Search

EP-4377715-B1 - HEIGHT ESTIMATION METHOD FOR OBJECTS BY MEANS OF ULTRASONIC SENSOR SYSTEM

EP4377715B1EP 4377715 B1EP4377715 B1EP 4377715B1EP-4377715-B1

Inventors

  • SULEIMAN, WASSIM
  • BROWN, CHRISTOPHER
  • Glatz, Nicolai

Dates

Publication Date
20260506
Application Date
20220629

Claims (15)

  1. A method for the estimation of the height of an object (3) by means of an ultrasonic sensor system of a vehicle (1), comprising the following steps: a) receiving at least two ultrasonic signals by way of at least one ultrasonic sensor (2), wherein a single ultrasonic sensor (2) has different sensor positions relative to the object (3) due to a vehicle movement, or wherein multiple ultrasonic sensors (2) have different sensor positions (P1, P2) relative to the object (3) due to the vehicle movement and/or a different arrangement on the vehicle (1) (S10); b) calculating a first item of height information (H), which is a measure of the squared height (h) of the object, based on two items of distance information (r1, r2) measured between the respective sensor position (P1, P2) and the object (3) based on the at least two ultrasonic signals and an item of distance information (s) measured horizontally between the sensor positions (S11); c) calculating the variance of the first item of height information (S12); d) receiving at least one further ultrasonic signal by way of the at least one ultrasonic sensor (2) and calculating a second item of height information, which is a measure of the squared height (h) of the object, based on two items of distance information (r2, r3) measured between the respective sensor position (P2, P3) and the object (3), at least partially based on the at least one further ultrasonic signal, and an item of distance information (s) measured horizontally between the sensor positions (S13); e) calculating the variance of the second item of height information (S14); f) calculating an averaged item of height information by combining the first and second item of height information and an averaged variance of the item of height information by combining the variance of the first and second item of height information (S15); g) classifying a detected object (3) into a height class by calculating at least one probability value (p) based on a normal distribution function, which has the averaged item of height information as a mean value and the averaged variance of the item of height information as a variance (S16).
  2. The method as claimed in claim 1, characterized in that further height information (H), which is a measure of the squared height (h) of the object, and variance information relating to this further height information are calculated iteratively, and in that the averaged item of height information is determined by combining the height information and an averaged variance of the height information is determined by combining the variances of the height information.
  3. The method as claimed in claim 1 or 2, characterized in that the first, second, and/or further item of height information (H) is calculated by the following formula: H = h 2 = r 1 2 − r 1 2 − r 2 2 + s 2 2 4 s 2 ; wherein the following applies: h: height difference between the at least one ultrasonic sensor and the object; r 1 : distance between a first transmitter position and the object in the transmitting and receiving directions; r 2 : distance between a second transmitter position and the object in the transmitting and receiving directions; s: item of distance information measured in the horizontal direction as the distance between the first and second sensor positions.
  4. The method as claimed in claim 1 or 2, characterized in that the variance of the first, second, and/or further item of height information is ascertained based on a first-order variation analysis.
  5. The method as claimed in any one of the preceding claims, characterized in that the calculation of the variance of the first, second, and/or further item of height information (H) is carried out by the following equation: Var H = dH dr 1 2 ⋅ Var r 1 + dH dr 2 2 ⋅ Var r 2 + dH ds 2 ⋅ Var s ; wherein the following applies: r 1 : distance between a first transmitter position and the object in the transmitting and receiving directions; r 2 : distance between a second transmitter position and the object in the transmitting and receiving directions; s: item of distance information measured in the horizontal direction as the distance between the first and second sensor positions; dH dr 1 : first derivative of the item of height information H according to r 1 ; dH dr 2 : first derivative of the item of height information H according to r 2 ; dH ds : first derivative of the item of height information H according to s; Var [ r 1 ]:variance of the distance between a first transmitter position and the object in the transmitting and receiving directions; Var [ r 2 ]:variance of the distance between a second transmitter position and the object in the transmitting and receiving directions; Var [ s ]: variance of the distance information measured in the horizontal direction as the distance between the first and second sensor positions.
  6. The method as claimed in any one of the preceding claims, characterized in that the calculating of the averaged item of height information is carried out based on the following formula: H ¯ = Var H " ⋅ H ′ + Var H ′ ⋅ H " Var H ′ + Var H " ; wherein the following applies: H': estimated item of height information from a first measurement cycle; H": estimated item of height information from a second measurement cycle; Var[H']: variance of the item of height information in the first measurement cycle; Var[H"]: variance of the item of height information in the second measurement cycle.
  7. The method as claimed in any one of the preceding claims, characterized in that the calculating of the averaged item of height information is carried out based on the following formula: Var H ¯ = 1 1 Var H ′ + 1 Var H " ; wherein the following applies: Var[H']: variance of the item of height information in the first measurement cycle; Var[H"]: variance of the item of height information in the second measurement cycle.
  8. The method as claimed in any one of claims 1 to 5, characterized in that the calculating of the averaged height information and of the averaged variance of the item of height information is carried out based on a method of averaged least square errors.
  9. The method as claimed in any one of the preceding claims, characterized in that the calculating of the probability value (p) for assignment of the object to a height class is carried out based on the following formula: p = ∫ a b , N x H ¯ Var H ¯ dx ; wherein the following applies: N ( x, H , Var [ H ] ): normal distribution; a: lower limit value for assignment to the respective height class; b: upper limit value for assignment to the respective height class; H : averaged item of height information; Var [ H ] : averaged variance.
  10. The method as claimed in any one of the preceding claims, characterized in that the calculation of the height class is based on a correction function which takes into consideration the deviation of the statistical distribution of the item of height information (H) from the normal function.
  11. The method as claimed in claim 10, characterized in that the correction function is estimated based on a data series of height information (H) which has been ascertained based on different distance information between the respective sensor position (P1, P2) and the object (3) and different horizontally measured distance information (s).
  12. The method as claimed in claim 10 or 11, characterized in that a lower and/or upper limit value used for the calculation of the probability value (p) is adapted based on the correction function.
  13. The method as claimed in any one of the preceding claims, characterized in that the object (3) is assumed to be a line object having a longitudinal orientation and the transmitting and receiving directions of the at least one ultrasonic sensor (2) and the direction in which the distance information (s) is measured are assumed to be perpendicular to this longitudinal alignment of the line object.
  14. The method as claimed in any one of claims 1 to 12, characterized in that the object (3) is replicated by an object contour line by means of information that has been ascertained by the ultrasonic sensor system in multiple detection cycles, and the distance information (s) is assumed to be the difference between the horizontally measured distance of the sensor positions (P1, P2) and the line.
  15. A system for the estimation of the height of an object (3), comprising an ultrasonic sensor system provided on a vehicle (1) and a computer unit (4), wherein the system is designed to carry out the following steps: a) receiving at least two ultrasonic signals by way of at least one ultrasonic sensor (2) of the ultrasonic sensor system, wherein a single ultrasonic sensor (2) has different sensor positions (P1, P2) relative to the object (3) due to a vehicle movement, or wherein multiple ultrasonic sensors (2) have different sensor positions (P1, P2) relative to the object (3) due to the vehicle movement and/or a different arrangement on the vehicle (1); b) calculating a first item of height information (H), which is a measure of the squared height (h) of the object, based on two items of distance information measured between the respective sensor position (P1, P2) and the object (3) based on the at least two ultrasonic signals and an item of distance information (s) measured horizontally between the sensor positions (P1, P2) by the computer unit (4); c) calculating the variance of the first item of height information (H) by way of the computer unit (4); d) receiving at least one further ultrasonic signal by way of the at least one ultrasonic sensor (2) and calculating a second item of height information, which is a measure of the squared height (h) of the object, based on two items of distance information (r2, r3) measured between the respective sensor position (P2, P3) and the object (3), at least partially based on the at least one further ultrasonic signal, and an item of distance information (s) measured horizontally between the sensor positions; e) calculating the variance of the second item of height information; f) calculating an averaged item of height information by combining the first and second item of height information and an averaged variance of the item of height information by combining the variance of the first and second item of height information; g) classification of a detected object (3) into a height class by calculating at least one probability value (p) by way of the computer unit (4) based on a normal distribution function, which has the averaged item of height information as a mean value and the averaged variance of the item of height information as a variance.

Description

Die Erfindung betrifft ein Verfahren zur Höhenschätzung von Objekten im Umgebungsbereich von Fahrzeugen mittels einer Ultraschallsensorik. Es ist bekannt, Umgebungsinformationen im Bereich eines Fahrzeugs mittels Ultraschallsensoren zu erfassen, beispielsweise um beim Einparken den Abstand zu anderen Objekten zu erfassen. Ultraschallsensoren liefern basierend auf der Laufzeit des Ultraschallsignals Abstandsinformationen, jedoch keine direkten Höheninformationen des detektierten Objekts, da der Empfangswinkel, aus dem die Rückreflexion erfolgt, nicht ermittelbar ist. Es ist jedoch bekannt, die Höhe des Objekts, an dem die Reflexion erfolgt, durch Auswertung der geometrischen Zusammenhänge aus mehreren Sende-Empfangszyklen zu ermitteln. So kann beispielsweise bei einer Bewegung des Fahrzeugs zwischen den beiden Sende-Empfangszyklen durch Auswertung der gemessenen Abstandsinformationen zwischen dem Ultraschallsensor und dem reflektierenden Objekt und einer Distanzinformation, die den Abstand der Sensorpositionen des Ultraschallsensors in horizontaler Richtung entlang der Ausbreitungsrichtung des Ultraschallsignals zwischen dem Ultraschallsensor und dem reflektierenden Objekt angibt, die Höhe des Objekts berechnet werden. Ein solches Verfahren ist beispielsweise aus DE 101 51 965 A1 bekannt. Ahnliche Verfahren finden sich in DE 10 2018 218007 A1 und in DE 10 2014 114999 A1. Problematisch dabei ist, dass sich durch Rauschen häufig schlechte Messergebnisse ergeben, die zu einer unzureichenden Höhenschätzung mit hohen Abweichungen führen. Ausgehend hiervon ist es Aufgabe der Erfindung, ein Verfahren zur Höhenschätzung von Objekten im Umgebungsbereich von Fahrzeugen mittels Ultraschallsensorik anzugeben, das eine zuverlässige Schätzung der Höhe des reflektierenden Objekts erlaubt. Die Aufgabe wird durch ein Verfahren mit den Merkmalen des unabhängigen Patentanspruchs 1 gelöst. Bevorzugte Ausführungsformen sind Gegenstand der Unteransprüche. Ein System zur Höhenschätzung eines Objekts ist Gegenstand des nebengeordneten Patentanspruchs 15. Gemäß einem ersten Aspekt bezieht sich die Erfindung auf ein Verfahren zur Höhenschätzung eines Objekts mittels einer Ultraschallsensorik eines Fahrzeugs. Das Verfahren umfasst die folgenden Schritte: Zunächst werden zumindest zwei Ultraschallsignale durch zumindest einen Ultraschallsensor empfangen. Dabei kann ein einziger Ultraschallsensor aufgrund einer Fahrzeugbewegung unterschiedliche Sensorpositionen relativ zum Objekt aufweisen. In anderen Worten werden also Empfangsinformationen eines einzelnen Ultraschallsensors ausgewertet, der aufgrund der Fahrzeugbewegung an zeitlich aufeinanderfolgenden Messzyklen unterschiedliche Positionen relativ zum Objekt, dessen Höhe ermittelt werden soll, einnimmt. Alternativ können auch mehrere Ultraschallsensoren die zumindest zwei Ultraschallsignale empfangen, wobei diese aufgrund der Fahrzeugbewegung und/oder einer unterschiedlichen Anordnung am Fahrzeug unterschiedliche Sensorpositionen relativ zum Objekt aufweisen. Anschließend wird eine erste Höheninformation berechnet, die ein Maß für die quadrierte Höhe des Objekts ist. Die Höhe des Objekts ist dabei vorzugsweise die relative Höhe des Objekts zu der Einbauhöhe des zumindest einen Ultraschallsensors an dem Fahrzeug. In anderen Worten wird damit nicht direkt die Höhe des Objekts bestimmt, sondern die quadrierte Höhe, um dadurch Fehlschätzungen, die durch Rauschen bedingt sind, zu vermeiden. Die Berechnung der ersten Höheninformation erfolgt basierend auf zwei zwischen der jeweiligen Sensorposition und dem Objekt gemessenen Abstandsinformationen und einer zwischen den Sensorpositionen horizontal gemessenen Distanzinformation. Die erste Höheninformation wird beispielsweise in einem ersten Messzyklus ermittelt. Anschließend erfolgt das Berechnen der Varianz der ersten Höheninformation. Daraufhin wird zumindest ein weiteres Ultraschallsignal durch den zumindest einen Ultraschallsensor empfangen. Es wird basierend auf dieser weiteren Ultraschallmessung eine zweite Höheninformation berechnet, die ein Maß für die quadrierte Höhe des Objekts ist. Die Berechnung erfolgt wiederum basierend auf zwei zwischen der jeweiligen Sensorposition und dem Objekt gemessenen Abstandsinformationen und einer zwischen den Sensorpositionen horizontal gemessenen Distanzinformation. Zumindest ein Teil dieser für die Berechnung nötigen Informationen stammt aus der weiteren Ultraschallmessung, die beispielsweise einen zweiten Messzyklus bildet. Der zweite Messzyklus kann dabei ein unmittelbar auf den ersten Messzyklus folgender Messzyklus sein oder zwischen dem ersten und zweiten Messzyklus können weitere Messzyklen liegen. Bei nicht unmittelbar aufeinander folgenden ersten und zweiten Messzyklen wird beispielsweise erreicht, dass sich das Fahrzeug zwischen dem ersten und zweiten Messzyklus auf einer längeren Wegstrecke bewegt, was die Genauigkeit der Höhenschätzung erhöht. Anschließend erfolgt das Berechnen der Varian