Method for determining the tilt angle of a vehicle

Abstract

A method for determining the tilting of a vehicle due to a load by determining the mass of the vehicle by determining the natural frequency of the vehicle during a braking process is described, which method comprises the following steps: Detection of a braking process, Detecting the resulting standstill of the vehicle, Determining the natural frequency of the vehicle using a sensor by measuring the vibration of the vehicle directly after it has come to a standstill and evaluating the sensor signal, Determining the total mass from the natural frequency using a vehicle model, Determining the number of people sitting in the vehicle and their distribution using a seat occupancy detection sensor, Determine the residual load by subtracting the number of people multiplied by an assumed average weight from the previously determined total mass, assuming that the residual load is in the trunk and Determining the tipping from the distribution of people and the residual load using a bar model.

Description

[0001] The aim is to determine the tilt of a vehicle due to a load, particularly in the rear area, e.g., in the trunk, but also due to people in the back seat. The determined vehicle tilt is to be used, for example, for headlight range control, chassis leveling, or a hill-hold assistant.

[0002] Until now, chassis sensors have been used for automatic headlight range adjustment to determine vehicle tilt. However, for cost reasons, attempts are being made to replace these sensors with other vehicle sensors. One possibility is to use inertial sensors such as acceleration sensors or yaw rate sensors. These are often referred to as IMU sensors (IMU inertial measurement unit). However, commercial IMU sensors have the disadvantage of high drift and also measure the inclination of the road surface. Other methods are required to compensate for these disadvantages. A headlight range control based on IMU sensors is disclosed, for example, in EP 2 402 212 B1. There, however, the headlight range is only adjusted when the vehicle is stationary, while the optical axis of the headlights is intended to remain unchanged when the vehicle is moving on a road.

[0003] US 2011 / 0 178 673 A1 concerns a vehicle having a wheel and a body connected by a suspension. The suspension contains a spring. A vibration detection unit measures how the body moves forward and backward. Another unit detects the body's natural frequency—the frequency at which it vibrates most strongly—based on the measured movements. A third unit then determines the body's weight based on this natural frequency.

[0004] DE 10 2011 081 395 A1 relates to a method for adjusting the beam range of a vehicle headlight. The method comprises a step of providing a signal representing a pitch angle of the vehicle and a step of changing the beam range of the headlight in response to the provided signal.

[0005] It is the object of the invention to provide a method for determining the tilt angle of a vehicle about its transverse axis due to a load, which can be carried out in a simple manner.

[0006] The task is solved by a method for determining the tilting of a vehicle due to a load by determining the mass of the vehicle by determining the natural frequency of the vehicle during a braking process with the following steps: Detection of a braking process, Detecting the resulting standstill of the vehicle, Determining the natural frequency of the vehicle using a sensor by measuring the vibration of the vehicle directly after it has come to a standstill and evaluating the sensor signal, Determining the total mass from the natural frequency using a vehicle model, Determining the number of people sitting in the vehicle and their distribution using sensors, in particular a seat occupancy detection sensor, Determine the residual load by subtracting the number of people multiplied by an assumed average weight from the previously determined total mass, assuming that the residual load is in the trunk and Determining the tipping from the distribution of people and the residual load using a bar model.

[0007] After a sufficiently strong braking maneuver, a vibration of the vehicle can be detected. Since no further forces act on the vehicle after it has come to a standstill, this vibration must represent the vehicle's natural frequency. Braking and standstill are determined using various characteristic signals, and the frequency is calculated using, for example, an FFT (fast Fourier transformation).

[0008] The natural frequency is largely determined by the vehicle's spring constants and the total mass, which results from the vehicle's weight and the weight of the payload. Since the spring constants do not change, it is possible to extract the mass from an equation describing this relationship and subsequently determine the vehicle's payload.

[0009] The load distribution is determined using a seat occupancy detection sensor, which is usually found in modern vehicles. A person is assumed to weigh, for example, 75 kg. The remaining load is assumed to be in the trunk, whose position and thus distance from the vehicle's center of gravity is known. The tilt can then be calculated using a physical beam model.

[0010] In an advantageous embodiment of the method, the sensor for determining the natural frequency is an IMU (inertial measurement unit) sensor.

[0011] Such sensors are usually present in today's vehicles, so that vehicle tilt can be advantageously determined without additional sensors.

[0012] The invention is described in more detail below using an exemplary embodiment with the aid of figures. Fig. 1 a measurement diagram of different vehicle sizes during a braking process, Fig. 2 a schematic representation of a simple vehicle model and Fig. 3 a schematic representation of a simple beam model. Fig. 4 more illustrative illustration using a vehicle sketch Fig. 5 Forces on a vehicle sketch from the side Fig. 6 Forces on a vehicle sketch on a slope

[0013] When a vehicle brakes to a standstill, an oscillation in acceleration can be measured at the moment the vehicle actually reaches the fully stopped state, as the brakes switch from sliding friction to static friction, resulting in much higher forces that induce an impulse into the mass of the vehicle. This is in the Fig. 1 can be recognized as a vibration.

[0014] The idea is that the frequency of this oscillation must be defined mainly by the vehicle mass with additional components such as spring forces and geometric distances from the center of gravity of the forces generated by the springs.

[0015] The following describes the calculation of the natural frequencies as it can be found on the website “Half Car Modeling (skill-lync.com)”. A corresponding vehicle model is available in the Fig. 2 is shown schematically.

[0016] Then a vertical oscillation (when the springs at the front and rear compress and rebound synchronously) and a pitch oscillation (positive when it compresses at the front and / or rebounds at the rear) are calculated according to the formulas Bounce: ωn1=(D1+D3)−(D1+D3)2−4⋅D42 Pitch ωn2=(D1+D3)−(D1+D3)2−4⋅D42 with the constants Constants: D1=kf+krms D2=b⋅kr−a⋅kfms D3=a2⋅kf+b2⋅krIy D4=(a+b)2⋅kf⋅krms⋅Iy and W s = 9.79 Sprung weight (kN) a = 106.7 Center of gravity behind the front axle (cm) l = 228.6 wheelbase (cm) b = 1- a Center of gravity in front of the rear axle (cm) k f = 24.52 * 10 3 Front spring stiffness (kN / m to N / m) k r = 26.27 * 10 3 Rear spring stiffness (kN / m to N / m) r y = 102.6 radius of gyration (cm) m s = W s * 103 / 9.81 Sprung mass (kN to kg) The moment of inertia (kg.cm^2) is calculated. The frictional forces of the damper are ignored.

[0017] Inserted into the above formula, this gives the frequency ω=kf+krms+a2⋅kf+b2⋅krms⋅ry2+(kf+krms+a2⋅kf+b2⋅krms⋅ry2)2−4⋅((a+b)2⋅kf⋅krms2⋅ry2) ω=(kf+kr)⋅ry2+a2⋅kf+b2⋅krms⋅ry2+((kf+kr)⋅ry2+a2⋅kf+b2⋅krms⋅ry2)2−4⋅((a+b)2⋅kf⋅krms2⋅ry2)

[0018] With z=(kf+kr)⋅ry2+a2⋅kf+b2⋅kr follows ω=zms⋅ry2+(zms⋅ry2)2−4⋅((a+b)2⋅kf⋅krms⋅ry2)2 (2⋅ω2−zms⋅ry2)2=(zms⋅ry2)2−4⋅((a+b)2⋅kf⋅krms⋅ry2) 4⋅ω4−4⋅ω2⋅zms⋅r2+(zms⋅ry2)2=(zms⋅ry2)2−4⋅((a+b)2⋅kf⋅krms⋅ry2) 4⋅ω4−4⋅ω2⋅zms⋅ry2=−4⋅(a+b)2⋅kf⋅krms2⋅ry2 4⋅ω4⋅ry2⋅ms2−4⋅ω2⋅z⋅ms+4⋅(a+b)2⋅kf⋅kr=0 (ω4⋅ry2)⋅ms2−(ω2⋅z)⋅ms+((a+b)2⋅kf⋅kr)=0

[0019] The solutions of a quadratic equation are: ax2+bx+c=0 xi=−b±b2−4ac2a

[0020] This results in the total mass: ms1,2=(ω2⋅z±ω4⋅z2−4ω4ry2⋅(a+b)2⋅kf⋅kr)2⋅ω4⋅ry2 with z=(kf+kr)⋅ry2+a2⋅kf+b2⋅kr

[0021] Both solutions must be verified, as one is only an apparent solution. It's possible that one solution is always the correct solution, since the only way to switch between the solutions is for the square root to be 0, as continuity is physically enforced.

[0022] With this calculated sprung mass m sThe vehicle's payload can be defined, and the load-related inclination of the vehicle can be determined using a so-called beam model. The calculated payload is distributed within the vehicle using seat occupancy detection, assuming, for example, a standard person with a mass of 75 kg. The remaining payload after distribution among the seats is located in the trunk.

[0023] The Fig. 3 shows a simple beam model in which the forces acting on a vehicle are entered, the Fig. Figures 4 to 6 show these forces in simple vehicle models. L1 = spring force front L2 = spring force rear Dr = weight of the driver Cdr = passenger weight P1, P2, P3 = weight forces of the passengers on the rear seat 1 to 3 Equalities: 1) Torque of the axle load is equal to the load L1⋅h1+L2⋅h2=(Dr+Cdr)⋅h3+(P1+P2+P3)⋅h4+Pk⋅h5 2) Suspension travel times lever travel is the same at the front and rear when nodding ΔZ1h1=ΔZ2h2→h1=ΔZ1⋅h2ΔZ2 3) Forces of the springs correspond to the load L1+L2=Dr+Cdr+P1+P2+P3+Pk 4) only nodding movement Z1−ΔZ1=Z2−ΔZ2 ΔZ2=Z2−Z1+ΔZ1 5) Spring equations L1=kf⋅(Z1+ΔZ1) L2=kr⋅(Z2+ΔZ2) 6) total torque around the axis of rotation G⋅gG=(Dr+Cdr)⋅h3+(P1+P2+P3)⋅h4+Pk⋅h5 G=Dr+Cdr+P1+P2+P3+Pk hG=(Dr+Cdr)⋅h3+(P1+P2+P3)⋅h4+Pk⋅h5G

[0024] 2) → 1) L1⋅ΔZ1⋅h2ΔZ2+L2⋅h2=(Dr+Cdr)⋅h3+(P1+P2+P3)⋅h4+Pk⋅h5

[0025] Inserting 6) L1⋅ΔZ1⋅h2ΔZ2+L2⋅h2=G⋅hG

[0026] Inserting 5), times ΔZ2 kf⋅(Z1+ΔZ1)⋅ΔZ1⋅h2+kr⋅(Z2+ΔZ2)⋅ΔZ2⋅h2=ΔZ2⋅G⋅hG

[0027] Inserting 4) kf⋅(Z1+ΔZ1)⋅ΔZ1⋅h2+k2⋅(2 Z2−Z1+ΔZ1)⋅(Z2−Z1+ΔZ1)⋅h2 =(Z2−Z1+ΔZ1)⋅G⋅hG

[0028] Solving for ΔZ1 leads to a quadratic equation: kfh2⋅Z1ΔZ1+kfh2⋅ΔZ12+krh2 ⋅(2 Z22−2Z1Z2+2Z2ΔZ1−Z1Z2+Z12−Z1ΔZ1+Z2ΔZ1−Z1ΔZ1 +ΔZ12)=hGZ2G−hGZ1G+hGΔZ1G kfh2⋅Z1ΔZ1+kfh2⋅ΔZ12+krh2 ⋅(2 Z22−3Z1Z2+3Z2ΔZ1−2Z1ΔZ1+Z12+ΔZ12) =hGZ2G−hGZ1G+hGΔZ1G kfh2⋅Z1ΔZ1+kfh2⋅ΔZ12+2krh2⋅Z22−3krh2⋅Z1Z2+3krh2⋅Z2ΔZ1−2krh2 ⋅Z2ΔZ1+krh2⋅Z12+krh2⋅ΔZ12−hGZ2G+hGZ1G−hGΔZ1G=0 ΔZ12⋅(kfh2+krh2)+ΔZ1⋅(kfh2Z1+3krh2⋅Z2−2krh2⋅Z1−hGG)+(2krh2 ⋅Z22−3krh2⋅Z1Z2+krh2⋅Z12−hGZ2G+hGZ1G)=0

[0029] This approach seems a bit too complicated for an algorithm that is supposed to run without further supervision.

[0030] Therefore, it is suggested that, as a simplification, instead of equation 4), a percentage distribution of the loads between the front and rear axles be used. For further simplification, the pivot point is assumed to be in front of the front axle or behind the rear axle, so that the term h G - h1 certainly cannot be 0, which would cause the algorithm to fail. G=L1+L2=G⋅ratio1+G⋅ratio2 ratio1+ratio2=1

[0031] This results in 5) G⋅ratio1=L1=kf⋅(Z1+ΔZ1) G⋅ratio2=L2=kr⋅(Z2+ΔZ2)

[0032] To further simplify the calculation, the zero point of the spring deflection is set to the position in which the vehicle would be deflected on a horizontal plane without any further change in load. Z1=Z2=0

[0033] It follows ΔZ1=G⋅ratio1 / kf ΔZ2=G⋅ratio2 / kr where ratio1=hG−h1l ratio2=l−(hG−h1)l

[0034] With h G from equation 6) hG=(Dr+Cdr)⋅h3+(P1+P2+P3)⋅h4+Pk⋅h5G the approach is good enough to allow the headlight range to be adjusted precisely enough.

[0035] These simplifications lead to deviations from the exact results. As long as the spring rates of the axles are approximately the same, the position of the pivot point is irrelevant for determining the angular change.

[0036] The angle of inclination β is determined using trigonometry.

[0037] The wheelbase proposed by the invention eliminates the need for chassis sensors and their wiring, offering significant savings potential. Chassis sensors are typically positioned in the vehicle where they are exposed to environmental influences, including stone chips, whereas the proposed IMU sensors can be mounted in the passenger compartment. IMU sensors are standard in every vehicle (ESP, airbag).

Claims

[1] Method for determining the tilting of a vehicle due to a load by determining the mass of the vehicle by means of determining the natural frequency of the vehicle during a braking process, comprising the steps: Detection of a braking process, Detecting the resulting standstill of the vehicle, Determining the natural frequency of the vehicle using a sensor by measuring the vibration of the vehicle directly after it has come to a standstill and evaluating the sensor signal, Determining the total mass from the natural frequency using a vehicle model, Determining the number of people sitting in the vehicle and their distribution using sensors, Determine the residual load by subtracting the number of people multiplied by an assumed average weight from the previously determined total mass, assuming that the residual load is in the trunk and Determining the tipping from the distribution of people and the residual load using a bar model. [2] Method according to claim 1, wherein the sensor for determining the natural frequency is an IMU sensor. [3] Method according to claim 1 or 2, wherein the sensor system for determining the persons sitting in the vehicle and their distribution is a seat occupancy detection sensor system. [4] Method according to one of the preceding claims, in which, for the sake of simplicity, the beam model is divided into percentages of the loads between the front and rear axles and, for further simplification, the pivot point is assumed to be in front of the front axle or behind the rear axle. [5] Method according to claim 4, in which, as a further simplification of the calculation, the zero point of the spring deflection is set to the position in which the vehicle would be deflected on a horizontal plane without any further change in load.

Images