Device and method for determining a vehicle inclination

Abstract

Tilt angle detection device in a vehicle, which has a chassis (3) with several vehicle wheels (10, 11, 12, 13) and several vehicle axles (8, 9) spaced apart from each other in a longitudinal direction (x) of the vehicle, each of which at least one of the vehicle wheels is assigned, and a vehicle body (2) supported by the chassis (3), wherein the tilt angle detection device (27) has several sensor units (30, 37), of which a first sensor unit (30) is provided on a first chassis component (14) assigned to a first of the vehicle axles (8) or on the vehicle body (2) and a second sensor unit (37) is provided on a second chassis component assigned to a second of the vehicle axles (9), and an evaluation unit (45) connected to the sensor units (30, 37),wherein the first sensor unit (37) can detect the orientation of the first chassis component (14) or the vehicle body (2) relative to the earth (E) or the height at the first vehicle axle (8) and first sensor signals characterizing this orientation or height can be provided, and wherein the second sensor unit (3) can detect the orientation of the second chassis component relative to the earth (E) and second sensor signals characterizing this orientation can be provided, wherein the first sensor signals can be provided in the form of at least one first time series and the second sensor signals in the form of at least one second time series, the evaluation unit (45) comprises a neural network (59) with an output layer (62) and an input layer (60), to which values ​​of the time series or values ​​derived therefrom can be supplied as input signals,and at least one output signal (S_&thgr;_N) characterizing an inclination (&thgr;_N) of the vehicle structure (2) can be estimated by the neural network (59) and made available at the output layer (62).

Description

The invention relates to a tilt angle detection device in a vehicle, which has a chassis with several vehicle wheels and several vehicle axles spaced apart from each other in a longitudinal direction, each of which is assigned at least one of the vehicle wheels, and a vehicle body supported by the chassis, wherein the tilt angle detection device has several sensor units, of which a first sensor unit is provided on a first chassis component assigned to a first of the vehicle axles or on the vehicle body and a second sensor unit is provided on a second chassis component assigned to a second of the vehicle axles, and an evaluation unit connected to the sensor units.wherein the first sensor unit can detect the orientation of the first chassis component or the vehicle body relative to the ground or the height at the first vehicle axle, and first sensor signals characterizing this orientation or height are available, and wherein the second sensor unit can detect the orientation of the second chassis component relative to the ground, and second sensor signals characterizing this orientation are available. Furthermore, the invention relates to a method for detecting the tilt angle in a vehicle. Up to now, the sensor signals have been evaluated using an analytical model. A disadvantage of this is that an analytical model can be inaccurate due to simplifications of real-world conditions. The use of neural networks for various purposes is generally known in chassis engineering. For example, reference can be made to each of the documents WO 2025 / 109 834 A1, DE 11 2021 001 704 T5, DE 11 2020 005 420 T5, and WO 2024 / 070 111 A1. Based on the relationship described at the beginning, the invention is primarily aimed at reducing inaccuracies in the evaluation of sensor signals. This problem is solved according to the invention by a tilt angle detection device according to claim 1 and by a method according to claim 11. Preferred embodiments of the invention are given in the dependent claims and in the following description. A tilt angle detection device in a vehicle, which has a chassis with several vehicle wheels and several vehicle axles spaced apart from each other in a longitudinal direction, each of which is assigned at least one of the vehicle wheels, and a vehicle body supported by the chassis, wherein the tilt angle detection device has several sensor units, of which a first sensor unit is provided on a first chassis component assigned to a first of the vehicle axles or on the vehicle body and a second sensor unit is provided on a second chassis component assigned to a second of the vehicle axles, and an evaluation unit connected to the sensor units,wherein the first sensor unit can detect the orientation of the first chassis component or the vehicle body relative to the earth or the height at the first vehicle axle and first sensor signals characterizing this orientation or height can be provided, and wherein the second sensor unit can detect the orientation of the second chassis component relative to the earth and second sensor signals characterizing this orientation can be provided, is further developed according to the invention in particular in that the first sensor signals can be provided in the form of at least one first time series and the second sensor signals in the form of at least one second time series, the evaluation unit comprises a neural network with an output layer and an input layer, to which values ​​of the time series or values ​​derived therefrom can be supplied as input signals,and that at least one output signal characterizing the inclination of the vehicle structure can be estimated by the neural network and made available at the output layer. A time series analysis performed by a neural network eliminates the need for an analytical model describing the vehicle's body tilting behavior. Consequently, the inaccuracies associated with such a model are also avoided. The invention further relates in particular to a method for detecting the tilt angle in a vehicle, which has a chassis with several vehicle wheels and several vehicle axles spaced apart from each other in a longitudinal direction, each of which is assigned at least one of the vehicle wheels, and a vehicle body supported by the chassis, wherein a tilt angle detection device has several sensor units, of which a first sensor unit is provided on a first chassis component assigned to a first of the vehicle axles or on the vehicle body and a second sensor unit is provided on a second chassis component assigned to a second of the vehicle axles, and an evaluation unit connected to the sensor units.wherein the first sensor unit detects the orientation of the first chassis component or the vehicle body relative to the earth or the height at the first vehicle axle and provides first sensor signals characterizing this orientation or height, and wherein the second sensor unit detects the orientation of the second chassis component relative to the earth and provides second sensor signals characterizing this orientation. In particular, the first sensor signals are provided in the form of at least one first time series and the second sensor signals in the form of at least one second time series, wherein the evaluation unit comprises a neural network with an output layer and an input layer, to which values ​​of the time series or values ​​derived therefrom are supplied as input signals.and wherein the neural network estimates at least one output signal characterizing the inclination of the vehicle structure and provides it at the output layer. The method according to the invention can be further developed according to all embodiments described in connection with the tilt angle detection device according to the invention. Furthermore, the tilt angle detection device according to the invention can be further developed according to all embodiments described in connection with the method according to the invention. The expression "at least one" also includes, in particular, the meaning of "one" or "exactly one". Preferably, the vehicle wheels rest on or roll on a surface. The tilt angle detection device is specifically provided in the vehicle. The neural network in question is, in particular, an artificial neural network. Preferably, the neural network has one or more intermediate layers. For example, the neural network is a convolutional neural network (CNN). Advantageously, the neural network is a recurrent neural network. Preferably, the neural network is a trained neural network. Preferably, the first sensor signals are provided or made available in the form of at least one first time series by means of the first sensor unit, a first scanning unit, or the evaluation unit. The first scanning unit can, for example, be integrated into the first sensor unit or the evaluation unit, or it can be a separate unit. Advantageously, the second sensor signals are provided or made available in the form of at least one second time series by means of the second sensor unit, a second scanning unit, or the evaluation unit. The second scanning unit can, for example, be integrated into the second sensor unit or the evaluation unit, or it can be a separate unit. The second scanning unit is, for example, formed by the first scanning unit, which in this case is preferably referred to simply as the scanning unit. According to an advantageous embodiment, at least two of the vehicle wheels are assigned to each vehicle axle, and these wheels are spaced apart from each other, particularly in a transverse direction. This is generally the case with passenger cars and trucks. The expression "at least two" also includes the meaning of "two" or "exactly two". The vehicle's transverse direction runs, in particular, perpendicular to the vehicle's longitudinal direction. A vehicle's vertical direction preferably runs perpendicular to both the vehicle's longitudinal direction and the vehicle's transverse direction. Advantageously, the forward direction of travel of the vehicle, or one of its most common directions, runs in the vehicle's longitudinal direction. The height level detected or detectable by the first sensor unit on the first vehicle axle is, for example, the height level of one of the vehicle wheel(s) assigned to the first vehicle axle. The number of vehicle axles is preferably at least two. Preferably, the vehicle axles comprise a front axle and a rear axle. For example, the first vehicle axle is the front axle and the second vehicle axle is the rear axle. Alternatively, the second vehicle axle is the front axle and the first vehicle axle is the rear axle. Preferably, the chassis comprises several vehicle springs. Preferably, the vehicle body is connected to unsprung components of the chassis via the vehicle springs. Advantageously, the unsprung components of the chassis are each assigned to one of the vehicle axles. In particular, the unsprung components of the chassis comprise the vehicle wheels. Preferably, the chassis comprises several wheel carriers. In particular, the chassis comprises one wheel carrier for each vehicle wheel. Advantageously, the wheel carriers are each assigned to one of the vehicle wheels. Preferably, the vehicle wheels are each mounted on one of the wheel carriers, in particular rotatably about a wheel axis. Advantageously, the unsprung components include the wheel carriers. Preferably, the chassis comprises several suspension links. Preferably, the vehicle wheels are articulated to the vehicle body by the suspension links. In particular, the wheel carriers are articulated to the vehicle body by the suspension links. Advantageously, the vehicle wheels are articulated to the vehicle body by means of, or via, the wheel carriers and / or the suspension links. In particular, the suspension links are each assigned to one of the vehicle wheels. Preferably, the first chassis component is formed by one of the unsprung components of the chassis assigned to the first vehicle axle or by one of the suspension links of the vehicle wheel or wheels assigned to the first vehicle axle. Advantageously, the second chassis component is formed by one of the unsprung components of the chassis assigned to the second vehicle axle or by one of the suspension links of the vehicle wheel or wheels assigned to the second vehicle axle. Preferably, the first chassis component is formed by a wheel carrier or a suspension link of one or more of the vehicle wheels assigned to the first vehicle axle. Advantageously, the second chassis component is formed by a wheel carrier or a suspension link of one or more of the vehicle wheels assigned to the second vehicle axle. According to an advantageous embodiment, a time window, preferably sliding, with a predetermined number of consecutive values ​​can be placed over the time series, preferably by means of a time window unit or by means of the evaluation unit, in particular before the input layer, wherein the values ​​of the time series lying within the time window or values ​​derived therefrom can be supplied to the input layer as, in particular, input signals. Preferably, the predetermined number of values ​​are past values. The current value(s), in particular the respective ones, can be included or excluded, for example. Preferably, the time window unit is located upstream of the input layer. Advantageously, the evaluation unit includes the time window unit or is connected to it. According to an advantageous further development, the values ​​of the time series, in particular before the input layer, can be or are subjected to filtering, such as high-pass filtering and / or low-pass filtering, preferably by means of a filter unit or by means of the evaluation unit. Preferably, the filter unit is located upstream of the input layer. Advantageously, the evaluation unit includes the filter unit or is connected to it. According to an advantageous embodiment, the values ​​of the time series, especially before the input layer, can be normalized or normalized, preferably by means of a normalizing unit or by means of the evaluation unit. Preferably, the normalization unit is located upstream of the input layer. Advantageously, the evaluation unit includes the normalization unit or is connected to it. According to an advantageous further development, the values ​​of the time series, especially before the input layer, can be or are subjected to feature extraction, preferably by means of a feature extraction unit or by means of the evaluation unit. Preferably, the feature extraction unit is located upstream of the input layer. Advantageously, the evaluation unit includes the feature extraction unit or is connected to it. Feature extraction includes, for example, the formation of root mean squares from several values ​​of each time series. According to an advantageous embodiment, the mass of the vehicle and / or the mass of a load on the vehicle, for example when the vehicle is stationary and / or while the vehicle is in motion, can be determined, preferably by means of a mass detection unit or the evaluation unit. Preferably, at least one mass signal or signals characterizing this mass or these masses can be additionally supplied to the input layer, preferably in the form of a time series, as an input signal or these input signals, preferably by means of the mass detection unit or the evaluation unit. A device and a method for determining a vehicle mass are known, for example, from DE 10 2013 211 243 A1 and DE 10 2005 014 569 A1. Preferably, the mass detection unit is located upstream of the input layer. Advantageously, the evaluation unit includes the mass detection unit or is connected to it. The vehicle preferably comprises at least one drive motor, by means of which the vehicle can be driven or is driven. Preferably, the at least one drive motor can deliver a drive torque. According to an advantageous embodiment, the drive torque of the at least one drive motor of the vehicle can be determined, preferably by means of a drive torque detection unit or an evaluation unit. Preferably, at least one drive torque signal or signals characterizing this drive torque can be additionally supplied to the input layer, preferably in the form of a time series, as an input signal or signals, preferably by means of the drive torque detection unit or the evaluation unit. The drive torque is also referred to, for example, as motor torque. Preferably, the drive torque sensing unit is located upstream of the input layer. Advantageously, the evaluation unit includes the drive torque sensing unit or is connected to it. Preferably, at least one of the vehicle wheels is steerable. Preferably, at least two of the vehicle wheels are steerable. Advantageously, every vehicle wheel of the front axle is steerable. According to an advantageous embodiment, a steering angle of at least one steerable vehicle wheel or at least one of the steerable vehicle wheels can be determined, preferably by means of a steering angle detection unit or by means of the evaluation unit. Preferably, at least one steering angle signal or steering angle signals characterizing this steering angle can be additionally supplied to the input layer, preferably in the form of a time series, as an input signal or input signals, preferably by means of the steering angle detection unit or by means of the evaluation unit. Preferably, the steering angle detection unit is located upstream of the input layer. Advantageously, the evaluation unit includes the steering angle detection unit or is connected to it. The vehicle preferably includes a vehicle braking system by means of which, in particular, a braking torque can be transmitted or is transmitted to at least one of the vehicle wheels. According to an advantageous embodiment, a braking torque of the vehicle acting on at least one of the vehicle wheels can be determined, preferably by means of a brake torque detection unit or the evaluation unit. Preferably, at least one brake torque signal or signals characterizing this braking torque can be additionally supplied to the input layer, preferably in the form of a time series, as an input signal or signals, preferably by means of the brake torque detection unit or the evaluation unit. Preferably, the brake torque detection unit is located upstream of the input layer. Advantageously, the evaluation unit includes the brake torque detection unit or is connected to it. The vehicle braking system is advantageously a hydraulic or pneumatic system in which the braking torque depends on the brake pressure. For this reason, the brake pressure can also be measured instead of the braking torque. In this case, the brake torque sensing unit can also be referred to as a brake pressure sensing unit. Furthermore, the brake torque signal(s) can also be referred to as the brake pressure signal(s). According to an advantageous embodiment, the at least one output signal describes or characterizes a pitch angle of the vehicle and / or at least one or more height levels of the vehicle, preferably at different vehicle axles. From at least two height levels occurring at different vehicle axles, the pitch angle of the vehicle can be determined or estimated. Preferably, the first sensor unit comprises an acceleration sensor, in particular one- or multi-dimensional, preferably three-dimensional. Advantageously, the first sensor unit also comprises a gyroscope, in particular one- or multi-dimensional, preferably three-dimensional. The gyroscope, in particular, increases the accuracy in detecting the orientation of the first chassis component or the vehicle body relative to the ground. The first sensor unit preferably comprises an inertial measurement unit (IMU) or is formed, for example, by one. Preferably, this inertial measurement unit is configured to detect several, for example, six kinematic degrees of freedom. The inertial measurement unit preferably comprises a multidimensional, preferably three-dimensional, accelerometer and a multidimensional, preferably three-dimensional, angular rate sensor. Alternatively, the first sensor unit includes or forms, for example, a height level sensor. Preferably, the second sensor unit comprises an accelerometer, in particular one- or multi-dimensional, preferably three-dimensional. Advantageously, the second sensor unit also comprises an angular rate sensor, in particular one- or multi-dimensional, preferably three-dimensional. In particular, the rotation rate sensor increases the accuracy in detecting the orientation of the second landing gear component relative to the earth. The second sensor unit preferably comprises an inertial measurement unit (IMU) or is formed, for example, by one. This inertial measurement unit is preferably configured to detect several, for example, six kinematic degrees of freedom. The inertial measurement unit preferably comprises a multidimensional, preferably three-dimensional, accelerometer and a multidimensional, preferably three-dimensional, angular rate sensor. According to an advantageous embodiment, a headlight assembly is provided with at least one adjustable headlight by means of which light can be emitted or is emitted, the beam angle of which can be controlled or is controlled depending on the at least one output signal. Preferably, the headlight assembly includes a headlight range control unit to which the output signal can be supplied or is supplied. The at least one output signal is thus, in particular, usable or is used for headlight range control. According to an advantageous embodiment, a third sensor unit is provided on a third chassis component assigned to the second vehicle axle, wherein the orientation of the third chassis component relative to the earth can be detected or is detected by means of the third sensor unit and third sensor signals characterizing this orientation can be provided or are provided. Preferably, the third chassis component is formed by another unsprung component of the chassis assigned to the second vehicle axle or by one of the suspension links of another vehicle wheel assigned to the second vehicle axle. More preferably, the third chassis component is formed by a wheel carrier or a suspension link of the vehicle wheel or another vehicle wheel assigned to the second vehicle axle. Preferably, the third sensor signals are or are provided in the form of at least one third time series. Preferably, the third sensor unit comprises an accelerometer, in particular one- or multi-dimensional, preferably three-dimensional. Advantageously, the third sensor unit also comprises a gyroscope, in particular one- or multi-dimensional, preferably three-dimensional. The gyroscope, in particular, increases the accuracy in detecting the orientation of the third landing gear component relative to the ground. The third sensor unit preferably comprises an inertial measurement unit (IMU) or is formed, for example, by one. This inertial measurement unit is preferably configured to detect several, for example, six kinematic degrees of freedom. The inertial measurement unit preferably comprises a multidimensional, preferably three-dimensional, accelerometer and a multidimensional, preferably three-dimensional, angular rate sensor. Preferably, the first sensor unit is provided on the vehicle body, wherein preferably the orientation of the vehicle body relative to the earth can be detected or is detected by means of the first sensor unit and sensor signals characterizing this orientation can be provided or are provided as the first sensor signals. According to an advantageous further development, a fourth or further sensor unit is provided, by means of which a height level on the first vehicle axle can be detected or is detected, and fourth or further sensor signals characterizing this height level can be provided or are provided. Preferably, the fourth or further sensor signals are or are provided in the form of at least a fourth or further time series. The height level detected or detectable by means of the fourth or further sensor unit on the first vehicle axle is in particular the height level of one of the vehicle wheel(s) assigned to the first vehicle axle. According to an advantageous embodiment, a fifth or additional sensor unit is provided, by means of which a, preferably different, height level on the first vehicle axle can be detected or is detected, and fifth or additional sensor signals characterizing this, preferably different, height level can be provided or are provided. Preferably, the fifth or additional sensor signals are or are provided in the form of at least one fifth or additional time series. The height level detected or detectable by means of the fifth or additional sensor unit on the first vehicle axle, preferably other, is in particular the height level of one, preferably other, of the vehicle wheels assigned to the first vehicle axle. Preferably, the neural network can be trained or trained, preferably before its use or in advance, in particular with reference data, at or with at least one loading state or different loading states of the vehicle. The invention is described below with reference to a preferred embodiment and the drawing. The drawing shows: Fig. 1 a schematic top view of a vehicle, Fig. 2 a schematic view of a wheel suspension of the vehicle, Fig. 3 a schematic view of a tilt angle detection device for determining a pitch angle according to one embodiment, Fig. 4 a schematic side view of the vehicle on an inclined surface, Fig. 5 a view of an evaluation unit of the tilt angle detection device with a schematically indicated neural network, and Fig. 6 several time series of measurement data with a sliding time window for selecting input data for the neural network. Figure 1 shows a schematic top view of a vehicle 1, which has a vehicle body 2 and a chassis 3 with several wheel suspensions 4, 5, 6 and 7, of which wheel suspensions 4 and 5 are assigned to a front axle 8 and wheel suspensions 6 and 7 to a rear axle 9. Each wheel suspension comprises a vehicle wheel, with wheel suspension 4 comprising vehicle wheel 10, wheel suspension 5 comprising vehicle wheel 11, wheel suspension 6 comprising vehicle wheel 12 and wheel suspension 7 comprising vehicle wheel 13. A coordinate system with a vehicle longitudinal axis x, a vehicle transverse axis y and a vehicle vertical axis z is also shown. Figure 2 shows a schematic view of the wheel suspension 4, which has a wheel carrier 14 connected by a joint 15, preferably designed as a ball joint, to a suspension link 16 designed as a transverse arm, the end of which is articulated to the vehicle body 2 by at least one or two joints 17, preferably designed as rubber bushings. Furthermore, the wheel carrier 14 is connected, preferably rigidly, to a strut 18, the end of which is connected to the vehicle body 2 by a strut mount 19. The strut 18 comprises a vehicle spring 20 and a damper 21, which is preferably surrounded by the vehicle spring 20, which is preferably designed as a coil spring. A wheel bearing 22 is attached to the wheel carrier 14, by means of which the vehicle wheel 10 is rotatably mounted on the wheel carrier 14 about a wheel axis 23.Furthermore, a tie rod 24 is connected to the wheel carrier 14 by means of a joint 25, preferably designed as a ball joint. The vehicle wheel 10 is in contact with a surface 26, which is, for example, a road or roadway. It should be noted that the vehicle coordinate system shown in Fig. 1 is shown shifted in Fig. 2. The vehicle 1 comprises a tilt angle detection device 27, shown schematically in Fig. 3, which includes a height sensor 28 provided at the joint 15. By measuring an angle α enclosed between the wheel carrier 14 and the suspension link 16, the height h of the vehicle wheel 10 relative to a reference position h_ref can be detected, and a height signal HLS_FR characterizing this height h can be provided in the form of a time series. The reference position h_ref is, in particular, fixed relative to the vehicle body 2. The height sensor 28 is thus connected between the suspension link 16 and the wheel carrier 14. Alternatively, the height sensor 28 can also be provided at another location in the wheel suspension 4 and, for example, connected between the suspension link 16 and the vehicle body 2. The wheel suspension 5 is arranged in a mirror image of the wheel suspension 4 with respect to a longitudinal vertical plane Exz of the vehicle 1 and has a height level sensor 29, preferably corresponding to the height level sensor 27, by means of which a height level of the vehicle wheel 11 can be detected and a height level signal HLS_FL characterizing this height level can be provided in the form of a time series (see Fig. 3 ). The vehicle wheels 10 and 11 of the wheel suspensions 4 and 5 of the front axle 8 are steerable. The vehicle wheels 12 and 13 of the wheel suspensions 6 and 7 of the rear axle 9 are, for example, not steerable or only steerable to a limited extent. Apart from these differences, the wheel suspensions 6 and 7 are preferably, in particular at least approximately, constructed in the same or a similar manner to the wheel suspensions 4 and 5. The vehicle body 2 is equipped with a measuring unit 30 in the form of an inertial measuring unit, which includes three translational acceleration sensors 31, 32, and 33 and three yaw rate sensors 34, 35, and 36 (see Fig. 3). Translational acceleration sensor 31 provides an acceleration signal 6DoF-Body_x, characterizing the longitudinal acceleration of the vehicle body 2, in the form of a time series. Translational acceleration sensor 32 provides an acceleration signal 6DoF-Body_y, characterizing the lateral acceleration of the vehicle body 2, in the form of a time series. Finally, translational acceleration sensor 33 provides an acceleration signal 6DoF-Body_z, characterizing the vertical acceleration of the vehicle body 2, in the form of a time series.The yaw rate sensor 34 provides a yaw rate signal, Yaw-rate-Body_x, in the form of a time series, characterizing the rotational speed or rate of the vehicle body 2 about the vehicle's longitudinal axis x. Furthermore, the yaw rate sensor 36 provides a yaw rate signal, Yaw-rate-Body_z, in the form of a time series, characterizing the rotational speed or rate of the vehicle body 2 about the vehicle's vertical axis z. Preferably, the yaw rate sensor 35 provides a yaw rate signal, Yaw-rate-Body_y, in the form of a time series, characterizing the rotational speed or rate of the vehicle body 2 about the vehicle's transverse axis y. A sensor unit 37 in the form of a multidimensional accelerometer is provided on a wheel carrier 14 of the vehicle wheel 12 of the wheel suspension 6. The accelerometer comprises three translational accelerometers 38, 39, and 40 (see Fig. 3). Translational accelerometer 38 provides an acceleration signal Knuckle_RR_x, characterizing the longitudinal acceleration of the vehicle wheel 12, in the form of a time series. Translational accelerometer 39 provides an acceleration signal Knuckle_RR_y, characterizing the lateral acceleration of the vehicle wheel 12, in the form of a time series. Finally, translational accelerometer 40 provides an acceleration signal Knuckle_RR_z, characterizing the vertical acceleration of the vehicle wheel 12, in the form of a time series. Optionally, the sensor unit 37 can be configured as an inertial measuring unit and may additionally include several, preferably three, yaw rate sensors.Alternatively, the sensor unit 37 can be provided not on the wheel carrier 14 of the vehicle wheel 12, but on a suspension link 16 of the wheel suspension 6. Furthermore, a sensor unit 41 in the form of a multidimensional accelerometer is provided on a wheel carrier 14 of the vehicle wheel 13 of the wheel suspension 7. This sensor unit comprises three translational accelerometers 42, 43, and 44 (see Fig. 3). Translational accelerometer 42 provides an acceleration signal Knuckle_RL_x, characterizing the longitudinal acceleration of the vehicle wheel 13, in the form of a time series. Translational accelerometer 43 provides an acceleration signal Knuckle_RL_y, characterizing the lateral acceleration of the vehicle wheel 13, in the form of a time series. Finally, translational accelerometer 44 provides an acceleration signal Knuckle_RL_z, characterizing the vertical acceleration of the vehicle wheel 13, in the form of a time series. Optionally, the sensor unit 41 can be configured as an inertial measuring unit and may additionally include several, preferably three, yaw rate sensors.Alternatively, the sensor unit 41 can be provided not on the wheel carrier 14 of the vehicle wheel 13, but on a suspension link 16 of the wheel suspension 7. As can be seen from Fig. 3, which shows a schematic view of the tilt angle detection device 27, the height level sensors 28 and 29 as well as the sensors of the sensor units 30, 37 and 41 are connected to an evaluation unit 45, by means of which a pitch angle θ_N of the vehicle 1 can be determined and a pitch angle signal S_θ_N characterizing this pitch angle can be provided. The vehicle 1 comprises a drive motor 46 with a motor shaft 47, which is coupled to the vehicle wheels 10 and 11, for example via at least one vehicle transmission. Additionally or alternatively, the motor shaft 47 is coupled to the vehicle wheels 12 and 13, for example via at least one vehicle transmission. The drive motor 46 is connected to a control unit 48, which in turn is connected to a bus system 49. The control unit 48 is, or preferably comprises, an engine control unit, or is, for example, connected to an engine control unit. The bus system 49 comprises, for example, a CAN bus. Preferably, the control unit 48 provides a drive torque signal Smotor, characterizing the drive torque output by the drive motor 46, via the bus system 49. Furthermore, the vehicle 1 comprises a vehicle braking system with a brake pedal 50 for actuating the same, wherein the vehicle braking system has a brake pressure sensor 51 connected to the control unit 48, by means of which a hydraulic or pneumatic brake pressure in the braking system can be detected and a brake pressure signal Sb characterizing this pressure can be provided. Preferably, the control unit 48 provides the brake pressure signal Sb via the bus system 49. The evaluation unit 45 is preferably connected to the bus system 49, by means of which the drive torque signal Smotor and the brake pressure signal Sb can be supplied to the evaluation unit 45. Preferably, a mass detection unit 52 is provided, by means of which the mass of the vehicle 1 or of a load 53 of the vehicle can be determined when stationary and / or while in motion, and at least one mass signal Smass characterizing this mass can be provided and advantageously supplied to the evaluation unit 45. Alternatively, the mass detection unit is formed, for example, by the evaluation unit. The vehicle 1 further comprises a front headlight assembly 54 with two adjustable front headlights 55 and 56, by means of which light 57 can be emitted, the beam angle φ of which can be controlled as a function of the pitch angle θ_N (see Fig. 4). For this purpose, the front headlight assembly 54 has a headlight range control unit 58, to which the pitch angle signal S_θ_N can be supplied. In a state resting on an inclined surface 26, the vehicle 1 is shown in Fig. 4, where θ_S denotes an angle characterizing the orientation of the surface 26 relative to the ground E, and θ_B denotes an angle characterizing the orientation of the vehicle body 2 relative to the ground E. The load 53 of the vehicle 1, which also contributes to the pitch angle θ_N, is also shown schematically. The pitch angle θ_N is the difference between θ_B and θ_S. The angle θ_S can be detected, for example, by sensor units 37 and 41. Furthermore, the angle θ_B can be detected, for example, by sensor unit 30. The evaluation unit 45 comprises a neural network 59, preferably a trained one, with an input layer 60, several intermediate layers 61, and an output layer 62, wherein values ​​of the time series or values ​​derived therefrom can be supplied to the input layer 60 as input signals. The pitch angle θ_N can be estimated by the neural network 59, and the pitch angle signal S_θ_N characterizing this angle can be provided as an output signal at the output layer 62. The sensor unit 30 forms, in particular, a first sensor unit of the tilt angle detection device 27, the sensor unit 37 forms, in particular, a second sensor unit of the tilt angle detection device 27, and the sensor unit 41 forms, in particular, a third sensor unit of the tilt angle detection device 27. A fourth or further sensor unit of the tilt angle detection device 27 preferably comprises the height level sensor 28 or is formed by it. A fifth or additional sensor unit of the tilt angle detection device 27 preferably comprises the height level sensor 29 or is formed by it. The vehicle axle 8 forms in particular a first vehicle axle of the vehicle 1 and the vehicle axle 9 forms in particular a second vehicle axle of the vehicle 1. The evaluation of the signals recorded by the sensors is explained in more detail below. Figure 6 shows time series of the signals recorded by the sensors. These signals include the height signal HLS_FR of vehicle wheel 10, the height signal HLS_FL of vehicle wheel 11, the translational acceleration signals 6DoF-Body_x, 6DoF-Body_y, 6DoF-Body_z of vehicle body 2, the yaw rate signals Yaw-rate-Body_x and Yaw-rate-Body_z of vehicle body 2, the translational acceleration signals Knuckle_RR_x, Knuckle_RR_y and Knuckle_RR_z of vehicle wheel 12 and the translational acceleration signals Knuckle_RL_x, Knuckle_RL_y and Knuckle_RL_z of vehicle wheel 13. In Fig. 6, the hyphens between the terms “6DoF” and “Body” and between the terms “Yaw”, “rate” and “Body” have been omitted from the signal designations. By means of the evaluation unit 45, a sliding time window 63 with a predetermined number of consecutive values ​​is placed over the time series HLS_FR , HLS_FL, 6DoF-Body_x, 6DoF-Body_y, 6DoF-Body_z, Yaw-rate-Body_x, Yaw-rate-Body_z, Knuckle_RR_x, Knuckle_RR_y, Knuckle_RR_z, Knuckle_RL_x, Knuckle_RL_y and Knuckle_RL_z, whereby the values ​​of the time series lying within the time window 63 are supplied to the input layer 60 as input signals. Preferably, the input layer 60 is additionally supplied with the mass signal Smass, the drive torque signal Smotor, and the brake pressure signal Sb. Optionally, a steering angle signal Slw, characterizing the steering angle of the vehicle wheel 10 or 11, is also supplied to the input layer 60. Preferably, the time window 63 is also applied to the signals Smotor, Sb, and Slw, which are particularly available in the form of time series, in a corresponding manner. The neural network 59 determines the currently estimated pitch angle θ_N from the data from time window 63 in the form of the pitch angle signal S_θ_N and provides this as an output signal at the output layer 62. Before the neural network 59 is used, it is trained with reference data at different vehicle load states 1. For this purpose, pitch angle reference data, shown in the bottom row of Fig. 6 and also referred to there as "pitch angle", are provided. A deviation between the angle θ_N estimated by the neural network 59 and a reference value θref at the end of the time window 63 can be used to train the neural network 59. In addition, with reference to Fig. 5, it should be noted that the signals acquired by the sensors can be subjected to preprocessing 64 before being fed to the neural network 59. This preprocessing 64 includes, in particular, normalization of the signals. Additionally, the preprocessing can include filtering of the signals. Optionally, feature extraction can also be provided, such as calculating root mean squares from the acquired and / or preprocessed signals. According to the described embodiment, five sensor units are provided, namely the inertial measuring unit 30, the multidimensional accelerometer or inertial measuring unit 37, the multidimensional accelerometer or inertial measuring unit 41, the height sensor 28, and the height sensor 29. Additionally, a height sensor may be provided on the wheel suspension 6 and / or a height sensor on the wheel suspension 7. Furthermore, a multidimensional accelerometer or an inertial measuring unit may be provided on the wheel suspension 4 and / or a multidimensional accelerometer or an inertial measuring unit on the wheel suspension 5. At a minimum, only two sensor units are required, for example, the first sensor unit 30 and the second sensor unit 37 or the third sensor unit 41, or, for example, the second sensor unit 37 or the third sensor unit 41 and the height sensor 28 or the height sensor 29. Reference sign 1 Vehicle 2 Vehicle body 3 Chassis 4 Wheel suspension 5 Wheel suspension 6 Wheel suspension 7 Wheel suspension 8 Front axle 9 Rear axle 10 Vehicle wheel 11 Vehicle wheel 12 Vehicle wheel 13 Vehicle wheel 14 Wheel carrier 15 Joint 16 Suspension link 17 Joint 18 Strut 19 Strut mount 20 Vehicle spring 21 Damper 22 Wheel bearing 23 Wheel pivot 24 Tie rod 25 Joint 26 Ground surface 27 Tilt angle detection device 28 Height sensor 29 Height sensor 30 Sensor unit / inertial measuring unit 31 Translational acceleration sensor 32 Translational acceleration sensor 33 Translational acceleration sensor 34 Yaw rate sensor 35 Yaw rate sensor 36 Yaw rate sensor 37 Sensor unit / multidimensional acceleration sensor 38 Translational acceleration sensor 39 Translational acceleration sensor 40 Translational acceleration sensor 41 Sensor unit / multidimensional accelerometer 42 Translational accelerometer 43 Translational44 Acceleration sensor 45 Translational acceleration sensor 46 Evaluation unit 46 Drive motor 47 Drive motor shaft 48 Control unit 49 Bus system 50 Brake pedal 51 Brake pressure sensor 52 Mass detection unit 53 Vehicle load 54 Headlight assembly 55 Headlights 56 Headlights 57 Light 58 Headlight range control unit 59 Neural network 60 Input layer 61 Intermediate layer 62 Output layer 63 Time window 64 Signal preprocessing α Angle θ_N Pitch angle θref Reference value φ Beam angle h Altitude h_ref Reference position HLS_FR Altitude signal HLS_FL Altitude signal Slw Steering angle signal S_θ_N Output signal / Pitch angle signal Sb Brake pressure signal Smass Mass signal Smotor Drive torque signal 6DoF-Body_x Acceleration signal of the body 6DoF-Body_y Acceleration signal of the body 6DoF-Body_z Acceleration signal of the setup Yaw-rate-Body_x Rotation rate signal of the setup Yaw-rate-Body_y Rotation rate signal of the setup Yaw-rate-Body_zRotation rate signal of the setup Knuckle_RR_x Acceleration signal at the vehicle wheel Knuckle_RR_y Acceleration signal at the vehicle wheel Knuckle_RR_z Acceleration signal at the vehicle wheel Knuckle_RL_x Acceleration signal at the vehicle wheel Knuckle_RL_y Acceleration signal at the vehicle wheel Knuckle_RL_z Acceleration signal at the vehicle wheel

Claims

Tilt angle detection device in a vehicle (1) which has a chassis (3) with several vehicle wheels (10, 11, 12, 13) and several vehicle axles (8, 9) spaced apart from each other in a longitudinal direction (x) of the vehicle, each of which at least one of the vehicle wheels (10, 11, 12, 13) is assigned, and a vehicle body (2) supported by the chassis (3), wherein the tilt angle detection device (27) has several sensor units (30, 37), of which a first sensor unit (30) is provided on a first chassis component (14) assigned to a first of the vehicle axles (8) or on the vehicle body (2) and a second sensor unit (37) is provided on a second chassis component assigned to a second of the vehicle axles (9), and has an evaluation unit (45) connected to the sensor units (30, 37),wherein the first sensor unit (30) can detect the orientation of the first chassis component (14) or the vehicle body (2) relative to the earth (E) or the height at the first vehicle axle (8) and first sensor signals characterizing this orientation or height can be provided, and wherein the second sensor unit (37) can detect the orientation of the second chassis component relative to the earth (E) and second sensor signals characterizing this orientation can be provided, characterized in that the first sensor signals can be provided in the form of at least one first time series and the second sensor signals in the form of at least one second time series, the evaluation unit (45) comprises a neural network (59) with an output layer (62) and an input layer (60), to which values ​​of the time series or values ​​derived therefrom can be supplied as input signals,and at least one output signal (S_θ_N) characterizing an inclination (θ_N) of the vehicle structure (2) can be estimated by the neural network (59) and made available at the output layer (62). Inclination angle detection device according to claim 1, characterized in that a sliding time window (63) with a predetermined number of consecutive values ​​can be placed over the time series by means of the evaluation unit (45) and the values ​​of the time series lying within the time window or values ​​derived therefrom can be supplied to the input layer (60) as the input signals. Device according to claim 1 or 2, characterized in that the values ​​of the time series before the input layer (60) can be normalized by means of the evaluation unit (45). Device according to one of the preceding claims, characterized in that the evaluation unit (45) comprises or is connected to at least one mass detection unit (52) by means of which a mass of the vehicle (1) or a load (53) of the vehicle (1) can be determined and at least one mass signal (Smass) characterizing this mass can additionally be supplied to the input layer (60) as an input signal. Device according to claim 4, characterized in that the neural network can be trained with different loading states of the vehicle (1). Device according to one of the preceding claims, characterized in that the output signal (S_θ_N) describes a pitch angle (θ_N) of the vehicle (1) or one or more height positions of the vehicle (1). Device according to one of the preceding claims, characterized in that the first sensor unit (37) comprises an inertial measuring unit. Device according to one of the preceding claims, characterized in that the second sensor unit (30) comprises a multidimensional acceleration sensor or an inertial measuring unit. Device according to one of the preceding claims, characterized in that a height level (h) of the first vehicle axle (8) can be detected by means of a further sensor unit (28) and further sensor signals (HLS_FR) characterizing this height level can be provided in the form of at least one further time series. Device according to one of the preceding claims, characterized by a front headlight device (54) with at least one adjustable front headlight (55, 56) by means of which light (57) can be emitted, the beam angle (φ) of which can be controlled depending on the at least one output signal (S_θ_N). Method for detecting the tilt angle in a vehicle (1) which has a chassis (3) with several vehicle wheels (10, 11, 12, 13) and several vehicle axles (8, 9) spaced apart from each other in a longitudinal direction (x) of the vehicle, each of which at least one of the vehicle wheels (10, 11, 12, 13) is assigned, and a vehicle body (2) supported by the chassis (3), wherein a tilt angle detection device (27) has several sensor units (30, 37), of which a first sensor unit (30) is provided on a first chassis component (14) assigned to a first of the vehicle axles (8) or on the vehicle body (2) and a second sensor unit (37) is provided on a second chassis component assigned to a second of the vehicle axles (9), and an evaluation unit (45) connected to the sensor units (30, 37),wherein the first sensor unit (30) detects the orientation of the first chassis component (14) or the vehicle body (2) relative to the earth (E) or the height at the first vehicle axle (8) and provides first sensor signals characterizing this orientation or height, and wherein the second sensor unit (37) detects the orientation of the second chassis component (14) relative to the earth (E) and provides second sensor signals characterizing this orientation, characterized in that the first sensor signals are provided in the form of at least one first time series and the second sensor signals in the form of at least one second time series, the evaluation unit (45) comprises a neural network (59) with an output layer (62) and an input layer (60), to which values ​​of the time series or values ​​derived therefrom are supplied as input signals,and at least one output signal (S_θ_N) characterizing an inclination (θ_N) of the vehicle body (2) is estimated by the neural network (59) and provided at the output layer (62). Method according to claim 11, characterized in that a mass of the vehicle (1) or of a load (53) of the vehicle (1) is determined and at least one mass signal (Smass) characterizing this mass is additionally supplied to the input layer (60) as an input signal. Method according to claim 12, characterized in that the neural network is trained with different loading states of the vehicle (1).

Images