des Lenkradwinkels.
• 8 und 9 sind Schemata, die zwei Ausführungsvarianten des zweiten Sensors und des dritten Sensors in der Form von Positionssensoren mit Wirbelstrom veranschaulichen.

Various other properties emerge from the following description with reference to the attached drawings, which show exemplary and non-limiting embodiments of the subject matter of the invention.

• 1 Figure 3 is the diagram illustrating the detection system according to the invention, which enables the measurement of the torque and the absolute steering wheel angle over more than one mechanical revolution for a steering of a vehicle.
• 2 until 4th are synoptic diagrams that illustrate three design variants for calculating the torque exerted.
• 5 illustrates the form of the signals supplied by the three angular position sensors converted by the detection system according to the invention as a function of the steering wheel angle in degrees and for a maximum torque exerted.
• 6th illustrates the first calculated signal θ̂ ₂ proportional to the absolute steering wheel angle over more than one mechanical revolution for different values of the torque.
• 7th illustrates the second calculated signal T̂ proportional to the torque for different values of the steering wheel angle.
• 8th and 9 are diagrams which illustrate two embodiment variants of the second sensor and the third sensor in the form of position sensors with eddy currents.

How it looks more precisely 1 As can be seen, the subject matter of the invention relates to a detection system 1 which is the measurement of the torque T and the absolute steering wheel angle θ ₂ over more than one mechanical turn of a steering 2 of a vehicle in the general sense. According to a preferred exemplary embodiment, which is described below in the description, the detection system enables 1 to measure the torque and the absolute steering wheel angle with several revolutions of an electric power steering of a vehicle. It goes without saying that the acquisition system 1 is also designed to measure the torque and the absolute steering wheel angle with several revolutions of a steering system of a vehicle for which the steering wheel is mechanically separated from the wheels.

This steering 2 includes an electric motor 3 that with a reduction gear 4th of all known types is provided, which has a reduction ratio R _red , which exerts a torque on the steering of the vehicle, namely the steering column 5 in the illustrated example, the one with a steering wheel 6th Is provided. Classically, the steering wheel can 6th Turn approximately three turns from a left end to a right end. The electric motor 3 exerts an assisting torque on the steering of the vehicle in the case of electric power steering and a resisting torque in the case of steering in which the steering wheel is mechanically separated from the wheels.

It goes without saying that the reduction gear 4th can exert the assisting torque on the steering rack, which is not shown and which serves to bring the wheels of the vehicle to rotate.

This steering 2 also includes a torsion bar 7th , which in the illustrated example is on the steering column 5 between the reduction gear 4th and the steering wheel 6th is mounted. This torsion bar 7th is implemented in a suitable manner to enable the measurement by angular deformation of the torque exerted on the steering. The torque T acting on the torsion bar 7th is the first quantity to be measured.

According to the invention, the detection system includes 1 a first angular position sensor 11 , which has N1 pole pairs, where N1 is an integer greater than or equal to 1. This first angular position sensor 11 measures the angle θ ₁ of the rotor of the electric motor 3 and delivers a first signal a ₁ . It should be noted that the first angular position sensor 11 Generally available in a power steering, as it is the control of the electric motor 3 the power steering allows.

The registration system 1 includes a second angular position sensor 12th , which has N2 pole pairs, where N2 is an integer greater than or equal to 1. This second angular position sensor 12th measures the angle θ ₂ the steering, which is between the reduction gear 4th and a first side of the torsion bar 7th located, namely the side opposite the steering wheel 6th is located. That angle θ ₂ the steering is located between the reduction gear 4th and the torsion bar 7th , so it is in front of the torsion bar. This second angular position sensor 12th supplies a second signal α ₂ . It should be noted that the second angular position sensor 12th the angle θ ₂ the steering measures in relation to the chassis of the vehicle.

The fact of an integer number of pole pairs N2 Choosing greater than or equal to 1 enables this second sensor to be simple in design and inexpensive. The consequence of this is that the second signal α _{2 is} not absolute over more than one mechanical revolution.

As will be explained in the following description, the detection system aims 1 to determine the corresponding reference steering wheel angle, which by default is the absolute steering wheel angle over more than one mechanical revolution θ ₂ is equivalent to. This absolute steering wheel angle with several turns θ ₂ is the second quantity to be measured.

The registration system 1 includes a third angular position sensor 13th , which has N3 pole pairs, where N3 is an integer greater than or equal to 1. This third angular position sensor 13th measures the angle θ ₃ the steering, which is on a second side of the torsion bar 7th located opposite the first side of the torsion bar 7th is located. This third angular position sensor 13th that is on the part of the steering column directly related to the steering wheel 6th is mounted, measures the steering wheel angle in a non-absolute way after the torsion bar 7th . This third angular position sensor 13th delivers a third signal a ₃ . It should be noted that the third angular position sensor 13th the angle θ ₃ the steering measures in relation to the chassis of the vehicle.

wobei mod der Modulo-Operator ist, der den Rest einer Division angibt. Die Funktion f_s ist eine Funktion in Sägezahnform mit Steigung gleich 1, enthalten zwischen -180° und 180° (nicht eingeschlossen) und null durchlaufend, wenn x gleich null ist. Standardmäßig verwendet dieses Dokument Winkel, die in Grad definiert sind, aber es ist auch möglich, die gleiche Funktion mit anderen Winkeleinheiten zu definieren. Zum Beispiel ist die Funktion f_s, die einen beliebigen Winkel, der in Bogenmaß definiert ist, in einen Winkel transformiert, der zwischen -π und π (nicht eingeschlossen) enthalten ist, definiert durch: $$f_s:x\to \left(\left(x+\pi \right)mod2\pi \right)-\pi$$

Wenn die Funktion f_s mit Bogenmaßwinkeln definiert ist, kann die Funktion f_s alternativ auch auf vollkommen äquivalente Weise definiert sein durch: $$f_s:x\to atan2\left(\text{sin}\left(x\right),\text{ cos}\left(x\right)\right)$$

wobei sin und cos die klassischen trigonometrischen Funktionen sind und wobei atan2 die Funktion ist, welche die Winkelkoordinierung (definiert zwischen -π und π nicht eingeschlossen) eines Punktes in der euklidischen Ebene angibt.According to the invention, the mathematical function f _{s is} defined, which is a mathematical function in a sawtooth shape with a slope equal to 1 and whose value is zero when the input data is zero. This mathematical function f _s allows any angle, defined in degrees, to be transformed into an angle between -180 ° and 180 ° (not included): $$f_s:x\to \left(\left(x+180\right)mOd360\right)-180$$

where mod is the modulo operator that specifies the remainder of a division. The function f _s is a sawtooth function with a slope equal to 1, contained between -180 ° and 180 ° (not included) and running through zero when x is equal to zero. By default, this document uses angles that are defined in degrees, but it is also possible to define the same function with other angle units. For example, the function f _s that transforms any angle defined in radians to an angle contained between -π and π (not included) is defined by: $$f_s:x\to \left(\left(x+\pi \right)m\mathrm{<figure-callout\; id="2"\; label="O\; d"\; filenames="DE112019005731T5\_0084.png"\; state="\{\{state\}\}">O</figure-callout>}d2\pi \right)-\pi$$

If the function f _{s is} defined with radian angles, the function f _{s can} alternatively be defined in a completely equivalent way by: $$f_s:x\to atan2\left(\text{sin}\left(x\right),\text{cos}\left(x\right)\right)$$

where sin and cos are the classical trigonometric functions and where atan2 is the function which gives the angular coordination (defined between -π and π not included) of a point in the Euclidean plane.

It goes without saying that it is possible to use further units such as the number of bits.

wobei G die Steifigkeit des Torsionsstabs 7 ist.The angular deformation of the torsion bar 7th is referred to as θ _shift . The relationship between the angular deformation θ _shift and the torque T is: $${\theta }_{sHift}=\frac{T}{G}$$

where G is the stiffness of the torsion bar 7th is.

By default, it is considered that the angle θ ₂ represents the reference steering wheel angle. According to the foregoing, the mathematical relationships between the independent mechanical angles θ ₂ , θ _shift and the dependent mechanical angles θ ₁ , θ ₃ can be modeled in the following way: $$\begin{array}{c}{\mathrm{<figure-callout\; id="1"\; label="\theta "\; filenames="DE112019005731T5\_0084.png"\; state="\{\{state\}\}">\theta </figure-callout>}}_1={\mathrm{R.}}_{red}\cdot {\mathrm{<figure-callout\; id="2"\; label="\theta "\; filenames="DE112019005731T5\_0084.png"\; state="\{\{state\}\}">\theta </figure-callout>}}_2\\ {\mathrm{<figure-callout\; id="3"\; label="\theta "\; filenames="DE112019005731T5\_0084.png"\; state="\{\{state\}\}">\theta </figure-callout>}}_3={\mathrm{<figure-callout\; id="2"\; label="\theta "\; filenames="DE112019005731T5\_0084.png"\; state="\{\{state\}\}">\theta </figure-callout>}}_2+{\theta }_{sHift}\end{array}$$

The reference steering wheel angle to be measured θ ₂ has a noticeable peak-to-peak variation Δθ ₂ . The torque to be measured has an observed peak-to-peak variation ΔT. This implies that the angular deformation of the torsion bar to be measured 7th has an observed peak-to-peak variation Δθ _shift equal to ΔT / G.

Aus dem Vorhergehenden folgt, dass jedes der Signale α₁, α₂, α₃ in dem Intervall [-180°, 180° [enthalten ist, was für die mechanischen Winkel θ₁ , θ₂ , θ₃ im Allgemeinen nicht der Fall ist.The first, the second and the third signal α ₁ , α ₂ , α ₃ , each by the first, the second and the third angular position sensor 11 , 12th , 13th are the same: $$\begin{array}{c}{\mathrm{<figure-callout\; id="1"\; label="\alpha "\; filenames="DE112019005731T5\_0084.png"\; state="\{\{state\}\}">\alpha </figure-callout>}}_1=f_s\left({\mathrm{<figure-callout\; id="1"\; label="N"\; filenames="DE112019005731T5\_0084.png"\; state="\{\{state\}\}">N</figure-callout>}}_1\cdot {\mathrm{\theta }}_1\right)\\ {\mathrm{<figure-callout\; id="2"\; label="\alpha "\; filenames="DE112019005731T5\_0084.png"\; state="\{\{state\}\}">\alpha </figure-callout>}}_2=f_s\left({\mathrm{<figure-callout\; id="2"\; label="N"\; filenames="DE112019005731T5\_0084.png"\; state="\{\{state\}\}">N</figure-callout>}}_2\cdot {\mathrm{\theta }}_2\right)\\ {\mathrm{<figure-callout\; id="3"\; label="\alpha "\; filenames="DE112019005731T5\_0084.png"\; state="\{\{state\}\}">\alpha </figure-callout>}}_3=f_s\left({\mathrm{<figure-callout\; id="3"\; label="N"\; filenames="DE112019005731T5\_0084.png"\; state="\{\{state\}\}">N</figure-callout>}}_3\cdot {\mathrm{\theta }}_3\right)\end{array}$$

From the above it follows that each of the signals α ₁ , α ₂ , α _{3 is contained} in the interval [-180 °, 180 ° [, which is for the mechanical angle θ ₁ , θ ₂ , θ ₃ generally not the case.

$$S_{cos}=A\cdot \text{cos}\left(\frac{2\pi }{360}\alpha \right)$$

wobei A die Amplitude der Signale ist. Es versteht sich, dass die Verarbeitungseinheit 15 anschließend die Signale S_sin und S_cos dekodieren muss, um das Signal α zurückzubekommen. In dem vorhergehenden Beispiel wird die Dekodierung auf folgende Weise bewirkt: $$\alpha =atan2\left(S_{sin},S_{cos}\right)\cdot \frac{360}{2\pi }$$

wobei atan2 die Funktion ist, welche die Winkelkoordinierung (definiert zwischen -π und π nicht eingeschlossen) eines Punktes in der euklidischen Ebene angibt.It is also important to note that the present invention does not relate to angular position sensors 11 , 12th , 13th is limited, which explicitly sends the signals a ₁ , a ₂ , a ₃ to the processing unit 15th transfer. Often times, as part of a physical implementation, the angular position sensors transmit 11 , 12th , 13th the information a ₁ , α ₂ , α ₃ in a coded form in order to simplify the transmission and to optimize the robustness of these signals against noise and any source of interference. For example, a signal α can be _{encoded in the form of two signals S sin} and S _cos , defined by: $${\mathrm{S.}}_{sin}=\mathrm{A.}\cdot \text{sin}\left(\frac{2\pi }{360}\alpha \right)$$

$${\mathrm{S.}}_{cOs}=\mathrm{A.}\cdot \text{cos}\left(\frac{2\pi }{360}\alpha \right)$$

where A is the amplitude of the signals. It goes without saying that the processing unit 15th then has to _{decode the signals S sin} and S _{cos in} order to get the signal α back. In the previous example, the decoding is done in the following way: $$\alpha =atan2\left({\mathrm{S.}}_{sin},{\mathrm{S.}}_{cOs}\right)\cdot \frac{360}{2\pi }$$

where atan2 is the function that gives the angular coordination (defined between -π and π not included) of a point in the Euclidean plane.

According to the invention, the detection system includes 1 also a processing unit 15th , which receives the first signal α ₁ , the second signal α ₂ and the third signal α ₃ and is configured to effect calculations to produce a first calculated signal θ̂ ₂ which is proportional to the absolute steering wheel angle over more than one mechanical revolution θ ₂ and to generate a second calculated signal T̂̂ proportional to the torque T exerted. This processing unit 15th is implemented by all computer systems that are programmed and designed to implement the processing and calculation operations according to the invention.

wobei:

• - $\mathrm{\Theta }=f_s\left({\mathrm{<figure-callout\; id="1"\; label="p"\; filenames="DE112019005731T5\_0084.png"\; state="\{\{state\}\}">p</figure-callout>}}_1{\mathrm{<figure-callout\; id="1"\; label="\alpha "\; filenames="DE112019005731T5\_0084.png"\; state="\{\{state\}\}">\alpha </figure-callout>}}_1+{\mathrm{<figure-callout\; id="2"\; label="p"\; filenames="DE112019005731T5\_0084.png"\; state="\{\{state\}\}">p</figure-callout>}}_2{\alpha }_2\right)$
  
• - f_s die vorstehend definierte mathematische Funktion ist,
• - q₁ und q₂ feste Gewichtungskoeffizienten sind, die der Entwickler auswählen kann, um den Sensorfehler zu minimieren. Die einfachste Wahl (q₁ = 0, q₂ = 0) definiert ein funktionelles Signal θ̂₂, das aber nicht immer optimal in Bezug auf Abweisung von Rauschen und von Messfehlern ist. Es gibt zahlreiche weitere Möglichkeiten, wie zum Beispiel (q₁ = 0, q₂ = 1/(N₂N_turns)) oder (q₁ = 1/(N₁R_redN_turns), q₂ = 0). Auf allgemeine Weise hängt die optimale Wahl der Koeffizienten q₁ und q₂ von der Präzision der zwei Winkelpositionssensoren 11, 12 ab, die für den Vernier-Effekt verwendet werden.
• - α₁ das Signal ist, das durch den ersten Winkelpositionssensor 11 erzeugt wird, α₂ das Signal, das durch den zweiten Winkelpositionssensor 12 erzeugt wird, Die Parameter p₁, p₂, und N_turns sind numerische Koeffizienten, die von dem Entwickler ausgewählt werden und die folgenden Bedingungen erfüllen müssen:

$$\{\begin{array}{c}{p}_{1}und{p}_{2}relativeganzeZahlen\\ \left(R_{red}{N}_1{\mathrm{<figure-callout\; id="1"\; label="p"\; filenames="DE112019005731T5\_0084.png"\; state="\{\{state\}\}">p</figure-callout>}}_1+N_2{p}_2\right)\cdot N_{turns}=1\\ N_{turns}\ge \frac{\mathrm{<figure-callout\; id="2"\; label="\Delta \; \theta "\; filenames="DE112019005731T5\_0084.png"\; state="\{\{state\}\}">\Delta </figure-callout>}{\theta }_{}{}_2}{360}\end{array}$$

wobei Δθ₂ die Spitze-zu-Spitze-Variation des absoluten Lenkradwinkels über mehr als eine mechanische Umdrehung ist und R_red das Untersetzungsverhältnis des Untersetzungsgetriebes 4 ist.This processing unit 15th causes on the one hand a Vernier calculation on the basis of at least the first signal α ₁ and the second signal α ₂ in order to generate a first calculated signal θ̂̂ ₂ , which is proportional to the absolute steering wheel angle over more than one mechanical revolution, and on the other hand a weighted angular sum of the third signal α ₃ and one of the signals from the first signal α ₁ , the second signal α ₂ and the first calculated signal θ̂ ₂ to generate a second calculated signal T proportional to the torque T.
According to a first variant embodiment, the processing unit effects 15th a vernier computation to produce the first computed signal θ̂ ₂ , which is: $${\widehat{\theta }}_2=N_{turns}\cdot f_s\left(\mathrm{\Theta }-{\mathrm{<figure-callout\; id="1"\; label="q"\; filenames="DE112019005731T5\_0084.png"\; state="\{\{state\}\}">q</figure-callout>}}_1\cdot f_s\left({\mathrm{<figure-callout\; id="1"\; label="N"\; filenames="DE112019005731T5\_0084.png"\; state="\{\{state\}\}">N</figure-callout>}}_1{\mathrm{R.}}_{red}{N}_{turns}\mathrm{\Theta }-{\alpha }_1\right)-{\mathrm{<figure-callout\; id="2"\; label="q"\; filenames="DE112019005731T5\_0084.png"\; state="\{\{state\}\}">q</figure-callout>}}_2\cdot f_s\left(N_2{N}_{turns}\mathrm{\Theta }-{\alpha }_2\right)\right)$$

whereby:

• - $\mathrm{\Theta }=f_s\left(p_1{\alpha }_1+p_2{\alpha }_2\right)$
  
• - f _{s is} the mathematical function defined above,
• q ₁ and q _{2 are} fixed weighting coefficients that the designer can choose to minimize sensor error. The simplest choice (q ₁ = 0, q ₂ = 0) defines a functional signal θ̂ ₂ , which, however, is not always optimal with regard to the rejection of noise and measurement errors. There are numerous other possibilities, such as (q ₁ = 0, q ₂ = 1 / (N ₂ N _turns )) or (q ₁ = 1 / (N ₁ R _red N _turns ), q ₂ = 0). In general, the optimal choice of the coefficients q ₁ and q ₂ depends on the precision of the two angular position sensors 11 , 12th used for the vernier effect.
• - α _{1 is} the signal that is passed by the first angular position sensor 11 is generated, α ₂ the signal that is generated by the second angular position sensor 12th The parameters p ₁ , p ₂ , and N _turns are numerical coefficients that are selected by the developer and must meet the following conditions:

$$\{\begin{array}{c}{p}_{1}und{p}_{2}relativeGanzeZaHlen\\ \left({\mathrm{R.}}_{red}{N}_1{\mathrm{<figure-callout\; id="1"\; label="p"\; filenames="DE112019005731T5\_0084.png"\; state="\{\{state\}\}">p</figure-callout>}}_1+N_2{p}_2\right)\cdot N_{turns}=1\\ N_{turns}\ge \frac{\mathrm{<figure-callout\; id="2"\; label="\Delta \; \theta "\; filenames="DE112019005731T5\_0084.png"\; state="\{\{state\}\}">\Delta </figure-callout>}{\theta }_{}{}_2}{360}\end{array}$$

where Δθ _{2 is} the peak-to-peak variation of the absolute steering wheel angle over more than one mechanical revolution and R _{red is} the reduction ratio of the reduction gear 4th is.

It should be noted that the parameter N _{turns has} a physical meaning. It corresponds to the number of mechanical turns of the steering wheel at which the first calculated signal θ ₂ will be sought. For this reason, the restriction N _turns ≧ Δθ _2/360 is present, which means that the linear peak-to-peak variation θ of the first calculated signal to be ₂ at least greater than the peak-to-peak variation of the measured parameter must which is defined by the specification of the application. In addition, N _turns does not have to be an integer.

It should be noted that the subject invention uses the Vernier effect technique to construct the first calculated signal θ̂ _{2 of} the steering wheel angle.

wobei N₁, p₁, N₂, p₂, ganze Zahlen sindAnother important point to consider is that in order to have an absolute steering wheel angle sensor over several mechanical turns (N _turns > 1), it is essential to have an incomplete reduction ratio R _red . This is understood with the help of the restriction that is already defined above: $${\mathrm{R.}}_{red}{N}_1{\mathrm{<figure-callout\; id="1"\; label="p"\; filenames="DE112019005731T5\_0084.png"\; state="\{\{state\}\}">p</figure-callout>}}_1+N_2{\mathrm{<figure-callout\; id="2"\; label="p"\; filenames="DE112019005731T5\_0084.png"\; state="\{\{state\}\}">p</figure-callout>}}_2=\frac{1}{N_{turns}}$$

where N ₁ , p ₁ , N ₂ , p ₂ are integers

If N _turns > 1, it is easy to see that the only way to address this constraint is to have an incomplete reduction ratio R _red . This is generally the case with electric power steering. In order to be absolutely safe over several turns of the steering wheel (when the system is switched on), the current steering wheel angle sensors use gears in order to be able to differentiate between the mechanical turns. These additional transmissions are costly. Thanks to the proposed invention, it is possible to implement a sensor with several revolutions without adding gear, since the subject of the invention is the gear of the reduction gear 4th of the electric motor 3 that is already available will be used again.

$$\begin{array}{l}wenn{N}_2>N_1{R}_{red}\\ \{\begin{array}{c}{q}_1=\frac{1}{p_2{N}_{turns}}\cdot \frac{{\sigma }_2^2{p}_2{N}_1{R}_{red}-{\sigma }_1^2{p}_1{N}_2}{{\left({\sigma }_1{N}_2\right)}^2+{\left({\sigma }_2{N}_1{R}_{red}\right)}^2}\\ q_2=0\end{array}\end{array}$$

wobei σ₁ die typische Fehleramplitude des ersten Winkelpositionssensors ist und wobei σ₂ die typische Fehleramplitude des zweiten Winkelpositionssensors ist.According to an advantageous variant, the optimal values of the weighting coefficients q ₁ and q _{2 are} the following: $$\begin{array}{l}wenn{N}_2<{\mathrm{<figure-callout\; id="1"\; label="N"\; filenames="DE112019005731T5\_0084.png"\; state="\{\{state\}\}">N</figure-callout>}}_1{\mathrm{R.}}_{red}\\ \{\begin{array}{c}{\mathrm{<figure-callout\; id="1"\; label="q"\; filenames="DE112019005731T5\_0084.png"\; state="\{\{state\}\}">q</figure-callout>}}_1=0\\ {\mathrm{<figure-callout\; id="2"\; label="q"\; filenames="DE112019005731T5\_0084.png"\; state="\{\{state\}\}">q</figure-callout>}}_2=\frac{1}{p_1{N}_{turns}}\cdot \frac{{\mathrm{<figure-callout\; id="1"\; label="\sigma "\; filenames="DE112019005731T5\_0084.png"\; state="\{\{state\}\}">\sigma </figure-callout>}}_1^2{p}_1{N}_2-{\mathrm{<figure-callout\; id="2"\; label="\sigma "\; filenames="DE112019005731T5\_0084.png"\; state="\{\{state\}\}">\sigma </figure-callout>}}_2^2{p}_2{\mathrm{<figure-callout\; id="1"\; label="N"\; filenames="DE112019005731T5\_0084.png"\; state="\{\{state\}\}">N</figure-callout>}}_1{\mathrm{R.}}_{red}}{{\left({\sigma }_1{N}_2\right)}^2+{\left({\sigma }_2{\mathrm{<figure-callout\; id="1"\; label="N"\; filenames="DE112019005731T5\_0084.png"\; state="\{\{state\}\}">N</figure-callout>}}_1{\mathrm{R.}}_{red}\right)}^2}\end{array}\end{array}$$

$$\begin{array}{l}wenn{N}_2>{\mathrm{<figure-callout\; id="1"\; label="N"\; filenames="DE112019005731T5\_0084.png"\; state="\{\{state\}\}">N</figure-callout>}}_1{\mathrm{R.}}_{red}\\ \{\begin{array}{c}{\mathrm{<figure-callout\; id="1"\; label="q"\; filenames="DE112019005731T5\_0084.png"\; state="\{\{state\}\}">q</figure-callout>}}_1=\frac{1}{p_2{N}_{turns}}\cdot \frac{{\mathrm{<figure-callout\; id="2"\; label="\sigma "\; filenames="DE112019005731T5\_0084.png"\; state="\{\{state\}\}">\sigma </figure-callout>}}_2^2{p}_2{\mathrm{<figure-callout\; id="1"\; label="N"\; filenames="DE112019005731T5\_0084.png"\; state="\{\{state\}\}">N</figure-callout>}}_1{\mathrm{R.}}_{red}-{\mathrm{<figure-callout\; id="1"\; label="\sigma "\; filenames="DE112019005731T5\_0084.png"\; state="\{\{state\}\}">\sigma </figure-callout>}}_1^2{p}_1{N}_2}{{\left({\sigma }_1{N}_2\right)}^2+{\left({\sigma }_2{\mathrm{<figure-callout\; id="1"\; label="N"\; filenames="DE112019005731T5\_0084.png"\; state="\{\{state\}\}">N</figure-callout>}}_1{\mathrm{R.}}_{red}\right)}^2}\\ {\mathrm{<figure-callout\; id="2"\; label="q"\; filenames="DE112019005731T5\_0084.png"\; state="\{\{state\}\}">q</figure-callout>}}_2=0\end{array}\end{array}$$

where σ _{1 is} the typical error amplitude of the first angular position sensor and where σ _{2 is} the typical error amplitude of the second angular position sensor.

wie folgt berechnet: $${\alpha }_2^{*}=f_s\left(f_s\left(c_2\cdot f_s\left({\alpha }_2+n_3\delta \right)+c_3\cdot f_s\left({\alpha }_3-n_2\delta \right)\right)-\frac{n_3}{n_2}\delta \right)$$

$$\text{wobei}:\delta =\frac{f_s\left(-n_3{\alpha }_2+n_2{\alpha }_3\right)}{n_2^2+n_3^2}$$

$$\{\begin{array}{c}{N}_{23}=\text{pgcd}\left(N_2,N_3\right)\\ n_2=\frac{N_2}{N_{23}}\\ n_3=\frac{N_3}{N_{23}}\end{array}$$

wobei pgcd der Operator „größter gemeinsamer Teiler“ ist. Die Parameter c₂ et c₃ sind numerische Koeffizienten, die von dem Entwickler ausgewählt werden und die folgenden Bedingungen erfüllen müssen: $$\{\begin{array}{l}{c}_{2}und{c}_{3}relativeganzeZahlen\hfill \\ c_2{n}_2+c_3{n}_3=1\hfill \end{array}$$

According to a second variant embodiment, the first signal θ̂ _{2 is} calculated on the basis of the signals α ₁ , α ₂ and α ₃ in the following way. First there is an intermediate signal ${\alpha }_{}^{}$${}_2^{*}$

calculated as follows: $${\mathrm{<figure-callout\; id="2"\; label="\alpha "\; filenames="DE112019005731T5\_0084.png"\; state="\{\{state\}\}">\alpha </figure-callout>}}_2^{*}=f_s\left(f_s\left({\mathrm{<figure-callout\; id="2"\; label="c"\; filenames="DE112019005731T5\_0084.png"\; state="\{\{state\}\}">c</figure-callout>}}_2\cdot f_s\left({\mathrm{<figure-callout\; id="2"\; label="\alpha "\; filenames="DE112019005731T5\_0084.png"\; state="\{\{state\}\}">\alpha </figure-callout>}}_2+{\mathrm{<figure-callout\; id="3"\; label="n"\; filenames="DE112019005731T5\_0084.png"\; state="\{\{state\}\}">n</figure-callout>}}_3\delta \right)+{\mathrm{<figure-callout\; id="3"\; label="c"\; filenames="DE112019005731T5\_0084.png"\; state="\{\{state\}\}">c</figure-callout>}}_3\cdot f_s\left({\alpha }_3-{\mathrm{<figure-callout\; id="2"\; label="n"\; filenames="DE112019005731T5\_0084.png"\; state="\{\{state\}\}">n</figure-callout>}}_2\delta \right)\right)-\frac{n_3}{{\mathrm{<figure-callout\; id="2"\; label="n"\; filenames="DE112019005731T5\_0084.png"\; state="\{\{state\}\}">n</figure-callout>}}_2}\delta \right)$$

$$\text{whereby}:\delta =\frac{f_s\left(-{\mathrm{<figure-callout\; id="3"\; label="n"\; filenames="DE112019005731T5\_0084.png"\; state="\{\{state\}\}">n</figure-callout>}}_3{\mathrm{<figure-callout\; id="2"\; label="\alpha "\; filenames="DE112019005731T5\_0084.png"\; state="\{\{state\}\}">\alpha </figure-callout>}}_2+{\mathrm{<figure-callout\; id="2"\; label="n"\; filenames="DE112019005731T5\_0084.png"\; state="\{\{state\}\}">n</figure-callout>}}_2{\alpha }_3\right)}{{\mathrm{<figure-callout\; id="2"\; label="n"\; filenames="DE112019005731T5\_0084.png"\; state="\{\{state\}\}">n</figure-callout>}}_2^2+{\mathrm{<figure-callout\; id="3"\; label="n"\; filenames="DE112019005731T5\_0084.png"\; state="\{\{state\}\}">n</figure-callout>}}_3^2}$$

$$\{\begin{array}{c}{\mathrm{<figure-callout\; id="23"\; label="N"\; filenames="DE112019005731T5\_0088.png"\; state="\{\{state\}\}">N</figure-callout>}}_{23}=\text{pgcd}\left({\mathrm{<figure-callout\; id="2"\; label="N"\; filenames="DE112019005731T5\_0084.png"\; state="\{\{state\}\}">N</figure-callout>}}_2,N_3\right)\\ {\mathrm{<figure-callout\; id="2"\; label="n"\; filenames="DE112019005731T5\_0084.png"\; state="\{\{state\}\}">n</figure-callout>}}_2=\frac{N_2}{N_{23}}\\ {\mathrm{<figure-callout\; id="3"\; label="n"\; filenames="DE112019005731T5\_0084.png"\; state="\{\{state\}\}">n</figure-callout>}}_3=\frac{N_3}{{\mathrm{<figure-callout\; id="23"\; label="N"\; filenames="DE112019005731T5\_0088.png"\; state="\{\{state\}\}">N</figure-callout>}}_{23}}\end{array}$$

where pgcd is the operator "greatest common divisor". The parameters c ₂ et c ₃ are numerical coefficients that are selected by the developer and must meet the following conditions: $$\{\begin{array}{l}{c}_{2}und{c}_{3}relativeGanzeZaHlen\hfill \\ c_2{\mathrm{<figure-callout\; id="2"\; label="n"\; filenames="DE112019005731T5\_0084.png"\; state="\{\{state\}\}">n</figure-callout>}}_2+c_3{\mathrm{<figure-callout\; id="3"\; label="n"\; filenames="DE112019005731T5\_0084.png"\; state="\{\{state\}\}">n</figure-callout>}}_3=1\hfill \end{array}$$

den gleichen absoluten Messfehler (in mechanischen Grad) wie das Signal α₂ besitzt, aber dass es eine schwächere Periodizität (gleich N₂₃) als die Periodizität des Signals α₂ (gleich N₂) besitzt. Anders gesagt hat die Verwendung des Signals α₃ es ermöglicht, den relativen Messfehler (in elektrischen Grad) von α₂ dank der Verringerung seiner Periodizität zu verringern. Dank dieser Verbesserung ist es in der Praxis in einem zweiten Schritt einfacher, a₁ mit θ̂₂ anstatt mit a₂ zu kombinieren, um ${\alpha }_2^{*}$

zu erhalten. Diese Rechnung ist wie folgt definiert: $${\widehat{\theta }}_2=N_{{}_{turns}}^{*}\cdot f_s\left({\mathrm{\Theta }}^{\mathrm{\ast }}-q_1^{*}\cdot f_s\left(N_1{R}_{red}{N}_{turns}^{*}{\mathrm{\Theta }}^{\mathrm{\ast }}-{\alpha }_1\right)-q_2^{*}\cdot f_s\left(N_{23}{N}_{turn}^{*}{\mathrm{\Theta }}^{\mathrm{\ast }}-{\alpha }_2^{*}\right)\right)$$

wobei: $${\mathrm{\Theta }}^{*}=f_s\left(p_1^{*}{\alpha }_1+p_2^{*}{\alpha }_2^{*}\right)$$

wobei $q_1^{*}$

und $q_2^{*}$

feste Gewichtungskoeffizienten sind, die der Entwickler auswählen kann, um den Sensorfehler zu minimieren. Die einfachste Wahl $\left(q_1^{*}=\mathrm{0,}{q}_2^{*}=0\right)$

definiert ein funktionelles Signal θ̂₂, das aber nicht immer optimal in Bezug auf Abweisung von Rauschen und von Messfehlern ist. Die Parameter $p_1^{*}$

, $p_2^{*}$

und $N_{turns}^{*}$

sind numerische Koeffizienten, die von dem Entwickler ausgewählt werden und die folgenden Bedingungen erfüllen müssen: $$\{\begin{array}{c}{p}_1^{*}und{p}_2^{*}relativeganzeZahlen\\ \left(R_{red}{N}_1{p}_1^{*}+N_{23}{p}_2^{*}\right)\cdot N_{turns}^{*}=1\\ N_{turns}^{*}\ge \frac{\mathrm{\Delta }{\theta }_2}{360}\end{array}$$

wobei Δθ₂ die Spitze-zu-Spitze-Variation des absoluten Lenkradwinkels über mehr als eine mechanische Umdrehung ist und R_red das Untersetzungsverhältnis des Untersetzungsgetriebes 4 ist.It is possible to demonstrate that the intermediate signal ${\alpha }_{}^{}$${}_2^{*}$

has the same absolute measurement error (in mechanical degrees) as the signal α ₂ , but that it has a weaker periodicity (equal to N ₂₃ ) than the periodicity of the signal α ₂ (equal to N ₂ ). In other words, the use of the signal α _{3 made} it possible to reduce the relative measurement error (in electrical degrees) of α ₂ thanks to the reduction in its periodicity. Thanks to this improvement, in practice, in a second step, it is easier _{to combine a 1} with θ̂ ₂ instead of with a ₂ in order to ${\alpha }_2^{*}$

to obtain. This calculation is defined as follows: $${\widehat{\theta }}_2=N_{{}_{turns}}^{*}\cdot f_s\left({\mathrm{\Theta }}^{\mathrm{\ast }}-{\mathrm{<figure-callout\; id="1"\; label="q"\; filenames="DE112019005731T5\_0084.png"\; state="\{\{state\}\}">q</figure-callout>}}_1^{*}\cdot f_s\left({\mathrm{<figure-callout\; id="1"\; label="N"\; filenames="DE112019005731T5\_0084.png"\; state="\{\{state\}\}">N</figure-callout>}}_1{\mathrm{R.}}_{red}{N}_{turns}^{*}{\mathrm{\Theta }}^{\mathrm{\ast }}-{\alpha }_1\right)-{\mathrm{<figure-callout\; id="2"\; label="q"\; filenames="DE112019005731T5\_0084.png"\; state="\{\{state\}\}">q</figure-callout>}}_2^{*}\cdot f_s\left(N_{23}{N}_{turn}^{*}{\mathrm{\Theta }}^{\mathrm{\ast }}-{\mathrm{<figure-callout\; id="2"\; label="\alpha "\; filenames="DE112019005731T5\_0084.png"\; state="\{\{state\}\}">\alpha </figure-callout>}}_2^{*}\right)\right)$$

whereby: $${\mathrm{\Theta }}^{*}=f_s\left({\mathrm{<figure-callout\; id="1"\; label="p"\; filenames="DE112019005731T5\_0084.png"\; state="\{\{state\}\}">p</figure-callout>}}_1^{*}{\mathrm{<figure-callout\; id="1"\; label="\alpha "\; filenames="DE112019005731T5\_0084.png"\; state="\{\{state\}\}">\alpha </figure-callout>}}_1+{\mathrm{<figure-callout\; id="2"\; label="p"\; filenames="DE112019005731T5\_0084.png"\; state="\{\{state\}\}">p</figure-callout>}}_2^{*}{\mathrm{<figure-callout\; id="2"\; label="\alpha "\; filenames="DE112019005731T5\_0084.png"\; state="\{\{state\}\}">\alpha </figure-callout>}}_2^{*}\right)$$

whereby ${\mathrm{<figure-callout\; id="1"\; label="q"\; filenames="DE112019005731T5\_0084.png"\; state="\{\{state\}\}">q</figure-callout>}}_1^{*}$

and ${\mathrm{<figure-callout\; id="2"\; label="q"\; filenames="DE112019005731T5\_0084.png"\; state="\{\{state\}\}">q</figure-callout>}}_2^{*}$

are fixed weighting coefficients that the designer can choose to minimize sensor error. The easiest choice $\left({\mathrm{<figure-callout\; id="1"\; label="q"\; filenames="DE112019005731T5\_0084.png"\; state="\{\{state\}\}">q</figure-callout>}}_1^{*}=\mathrm{0,}{\mathrm{<figure-callout\; id="2"\; label="q"\; filenames="DE112019005731T5\_0084.png"\; state="\{\{state\}\}">q</figure-callout>}}_2^{*}=0\right)$

defines a functional signal θ̂ ₂ , which, however, is not always optimal with regard to the rejection of noise and measurement errors. The parameters $p_{}^{}$${}_1^{*}$

, ${\mathrm{<figure-callout\; id="2"\; label="p"\; filenames="DE112019005731T5\_0084.png"\; state="\{\{state\}\}">p</figure-callout>}}_2^{*}$

and $N_{turns}^{*}$

are numerical coefficients that are selected by the developer and must meet the following conditions: $$\{\begin{array}{c}{\mathrm{<figure-callout\; id="1"\; label="p"\; filenames="DE112019005731T5\_0084.png"\; state="\{\{state\}\}">p</figure-callout>}}_1^{*}u\mathrm{<figure-callout\; id="2"\; label="n\; d\; p"\; filenames="DE112019005731T5\_0084.png"\; state="\{\{state\}\}">n</figure-callout>}d{p}_{}^{}{}_2^{*}relativeGanzeZaHlen\\ \left({\mathrm{R.}}_{red}{N}_1{\mathrm{<figure-callout\; id="1"\; label="p"\; filenames="DE112019005731T5\_0084.png"\; state="\{\{state\}\}">p</figure-callout>}}_1^{*}+N_{23}{\mathrm{<figure-callout\; id="2"\; label="p"\; filenames="DE112019005731T5\_0084.png"\; state="\{\{state\}\}">p</figure-callout>}}_2^{*}\right)\cdot N_{turns}^{*}=1\\ N_{turns}^{*}\ge \frac{\mathrm{<figure-callout\; id="2"\; label="\Delta \; \theta "\; filenames="DE112019005731T5\_0084.png"\; state="\{\{state\}\}">\Delta </figure-callout>}{\theta }_{}{}_2}{360}\end{array}$$

where Δθ _{2 is} the peak-to-peak variation of the absolute steering wheel angle over more than one mechanical revolution and R _{red is} the reduction ratio of the reduction gear 4th is.

Die Verarbeitungseinheit 15 bewirkt eine gewichtete Winkelsumme der unterschiedlichen verfügbaren Signale, um das zweite berechnete Signal T zu erzeugen, das proportional zu dem Drehmoment T ist. Eine gewichtete Winkelsumme der Signale ist eine lineare Kombination von Signalen, wobei das finale Ergebnis in dem Intervall [-180, 180° [(oder in dem Intervall [-π;ττ [gemäß der Wahl der Winkeleinheit) verringert wird.If all the conditions described above are observed for a given embodiment, it is possible to demonstrate the approximate measurement errors and the approximate slope indications that the first calculated signal θ̂̂ ₂ equals the absolute steering wheel angle over more than one mechanical revolution θ ₂ is, namely: $${\widehat{\theta }}_2={\mathrm{<figure-callout\; id="2"\; label="\theta "\; filenames="DE112019005731T5\_0084.png"\; state="\{\{state\}\}">\theta </figure-callout>}}_2$$

The processing unit 15th effects a weighted angular sum of the different available signals to produce the second calculated signal T which is proportional to the torque T. A weighted angle sum of the signals is a linear combination of signals, the final result being reduced in the interval [-180, 180 ° [(or in the interval [-π; ττ [according to the choice of the angle unit).

The processing unit 15th causes a weighted angle sum of the third signal α ₃ and of one of the signals from the first signal α ₁ , second signal α ₂ and the first calculated signal θ̂ ₂ .

wobei f_s die vorstehend definierte mathematische Funktion in Sägezahnform und mit Steigung gleich 1 ist, a₁ das Signal ist, das durch den ersten Winkelpositionssensor 11 erzeugt wird, a₃ das Signal ist, das durch den dritten Winkelpositionssensor 13 erzeugt wird, und G die Steifigkeit des Torsionsstabs 7 ist, und wobei k₁, k₃ numerische Koeffizienten sind, die von dem Entwickler ausgewählt werden und die folgenden Bedingungen erfüllen müssen: $$\{\begin{array}{l}{k}_{1}und{k}_{3}relativeganzeZahlen\hfill \\ R_{red}{N}_1{k}_1+N_3{k}_3=0\hfill \\ \frac{360}{\left|k_3{N}_3\right|}>\mathrm{\Delta }{\theta }_{shift}\hfill \end{array}$$

wobei Δθ_shift die Spitze-zu-Spitze-Variation der Winkelverformung des Torsionsstabs 7 ist.According to a first variant embodiment, which is shown in 2 is illustrated, the second calculated signal T̂̂ is calculated using the third signal a ₃ and the first signal α ₁ . The second calculated signal T is: $$\widehat{T}=\frac{G}{\left|k_3{\mathrm{<figure-callout\; id="3"\; label="N"\; filenames="DE112019005731T5\_0084.png"\; state="\{\{state\}\}">N</figure-callout>}}_3\right|}\cdot f_s\left({\mathrm{<figure-callout\; id="1"\; label="k"\; filenames="DE112019005731T5\_0084.png"\; state="\{\{state\}\}">k</figure-callout>}}_1{\mathrm{<figure-callout\; id="1"\; label="\alpha "\; filenames="DE112019005731T5\_0084.png"\; state="\{\{state\}\}">\alpha </figure-callout>}}_1+{\mathrm{<figure-callout\; id="3"\; label="k"\; filenames="DE112019005731T5\_0084.png"\; state="\{\{state\}\}">k</figure-callout>}}_3{\alpha }_3\right)$$

where f _{s is} the mathematical function defined above in sawtooth form and with a slope equal to 1, a _{1 is} the signal that is transmitted by the first angular position sensor 11 is generated, a _{3 is} the signal generated by the third angular position sensor 13th is generated, and G is the rigidity of the torsion bar 7th and where k ₁ , k _{3 are} numerical coefficients selected by the designer and which must meet the following conditions: $$\{\begin{array}{l}{k}_{1}und{k}_{3}relativeGanzeZaHlen\hfill \\ {\mathrm{R.}}_{red}{N}_1{\mathrm{<figure-callout\; id="1"\; label="k"\; filenames="DE112019005731T5\_0084.png"\; state="\{\{state\}\}">k</figure-callout>}}_1+N_3{\mathrm{<figure-callout\; id="3"\; label="k"\; filenames="DE112019005731T5\_0084.png"\; state="\{\{state\}\}">k</figure-callout>}}_3=0\hfill \\ \frac{360}{\left|k_3{\mathrm{<figure-callout\; id="3"\; label="N"\; filenames="DE112019005731T5\_0084.png"\; state="\{\{state\}\}">N</figure-callout>}}_3\right|}>\mathrm{\Delta }{\theta }_{sHift}\hfill \end{array}$$

where Δθ _{shift is} the peak-to-peak variation in angular deformation of the torsion bar 7th is.

wobei f_s die vorstehend definierte mathematische Funktion in Sägezahnform und mit Steigung gleich 1 ist, α₂ das Signal ist, das durch den zweiten Winkelpositionssensor 12 erzeugt wird, a₃ das Signal ist, das durch den dritten Winkelpositionssensor 13 erzeugt wird, und G die Steifigkeit des Torsionsstabs 7 ist, und wobei k₂, k₃ numerische Koeffizienten sind, die von dem Entwickler ausgewählt werden und die folgenden Bedingungen erfüllen müssen: $$\{\begin{array}{l}{k}_{2}und{k}_{3}relativeganzeZahlen\hfill \\ N_2{k}_2+N_3{k}_3=0\hfill \\ \frac{360}{\left|k_3{N}_3\right|}>\mathrm{\Delta }{\theta }_{shift}\hfill \end{array}$$

wobei Δθ_shift die Spitze-zu-Spitze-Variation der Winkelverformung des Torsionsstabs 7 ist.According to a second embodiment variant, which is shown in 3 is illustrated, the second calculated signal T is calculated based on the third signal α ₃ and the second signal α ₂ . The second calculated signal T̂ is: $$\widehat{T}=\frac{G}{\left|k_3{\mathrm{<figure-callout\; id="3"\; label="N"\; filenames="DE112019005731T5\_0084.png"\; state="\{\{state\}\}">N</figure-callout>}}_3\right|}\cdot f_s\left({\mathrm{<figure-callout\; id="2"\; label="k"\; filenames="DE112019005731T5\_0084.png"\; state="\{\{state\}\}">k</figure-callout>}}_2{\mathrm{<figure-callout\; id="2"\; label="\alpha "\; filenames="DE112019005731T5\_0084.png"\; state="\{\{state\}\}">\alpha </figure-callout>}}_2+{\mathrm{<figure-callout\; id="3"\; label="k"\; filenames="DE112019005731T5\_0084.png"\; state="\{\{state\}\}">k</figure-callout>}}_3{\alpha }_3\right)$$

where f _{s is} the mathematical function defined above in sawtooth form and with a slope equal to 1, α _{2 is} the signal that is transmitted by the second angular position sensor 12th is generated, a _{3 is} the signal generated by the third angular position sensor 13th is generated, and G is the rigidity of the torsion bar 7th and where k ₂ , k _{3 are} numerical coefficients selected by the designer and which must meet the following conditions: $$\{\begin{array}{l}{k}_{2}und{k}_{3}relativeGanzeZaHlen\hfill \\ N_2{\mathrm{<figure-callout\; id="2"\; label="k"\; filenames="DE112019005731T5\_0084.png"\; state="\{\{state\}\}">k</figure-callout>}}_2+N_3{\mathrm{<figure-callout\; id="3"\; label="k"\; filenames="DE112019005731T5\_0084.png"\; state="\{\{state\}\}">k</figure-callout>}}_3=0\hfill \\ \frac{360}{\left|k_3{\mathrm{<figure-callout\; id="3"\; label="N"\; filenames="DE112019005731T5\_0084.png"\; state="\{\{state\}\}">N</figure-callout>}}_3\right|}>\mathrm{\Delta }{\theta }_{sHift}\hfill \end{array}$$

where Δθ _{shift is} the peak-to-peak variation in angular deformation of the torsion bar 7th is.

wobei f_s die vorstehend definierte mathematische Funktion in Sägezahnform und mit Steigung gleich 1 ist, θ̂₂ das erste berechnete Signal ist, a₃ das Signal ist, das durch den dritten Winkelpositionssensor 13 erzeugt wird, und G die Steifigkeit des Torsionsstabs 7 ist, und wobei k₃ ein numerischer Koeffizient ist, der ausgewählt ist und die folgenden Bedingungen erfüllen muss: $$\{\begin{array}{l}{k}_{3}relativeganzeZahl\hfill \\ \frac{360}{\left|k_3{N}_3\right|}>\mathrm{\Delta }{\theta }_{shift}\hfill \end{array}$$

wobei Δθ_shift die Spitze-zu-Spitze-Variation der Winkelverformung des Torsionsstabs 7 ist.According to a third embodiment variant, which is shown in 4th is illustrated, the second calculated signal T is calculated based on the third signal α ₃ and the first calculated signal θ̂̂ ₂ . The second calculated signal T̂̂ is: $$\widehat{T}=\frac{G}{\left|k_3{\mathrm{<figure-callout\; id="3"\; label="N"\; filenames="DE112019005731T5\_0084.png"\; state="\{\{state\}\}">N</figure-callout>}}_3\right|}\cdot f_s\left(N_3{\mathrm{<figure-callout\; id="3"\; label="k"\; filenames="DE112019005731T5\_0084.png"\; state="\{\{state\}\}">k</figure-callout>}}_3{\widehat{\theta }}_2+{\mathrm{<figure-callout\; id="3"\; label="k"\; filenames="DE112019005731T5\_0084.png"\; state="\{\{state\}\}">k</figure-callout>}}_3{\alpha }_3\right)$$

where f _{s is} the mathematical function defined above in sawtooth form and with a slope equal to 1, θ̂ _{2 is} the first calculated signal, a _{3 is} the signal that is transmitted by the third angular position sensor 13th is generated, and G is the rigidity of the torsion bar 7th is, and where k _{3 is} a numerical coefficient that is selected and must meet the following conditions: $$\{\begin{array}{l}{k}_{3}relativeGanzeZaHl\hfill \\ \frac{360}{\left|k_3{\mathrm{<figure-callout\; id="3"\; label="N"\; filenames="DE112019005731T5\_0084.png"\; state="\{\{state\}\}">N</figure-callout>}}_3\right|}>\mathrm{\Delta }{\theta }_{sHift}\hfill \end{array}$$

where Δθ _{shift is} the peak-to-peak variation in angular deformation of the torsion bar 7th is.

Aus der vorstehenden Beschreibung geht hervor, dass das Referenzdrehmoment und der Referenzlenkradwinkel ohne „physischen“ Sensor des Drehmoments und des Lenkradwinkels bestimmt werden. Tatsächlich wird der absolute Lenkradwinkel über mehr als eine mechanische Umdrehung θ₂ anhand des ersten Signals a₁, das durch den ersten Winkelpositionssensor 11 erzeugt wird, und des zweiten Signals α₂ bestimmt, das durch den zweiten Winkelpositionssensor 12 erzeugt wird, während das ausgeübte Drehmoment T anhand des dritten Signals α₃, das durch den dritten Winkelpositionssensor 13 erzeugt wird, und anhand von einem der Signale von dem ersten berechneten Signal α₁, dem zweiten Signal α₂ und dem dritten berechneten Signal θ̂₂ bestimmt wird.If all the conditions described above are observed (for a given embodiment variant), it is possible to demonstrate the approximate measurement errors and the approximate slope indications that the second calculated signal T is equal to the torque exerted, namely: $$\widehat{T}=T$$

From the above description it can be seen that the reference torque and the reference steering wheel angle are determined without a “physical” sensor of the torque and the steering wheel angle. In fact, the absolute steering wheel angle becomes over more than one mechanical turn θ ₂ based on the first signal a ₁ , which is generated by the first angular position sensor 11 is generated, and the second signal α ₂ determined by the second angular position sensor 12th is generated while the exerted torque T based on the third signal α ₃ , which is generated by the third angular position sensor 13th is generated, and based on a of the signals from the first calculated signal α ₁ , the second signal α ₂ and the third calculated signal θ̂ _{2 is} determined.

The first angular position sensor 11 , the second angular position sensor 12th and the third angular position sensor 13th are angular position sensors with Ni pole pairs (where Ni is an integer greater than or equal to 1) of all known types. For example, are the first angular position sensor 11 , the second angular position sensor 12th and / or the third angular position sensor 13th Sensors or combinations of position sensors with Hall effect, magnetic resistance, fluxgates, induction, with eddy current or with variable reluctance.

The complex and expensive sensors of the prior art are being replaced by much simpler angular position sensors and by specific signal processing. This replacement was made thanks to the elegant reuse of the first signal from the first angular position sensor 11 which is already available on most of the electric power steering systems and which controls the electric motor 3 of the power steering, and also thanks to the elegant re-use of the reduction gear of the electric motor, which is already available on all electric power steering systems. In addition, the registration system requires 1 according to the invention no initialization at start-up and does not require a continuous control system. The first calculated signal and the second calculated signal are available as soon as the angular position sensors are switched on. No previous steering movement is required to have these two calculated signals.

5 until 7th illustrate the subject matter of the invention by implementing a numerical calculation according to the principles described above with the following dimensions: Δθ ₂ = 1000 ° mec, ΔT = 21 Nm, G = 3 Nm / °, R _red = 61/3, N ₁ = 1, N ₂ = 10, N ₃ = 20, p ₁ = 1, p ₂ = -2, N _turns = 3, k ₂ = -2, k ₃ = 1, q ₁ = 0, q ₂ = 0. The second variant was used to calculate the torque.

5 Fig. 10 shows the shape of the first, second and third signals a ₁ , a ₂ , a ₃ , each given by the three angular position sensors 11 , 12th , 13th are generated depending on the reference steering wheel angle in an ideal case where there is no noise and no measurement error.

6th illustrates the first calculated signal θ̂̂ ₂ proportional to the absolute steering wheel angle over more than one mechanical revolution for different values of the torque. The first calculated signal θ̂ ₂ depends linearly on the absolute steering wheel angle over more than one mechanical revolution θ ₂ from, whereby it is completely insensitive to the torque.

7th illustrates the second calculated signal T̂ proportional to the torque for different values of the steering wheel angle. The second calculated signal T depends linearly on the torque T exerted, being completely insensitive to the steering wheel angle.

This numerical modeling makes it possible to demonstrate that the first calculated signal and the second calculated signal each correspond to the absolute steering wheel angle over more than one mechanical revolution θ ₂ and the applied torque T.

It should be noted that the figures illustrate simulations implemented with no noise. Further simulations in the presence of noise have been implemented. These simulations have shown that the present invention is robust to this noise, including at mechanical angles or the first, second and third signals a ₁ , a ₂ , a ₃ in sawtooth shape that have interruptions. This means that the noise that is present at the input is transferred to the output without being significantly amplified. In the case in which the weighting coefficients q ₁ and q _{2 are} sensibly selected, the use of the Vernier effect even makes it possible to greatly improve the signal-to-noise ratio at the output.

It should be taken into account that the acquisition system according to the invention is based on input data which are output from the specifications of each desired application. Thus, the peak-to-peak variations Δθ ₂ and ΔT of the steering wheel angle and the torque correspond to the scope of the measurement system required by the specification. These input data can of course represent different values depending on the intended application.

Ausgehend von den Konzeptionsparametern ist es möglich, den folgenden Schwellenwert D_lim zu definieren: $$D_{lim}=\frac{\lambda 360}{2N_1{R}_{red}{N}_{turns}}\text{cos}\left(atan\left(\frac{N_2}{N_1{R}_{red}}\right)\right)$$

wobei λ ein fester Koeffizient ist, der zwischen 0 und 1 beträgt, der von dem Entwickler ausgewählt wird und der ermöglicht, die Schwere des Diagnosesystems zu regulieren. Somit ist, wenn die Verarbeitungseinheit in einem beliebigen Moment das Ereignis erfasst, $$D>D_{lim}$$

Somit liefert die Verarbeitungseinheit 15 ein Warnsignal.According to a particularly advantageous embodiment, the processing unit verifies 15th that the set of measured and calculated signals belongs to a set of permissible values. If this is not the case, the processing unit delivers 15th a warning signal if the set of measured and calculated signals does not belong to the set of permissible values. This approach enables certain categories of errors to be captured from unusually large measurements. For example, the processing unit can calculate the following signal D: $$\mathrm{D.}=\sqrt{f_s{\left({\mathrm{R.}}_{r\mathrm{<figure-callout\; id="1"\; label="e\; d\; N"\; filenames="DE112019005731T5\_0084.png"\; state="\{\{state\}\}">e</figure-callout>}d}{N}_{}{}_1{\widehat{\theta }}_2-{\alpha }_1\right)}^2+f_s{\left({\mathrm{<figure-callout\; id="2"\; label="N"\; filenames="DE112019005731T5\_0084.png"\; state="\{\{state\}\}">N</figure-callout>}}_2{\widehat{\theta }}_2-{\alpha }_2\right)}^2}$$

Based on the design parameters, it is possible to define _{the following threshold value D lim:} $${\mathrm{D.}}_{lim}=\frac{\lambda 360}{2{\mathrm{<figure-callout\; id="1"\; label="N"\; filenames="DE112019005731T5\_0084.png"\; state="\{\{state\}\}">N</figure-callout>}}_1{\mathrm{R.}}_{red}{N}_{turns}}\text{cos}\left(atan\left(\frac{N_2}{{\mathrm{<figure-callout\; id="1"\; label="N"\; filenames="DE112019005731T5\_0084.png"\; state="\{\{state\}\}">N</figure-callout>}}_1{\mathrm{R.}}_{red}}\right)\right)$$

where λ is a fixed coefficient, between 0 and 1, selected by the designer, which allows the severity of the diagnostic system to be regulated. Thus, if at any moment the processing unit detects the event, $$\mathrm{D.}>{\mathrm{D.}}_{lim}$$

Thus the processing unit delivers 15th a warning sign.

According to one embodiment variant, the second angular position sensor 12th and the third angular position sensor 13th Eddy current position sensors. 8th exemplifies the second angular position sensor 12th and the third angular position sensor 13th that are on either side of the torsion bar 7th are arranged. Any angular position sensor 12th , 13th contains on the one hand a respective goal 12 ₁ , 13 ₁ that is firmly on each side of the torsion bar 7th is mounted, and on the other hand, a respective detection probe 12 ₂ , 13 ₂ that in relation to the corresponding target outside in relation to the torsion bar 7th is placed.

According to an advantageous embodiment variant, which is shown in 9 illustrated have the second angular position sensor 12th with eddy current and the third angular position sensor 13th with eddy current a common detection probe 23 that between goals 12 ₁ , 13 ₁ the two sensors are placed, i.e. at the height of the torsion bar 7th . This common acquisition probe 23 includes a common support plate for the windings of the second angular position sensor 12th and the third angular position sensor 13th . Implementation of the windings of the two angular position sensors 12th , 13th enables manufacturing costs to be reduced using a single circuit carrier. The crosstalk between these two angular position sensors can be canceled by using a number of pole pairs N ₂ other than N ₃ , or by inserting a conductive and / or magnetic material between the windings that enables the magnetic field to be decoupled from these two sensors.

The invention is not restricted to the examples described and illustrated, since various modifications can be made without departing from the scope.

QUOTES INCLUDED IN THE DESCRIPTION

This list of the documents listed by the applicant was 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.

Patent literature cited

• FR 2964190 [0006]
• EP 3090921 [0007]
• FR 2872896 [0008]
• DE 102009039764 [0010]

Claims (12)

Detection system for a steering system (2) of a vehicle, which enables the torque (T) and the absolute steering wheel angle to be measured over more than one mechanical revolution (θ̂ ₂ ), this steering system including a torsion bar (7) and an electric motor (3) equipped with a reduction gear (4), this detection system comprising: - a first angular position sensor (11) having N1 pole pairs, where N1 is an integer greater than or equal to 1, this first angular position sensor (11) den Measures the angle of the electric motor (3) and _{delivers a first signal (α 1} ), - a second angular position sensor (12), which has N2 pole pairs, where N2 is an integer greater than or equal to 1, this second angular position sensor (12) measuring the angle the steering measures, which is located between the reduction gear (4) and a first side of the torsion bar (7), this second angular position sensor providing a second signal (α ₂ ) fert, - a third angular position sensor (13), which has N3 pole pairs, where N3 is an integer greater than or equal to 1, this third angular position sensor (13) measuring the angle of the steering, which is on a second side of the torsion bar (7) which is located opposite the first side and _{delivers a third signal (α 3} ), - and a processing unit (15) which, on the one hand, a Vernier calculation on the basis of at least the first signal (α ₁ ) and the second signal (α ₂ ) to generate a first calculated signal (θ̂ ₂ ) proportional to the absolute steering wheel angle over more than one mechanical revolution (θ̂ ₂ ), and on the other hand a weighted angle sum of the third signal (α ₃ ) and from one the signals from the first signal (α ₁ ), the second signal (α ₂ ) and the first calculated (θ̂ ₂ ) signal to produce a second calculated signal (f) proportional to the torque (T). System according to the preceding claim, wherein the processing unit (15) assumes that the first calculated (θ̂ ₂ ) signal corresponds to the absolute steering wheel angle over more than one mechanical revolution (θ̂ ₂ ) and that the second calculated (T̂) signal corresponds to the torque exerted (T) corresponds to. System according to any one of the preceding claims, wherein the processing unit (15) effects the vernier calculation to generate the first calculated signal (θ̂ ₂ ) which reads: $${\widehat{\theta }}_2=N_{turns}\cdot f_s\left(\mathrm{\Theta }-q_1\cdot f_s\left(N_1{\mathrm{R.}}_{red}{N}_{turns}\mathrm{\Theta }-{\alpha }_1\right)-q_2\cdot f_s\left(N_2{N}_{turns}\mathrm{\Theta }-{\alpha }_2\right)\right)$$

where: θ = f _s (p ₁ α ₁ + p ₂ α ₂ ) and f _{s is} the mathematical function in sawtooth form and with a slope equal to 1, where q ₁ and q _{2 are} fixed selected weighting coefficients, and α _{1 is} the signal that is generated by the first angular position sensor (11), α _{2 is} the signal generated by the second angular position sensor (12), and where p ₁ , p ₂ , and N _{turns are} numerical coefficients selected and satisfying the following conditions have to: $$\{\begin{array}{c}{p}_{1}und{p}_{2}relativeGanzeZaHlen\\ \left({\mathrm{R.}}_{red}{N}_1{p}_1+N_2{p}_2\right)\cdot N_{turns}=1\\ N_{turns}\ge \frac{\mathrm{\Delta }{\theta }_2}{360}\end{array}$$

where Δθ _{2 is} the peak-to-peak variation of the absolute steering wheel angle over more than one mechanical revolution and R _{red is} the reduction ratio of the reduction gear (4). System according to Claim 3 , where the weighting coefficients q ₁ and q ₂ are: $$\begin{array}{l}wenn{N}_2<N_1{\mathrm{R.}}_{red}\\ \{\begin{array}{c}{q}_1=0\\ q_2=\frac{1}{p_1{N}_{turns}}\cdot \frac{{\sigma }_1^2{p}_1{N}_2-{\sigma }_2^2{p}_2{N}_1{\mathrm{R.}}_{red}}{{\left({\sigma }_1{N}_2\right)}^2+{\left({\sigma }_2{N}_1{\mathrm{R.}}_{red}\right)}^2}\end{array}\end{array}$$

$$\begin{array}{l}wenn{N}_2>N_1{\mathrm{R.}}_{red}\\ \{\begin{array}{c}{q}_1=\frac{1}{p_2{N}_{turns}}\cdot \frac{{\sigma }_2^2{p}_2{N}_1{\mathrm{R.}}_{red}-{\sigma }_1^2{p}_1{N}_2}{{\left({\sigma }_1{N}_2\right)}^2+{\left({\sigma }_2{N}_1{\mathrm{R.}}_{red}\right)}^2}\\ q_2=0\end{array}\end{array}$$

where σ _{1 is} the typical error amplitude of the first angular position sensor and where σ _{2 is} the typical error amplitude of the second angular position sensor. System according to one of the Claims 1 or 2 wherein the processing unit (15) effects the vernier computation to generate the first computed signal (θ̂ ₂ ) which is: $${\widehat{\theta }}_2=N_{{}_{turns}}^{*}\cdot f_s\left({\mathrm{\Theta }}^{\mathrm{\ast }}-q_1^{*}\cdot f_s\left(N_1{\mathrm{R.}}_{red}{N}_{turns}^{*}{\mathrm{\Theta }}^{\mathrm{\ast }}-{\alpha }_1\right)-q_2^{*}\cdot f_s\left(N_{23}{N}_{turn}^{*}{\mathrm{\Theta }}^{\mathrm{\ast }}-{\alpha }_2^{*}\right)\right)$$

whereby: - $${\mathrm{\Theta }}^{*}=f_s\left(p_1^{*}{\alpha }_1+p_2^{*}{\alpha }_2^{*}\right)$$

- $${\alpha }_2^{*}=f_s\left(f_s\left(c_2\cdot f_s\left({\alpha }_2+n_3\delta \right)+c_3\cdot f_s\left({\alpha }_3-n_2\delta \right)\right)-\frac{n_3}{n_2}\delta \right)$$

- $$\delta =\frac{f_s\left(-n_3{\alpha }_2+n_2{\alpha }_3\right)}{n_2^2+n_3^2}$$

- $$\{\begin{array}{c}{N}_{23}=\text{pgcd}\left(N_2,N_3\right)\\ n_2=\frac{N_2}{N_{23}}\\ n_3=\frac{N_3}{N_{23}}\end{array}$$

and f _{s is} the mathematical function in sawtooth form and with a slope equal to 1, where pgcd is the operator "greatest common divisor", where $q_1^{*}$

and $q_2^{*}$

are fixed selected weighting coefficients, and α _{1 is} the signal generated by the first angular position sensor (11), α _{2 is} the signal generated by the second angular position sensor (12), α _{3 is} the signal generated by the third Angular position sensor (13) is generated, and where c ₂ , c ₃ , $p_1^{*},$

$p_2^{*}$

and N _{turns are} numeric coefficients that are selected and must meet the following conditions: $$\{\begin{array}{c}\begin{array}{l}{c}_{2}und{c}_{3}relativeGanzeZaHlen\\ p{*}_{1}undp{*}_{2}relativeGanzeZaHlen\end{array}\\ c_2{n}_2+c_3{n}_3=1\\ \left({\mathrm{R.}}_{red}{N}_1{p}_1^{*}+N_{23}{p}_2^{*}\right)\cdot N_{turns}^{*}=1\\ N_{turns}^{*}\ge \frac{\mathrm{\Delta }{\theta }_2}{360}\end{array}$$

where Δθ _{2 is} the peak-to-peak variation of the absolute steering wheel angle over more than one mechanical revolution and R _{red is} the reduction ratio of the reduction gear (4). System according to one of the preceding claims, wherein the second calculated signal (T̂) is: $$\widehat{T}=\frac{G}{\left|k_3{N}_3\right|}\cdot f_s\left(k_2{\alpha }_2+k_3{\alpha }_3\right)$$

where f _{s is} the mathematical function in sawtooth form and with a slope equal to 1, α _{2 is} the signal generated by the second angular position sensor (12), α _{3 is} the signal generated by the third angular position sensor (13), and G is the stiffness of the torsion bar (7), and where k ₂ , k _{3 are} numerical coefficients that are selected and must meet the following conditions: $$\{\begin{array}{l}{k}_{2}und{k}_{3}relativeGanzeZaHlen\hfill \\ N_2{k}_2+N_3{k}_3=0\hfill \\ \frac{360}{\left|k_3{N}_3\right|}>\mathrm{\Delta }{\theta }_{sHift}\hfill \end{array}$$

where Δθ _{shtft is} the peak-to-peak variation in angular deformation of the torsion bar. System according to one of the Claims 1 until 5 , where the second calculated signal (T̂) is: $$\widehat{T}=\frac{G}{\left|k_3{N}_3\right|}\cdot f_s\left(k_1{\alpha }_1+k_3{\alpha }_3\right)$$

where f _{s is} the mathematical function in sawtooth form and with a slope equal to 1, α _{1 is} the signal generated by the first angular position sensor (11), α _{3 is} the signal generated by the third angular position sensor (13), and G is the stiffness of the torsion bar (7), and where k ₁ , k _{3 are} numerical coefficients that are selected and must meet the following conditions: $$\{\begin{array}{l}{k}_{2}und{k}_{3}relativeGanzeZaHlen\hfill \\ {\mathrm{R.}}_{red}{N}_1{k}_1+N_3{k}_3=0\hfill \\ \frac{360}{\left|k_3{N}_3\right|}>\mathrm{\Delta }{\theta }_{sHift}\hfill \end{array}$$

where Δθ _{shift is} the peak-to-peak variation in angular deformation of the torsion bar. System according to one of the Claims 1 until 5 , where the second calculated signal (T̂) is: $$\widehat{T}=\frac{G}{\left|k_3{N}_3\right|}\cdot f_s\left(N_3{k}_3{\widehat{\theta }}_2+k_3{\alpha }_3\right)$$

where f _{s is} the mathematical function in sawtooth form and with a slope equal to 1, θ̂ _{2 is} the first calculated signal, a _{3 is} the signal generated by the third angular position sensor (13), and G is the stiffness of the torsion bar (7) , and where k _{3 is} a numerical coefficient that is selected and must meet the following conditions: $$\{\begin{array}{l}{k}_{3}relativeGanzeZaHl\hfill \\ \frac{360}{\left|k_3{N}_3\right|}>\mathrm{\Delta }{\theta }_{sHift}\hfill \end{array}$$

where Δθ _{shtft is} the peak-to-peak variation of the angular deformation of the torsion bar (7). System according to one of the preceding claims, wherein the processing unit (15) verifies that the set of measured and calculated signals belongs to a set of admissible values, wherein the processing unit (15) delivers a warning signal if the set of values does not belong to a set of admissible values listened to. System according to one of the preceding claims, characterized in that the first angular position sensor (11), the second angular position sensor (12) and / or the third angular position sensor (13) position sensors with Hall effect, magnetic resistance, flux gates, induction, with eddy current or with variable Are reluctance. System according to one of the preceding claims, characterized in that the second angular position sensor (12) and the third angular position sensor (13) are sensors in which combinations of eddy current position sensors include a common detection probe (23) which has a common support plate for the windings of the second angular position sensor (12) and the third angular position sensor (13). Steering with the detection system (1) according to one of the Claims 1 until 11 and which executes a steering command as a function of the absolute steering wheel angle over more than one mechanical revolution (θ̂ ₂ ) and the torque (T).

Images