Computer-implemented method for determining a road gradient of a roadway traveled by a vehicle

Abstract

The invention relates to a computer-implemented method (100) for determining a road gradient of a roadway traveled by a vehicle (300), comprising the steps of: determining (110) the road gradient using a Kalman filter based on an acceleration of the vehicle (300), wherein in a prediction step of the Kalman filter, the road gradient is estimated using a mathematical system model of the vehicle (300) and an associated process error, and in a subsequent correction step of the Kalman filter, the estimated road gradient is corrected by one or more parameter values of the vehicle (300) and associated measurement errors; providing (120) a plurality of characteristic maps for various parameters detectable by the vehicle (300), each of which is configured to output an adjusted process error and / or measurement error based on a parameter value detected for the vehicle (300);Detecting (130) at least one first parameter value of a first detectable parameter of the detectable parameters and determining (140) at least one associated first characteristic map of the plurality of characteristic maps; outputting (150) an adjusted process error and / or measurement error to the Kalman filter based on the first parameter value and the first characteristic map;

Description

[0001] The invention relates to a computer-implemented method and a device for determining a road gradient of a roadway traveled by a vehicle, as well as a computer program product.

[0002] When controlling a vehicle, for example when changing gear, a gradient of a road surface traveled by the vehicle can be taken into account in order to keep mechanical stress on the vehicle and its components as low as possible while simultaneously ensuring efficient control.

[0003] The road gradient can be calculated using IMUs (Inertial Measurement Units). IMUs combine various inertial sensors (e.g., gyroscopes and accelerometers) to calculate the angle around one or more spatial axes using the measured data (in this case, angular rate and acceleration).

[0004] Calculating the road gradient implemented on vehicle control units can also be done purely using acceleration sensors, using the measured acceleration from a vehicle's acceleration sensor. Gravitational acceleration "g" can be used as a guide when calculating the road gradient, as it points towards the center of gravity / Earth with sufficiently high accuracy.

[0005] To enable sufficiently accurate calculation of the road gradient even in dynamic situations such as driving with various "disturbances" (vehicle acceleration, vibrations, centrifugal acceleration when cornering, etc.), mathematical methods are used to estimate the current gradient, both with IMUs and pure acceleration sensors. A typical mathematical method here is the Kalman filter.

[0006] Here, gradient refers to both the gradient around the vehicle's lateral axis and the longitudinal axis. The angle around the lateral axis represents the vehicle's pitch angle, and the angle around the longitudinal axis represents the vehicle's roll angle. The road gradient is often equated with the vehicle's pitch angle.

[0007] DE 10 2014 201 769 A1 describes a method for calculating gradients using acceleration data in a vehicle via a Kalman filter. Vehicle body movements are also taken into account to improve the road gradient calculation. DE 10 2014 201 769 A1 discloses that a choice is made between a parameterization for small expected vehicle body movements and a parameterization for large expected vehicle body movements. This allows switching between a robust parameterization for large vehicle body movements and a fast parameterization for small vehicle body movements.

[0008] US 10 124 806 B2 relates to real-time vehicle mass and road gradient estimation. In particular, the present invention relates to the application of Kalman filter theory and the recursive least squares algorithm with forgetting factor to real-time vehicle mass and road gradient estimation. CN 1 16 674 571 A relates to a method for real-time estimation of the mass and gradient of an automobile based on a data confidence factor. DE 10 2017 209 747 A1 relates to a method for determining a gradient of a roadway. The invention also relates to a vehicle configured to perform this method. The invention is particularly advantageously applicable to two-wheeled vehicles, in particular motorcycles.

[0009] Highly dynamic driving situations and poor vehicle surfaces with significant and changing slip pose a challenge when calculating road gradients. These driving situations also occur very frequently in off-road environments (e.g., driving on wet ground, driving on unpaved surfaces such as gravel pits, etc.).

[0010] It is an object of the present invention to provide a computer-implemented method and a device for determining a road gradient of a roadway traveled by a vehicle, which at least improve one or more of the aforementioned disadvantages. In particular, it is an object of the present invention to provide an improved and more precise determination of the road gradient using a Kalman filter.

[0011] According to a first aspect, the problem is solved by a computer-implemented method for determining a road gradient of a roadway traveled by a vehicle. The method comprises the following steps: - Determining the road gradient by means of a Kalman filter based on an acceleration of the vehicle, wherein in a prediction step of the Kalman filter the road gradient is estimated by means of a mathematical system model of the vehicle and an associated process error and in a subsequent correction step of the Kalman filter the estimated road gradient is corrected by one or more parameter values of the vehicle and associated measurement errors; - Providing a plurality of characteristic maps for various parameters detectable by the vehicle, each of which is configured to output an adjusted process error and / or measurement error based on a parameter value detected for the vehicle; - detecting at least one first parameter value of a first detectable parameter of the detectable parameters and determining at least one associated first characteristic map of the plurality of characteristic maps; - Outputting an adjusted process error and / or measurement error to the Kalman filter based on the first parameter value and the first map.

[0012] One aspect of the present invention is that in situations where, for example, the system model has high / low inaccuracies, the process noise can be adjusted. Furthermore, in situations where the measurement has high / low inaccuracies, the measurement noise can be adjusted. This can now also be used to improve the gradient calculation via the Kalman filter in highly dynamic situations, for example in off-road areas. The vehicle can be a motorized two-wheeler, a passenger car, or a commercial vehicle that is operated on roads and / or off-road in what is known as off-road areas. Furthermore, the vehicle can be designed as a self-propelled work machine in the form of a construction, agricultural, or forestry machine.

[0013] The first recorded parameter, as well as additional recorded parameters, can characterize a vehicle's driving situation, and the Kalman filter parameters can be dynamically adjusted to the situation. This allows for a more precise determination of the road gradient. Instead of providing a multitude of maps, a single multidimensional map can be provided. Alternatively or additionally, several multidimensional maps can also be provided.

[0014] The road gradient can be the current road gradient. Road gradient here can refer to both the gradient around the vehicle's lateral axis and the longitudinal axis. The angle around the lateral axis represents the vehicle's pitch angle, and the angle around the longitudinal axis represents the vehicle's roll angle.

[0015] The roadway can be a surface on which the vehicle travels.

[0016] The system model of the vehicle can be a mathematical description of the vehicle. The system model can describe one or more states of the vehicle, particularly as a function of time. The system model can include acceleration.

[0017] The Kalman filter, which is used to determine the current road gradient taking acceleration into account, is essentially composed, within the meaning of the invention and in a manner known in principle to those skilled in the art, of a prediction step (also called the prediction step) and a correction step. During the prediction step, a previous estimate of the road gradient is subjected to state dynamics to obtain a prediction of the road gradient for the current point in time. This prediction is then corrected in the subsequent correction step using current measured values, in this case parameter values, and thus more closely aligned with the actual prevailing road gradient.This process is then repeated for the next discrete time point to determine the road gradient, whereby with an increasing number of time points, according to the characteristics of a Kalman filter, the estimate of the road gradient is increasingly adjusted to the actual conditions.

[0018] The determination of the road gradient using the Kalman filter can be done based on other vehicle parameters.

[0019] Providing the plurality of characteristic maps may include storing the plurality of characteristic maps. The plurality of characteristic maps may be storable in a memory and / or a controller of the vehicle.

[0020] The measurement and process errors can be filter parameterizations of the Kalman filter.

[0021] The characteristic maps of the plurality of characteristic maps can each be a set of data points of a sampled function F with N variables. By searching, interpolating and / or extrapolating within the set, the characteristic map can map input values, here the at least one first parameter value, to output values, here the process and / or measurement error. The characteristic map can characterize a predetermined driving situation of the vehicle. For example, a characteristic map can receive as an input value a speed detected by a speed sensor of the vehicle. If, for example, the gradient of the measured speed(s) exceeds / falls below a certain threshold value, this can be an indication of a spinning wheel / locking wheel. The higher / lower this gradient is, the weaker / stronger the adaptation of the filter parameterization can be.

[0022] Detecting at least the first parameter value may include detecting the first parameter value using a sensor of the vehicle and / or an external sensor or an external measuring device. Detecting may include or be receiving the first parameter value from the sensor of the vehicle and / or the external sensor or the external measuring device.

[0023] The multitude of characteristic maps can be indicative of the vehicle’s respective driving situations.

[0024] The first characteristic map can characterize predetermined adjusted process errors and / or measurement errors for predetermined parameter values. The first parameter value can be a parameter value that differs from the predetermined parameter values of the first characteristic map. The method can further comprise: - interpolating or extrapolating the predetermined parameter values and / or the associated adjusted process errors and / or measurement errors of the first characteristic map such that an interpolated or extrapolated adjusted process error and / or an interpolated or extrapolated adjusted measurement error can be assigned to the first parameter value; - Output the adjusted interpolated or extrapolated process error and / or measurement error to the Kalman filter.

[0025] For example, the predetermined parameter values of the first map may be 0, 1, 2, 3, 4, and 5, with the first parameter value being 1.4. Thus, the first parameter value lies between 1 and 2. By interpolating the predetermined parameter values and the process and / or measurement errors associated with the predetermined parameter values, an interpolated process and / or measurement error for the first parameter value of 1.2 may be determined.

[0026] The at least one first parameter may characterise one of the following: one or more differential locks engaged; a rotational speed of a wheel, axle or transmission drive of the vehicle; a speed of the vehicle determined by an external source; an acceleration of a vehicle axle of the vehicle; a reversing maneuver of the vehicle; a working process of the vehicle; a position of the vehicle; a course of a steering angle of the vehicle; or a subsurface condition of the road surface.

[0027] In cases where one or more of the vehicle's differential locks, particularly longitudinal and / or transverse locks, are engaged, it can be assumed that the vehicle is being driven on poor ground with significant slippage. This can be taken into account when determining the process and measurement error. At least one of the plurality of maps can be configured to determine an adjusted process and / or measurement error based on the number and / or type of differential locks engaged in the vehicle.

[0028] A map of the plurality of maps can be assigned to at least one parameter, two, three or more parameters.

[0029] The speed of the vehicle can be detected by means of a rotational speed of a rotational speed sensor of the vehicle. If the speed is determined indirectly via the rotational speed sensor on a wheel, axle, or transmission output of the vehicle, the determination of the vehicle speed is subject to errors in slip situations. Consequently, this does not correspond to the actual speed. The at least one characteristic map of the plurality of characteristic maps assigned to the rotational speed can be designed such that the process and / or measurement error is adjusted based on a gradient of detected rotational speeds of the wheel, axle, or transmission output. For this purpose, the method can comprise determining at least first to third rotational speeds and first and second gradients. The first gradient can be determined based on the first and second rotational speeds, and the second gradient can be determined based on the second and third rotational speeds.If the first gradient is higher than the second gradient, an adjustment of the process and / or measurement errors based on the first gradient may be smaller and / or weaker compared to the adjustment of the process and / or measurement errors based on the second gradient.

[0030] The speed or a second speed of the vehicle can be determined via an external source (e.g. via GNSS). This (second) speed can be used to determine the slip. In the slip situation described above, a difference can therefore arise between the speed determined using the external source and the speed determined using the speed sensor. At least one characteristic map of the plurality of characteristic maps can be assigned to the speed determined using the external source. This characteristic map can be used in particular when the speed determined using the external source differs from the speed determined using the speed sensor by a predetermined amount.This characteristic map can be designed to adapt the process and / or measurement error based on the vehicle speed determined by means of the external source and / or the vehicle speed determined by means of the speed sensor.

[0031] A first portion of the plurality of maps can be used to adjust the process and / or measurement error when the vehicle is stationary. Alternatively or additionally, a second portion of the plurality of maps can be used to adjust the process and / or measurement error when the vehicle is moving.

[0032] The acceleration can be detected using an acceleration sensor of the vehicle. If the acceleration sensor detects (parameterizable) strong acceleration patterns / peaks in the vehicle's vertical axis over a longer period of time, this can be an indication of an uneven surface. At least one of the plurality of characteristic maps associated with the acceleration can be configured such that the adjustments to the process and / or measurement error are based on the acceleration patterns and / or peaks.

[0033] When a vehicle performs a reversing maneuver (e.g., a D->R change: the vehicle decelerates sharply and continues driving in the opposite direction), this information can be used to select the appropriate Kalman filter parameters for this situation. This is a typical driving situation for wheel loaders (but also for other construction machinery such as dump trucks), where high vehicle dynamics are expected. Consequently, at least one of the numerous maps can be configured to adapt the process and / or measurement error based on the reversing maneuver.

[0034] Depending on the vehicle's operating process, particularly in the form of a work machine, corresponding measurement and process errors are to be expected in the Kalman filter. For this purpose, the method can further comprise determining, particularly in advance during a vehicle development process, the process and / or measurement errors before the vehicle is deployed. At least one of the plurality of characteristic maps can be assigned to one or more operating processes, with the respective characteristic map being configured to adapt the process and / or measurement errors based on the previously determined process and / or measurement errors. If the operating process of the vehicle's control system is known (e.g., as information already available on a signal interface), the most suitable filter parameterization can be selected.

[0035] The process error and / or the measurement error can be adjusted by a process error value and / or measurement error value using the plurality of characteristic maps. The process error value and / or the measurement error value can be calculated based on the acquired first parameter value.

[0036] The adjustment of the process and / or measurement errors can be performed dynamically. The adjustment of the process and / or measurement errors can be performed within predetermined time points. Alternatively or additionally, the adjustment of the process and / or measurement errors can be performed upon a change in at least the first parameter value.

[0037] The process error and / or the measurement error can be adjusted by means of the plurality of characteristic maps in such a way that the adjusted process error and / or the adjusted measurement error are adjusted sub-proportionately, proportionally or disproportionately to the first parameter value.

[0038] This allows the process and / or measurement errors to be adjusted to suit the driving situation.

[0039] The characteristic maps of the plurality of characteristic maps can be one- and / or multi-dimensional characteristic maps. A parameter value usable for the multi-dimensional characteristic maps can be a multi-dimensional parameter vector with respective entered parameter values. The dimensions and / or the number of dimensions of the parameter vector and the associated characteristic map can be identical.

[0040] The procedure may include: - detecting a second parameter value of a second detectable parameter of the detectable parameters and determining an associated second characteristic map of the plurality of characteristic maps; - weighting the process errors and / or measurement errors output by means of the first and second maps based on predetermined weight proportions of the first and second maps in order to determine a weighted adjusted process error and / or measurement error; - Output the weighted adjusted process error and / or measurement error to the Kalman filter.

[0041] If multiple driving situations and / or recorded parameter values (here, the first and second parameter values) are present simultaneously, they can be weighted. For example, the sum of the weightings can be 100%, with the sum being distributed in predetermined proportions among the respective maps.

[0042] The method may further comprise providing a plurality of Kalman filters that are at least partially different from one another. At least one characteristic map of the plurality of characteristic maps may be assigned to each of the plurality of Kalman filters, wherein the adjusted process error and / or measurement error is output to the Kalman filter assigned to the characteristic map used.

[0043] Consequently, a different Kalman filter can be used depending on the parameter value acquired.

[0044] According to a second aspect, the object is achieved by a device for determining a road gradient of a roadway traveled by a vehicle. The device comprises means for carrying out the method according to the first aspect. The means can comprise a controller for executing a computer program product mentioned below and / or a memory for storing the computer program product. The controller can be a processor and / or comprise such a processor. Alternatively or additionally, the means can comprise a communication unit designed to send and / or receive information. The communication unit can be designed to receive the at least one first parameter value.

[0045] According to a third aspect, the object is achieved by a computer program product comprising instructions that, when the program is executed by a processor, cause the processor to execute the method according to the first aspect. Alternatively or additionally, the computer program product may comprise instructions that, when the program is executed by the device according to the second aspect, in particular the control of the device, cause the device to execute the method according to the first aspect.

[0046] The object is achieved according to a fourth aspect by a vehicle comprising the device according to the second aspect and / or a memory for storing the computer program product according to the third aspect and a processor for executing the computer program product.

[0047] Preferred embodiments are explained using the accompanying figures. They show: Fig. 1 a schematic representation of a computer-implemented method for determining a road gradient; Fig. 2 a schematic representation of a device for determining a road gradient; and Fig. 3 a schematic representation of a vehicle with such a device.

[0048] In the figures, identical or essentially functionally identical or similar elements are designated by the same reference numerals.

[0049] Fig. 1 shows a schematic representation of a computer-implemented method 100 for determining a road gradient of a roadway traveled by a vehicle 300. The method 100 can be stored in the form of a computer program product in a memory 210 and / or a controller 220, for example, a device 200 mentioned below.

[0050] The method 100 includes determining 110 the road gradient using a Kalman filter based on an acceleration of the vehicle 300. Further parameters of the vehicle 300 can be used to determine 110 the road gradient. In a prediction step of the Kalman filter, the road gradient is estimated using a mathematical system model of the vehicle 300 and an associated process error. In a subsequent correction step of the Kalman filter, the estimated road gradient is corrected using one or more parameter values of the vehicle 300 and associated measurement errors.

[0051] The method 100 further comprises providing 120 a plurality of characteristic maps for various parameters detectable by the vehicle 300. The plurality of characteristic maps are each configured to output an adjusted process error and / or measurement error based on a parameter value of the detectable parameters detected for the vehicle 300.

[0052] The detectable parameters can characterize one or more engaged differential locks, a rotational speed of a wheel, an axle or a transmission drive, a speed determined by means of an external source, an acceleration of a vehicle axle, a reversing process, a work process, a position, a course of a steering angle and / or a surface condition of the roadway of the vehicle 300.

[0053] Furthermore, the method 100 comprises detecting 130 at least a first parameter value of a first detectable parameter of the detectable parameters. For example, the first parameter is the rotational speed of the wheel and the first parameter value is a number.

[0054] The method 100 further comprises determining 140 at least one associated first characteristic map of the plurality of characteristic maps. In the example of the rotational speed, at least one of the characteristic maps, here the first characteristic map, is configured to adapt the process and / or measurement error based on the detected rotational speed.

[0055] For example, if the gradient of the measured speed(s) exceeds / falls below a certain threshold, this may be an indication of a spinning wheel / locking wheel of vehicle 300. The higher / lower this gradient is, the weaker / stronger the adjustment of the filter parameterization using the first map may be.

[0056] The method 100 further comprises outputting 150 one or more of the adjusted process errors and / or measurement errors to the Kalman filter based on the first parameter value and the first characteristic map.

[0057] Due to the limited storage space on memories 210 and / or controllers 220 of vehicle 300, the characteristic maps of the plurality of characteristic maps can advantageously map predetermined parameter values to respective predetermined adjusted process and / or measurement errors. For example, the predetermined parameter values can be discrete values. If the detected first parameter value is a non-discrete parameter value, the predetermined parameter values can be interpolated or extrapolated. The interpolation or extrapolation can be performed such that the interpolation or extrapolation range encompasses the first parameter value. Thus, an interpolated or extrapolated process and / or measurement error can be determined.

[0058] This allows for a reduction in the required storage space. Furthermore, a corresponding adjusted process and / or measurement error can be determined for a large number of discrete and non-discrete recorded parameter values. This further enables the process and / or measurement errors to be adjusted to the respective driving situation, thus improving the accuracy of determining the road gradient.

[0059] If multiple parameter values are acquired, two or more of the multiple maps can be used. The process and / or measurement errors adjusted by the maps used can be weighted to determine a final process and / or measurement error based on the weighted process and / or measurement errors of each map used.

[0060] Fig. 2 shows a device 200 for determining a road gradient of a roadway traveled by a vehicle 300. Device 200 includes means 210, 220 for executing method 100.

[0061] The means can be provided in the form of a memory 210 and a controller 220. The memory 210 can be configured to store a computer program product comprising instructions which, when the program is executed by a processor and / or the controller 220, cause the processor and / or the controller to execute the method 100. The controller 220 can comprise or be the processor. The device 200 can comprise further units for executing the method 100. For example, the device 100 can further comprise a communication unit for sending and / or receiving information. The communication unit can be configured to receive at least the first parameter value. The communication unit can further be configured to send the adjusted process and / or measurement errors to the controller of the vehicle 300.

[0062] Fig.3 shows a vehicle 300, for example a work machine and / or a vehicle 300 intended for agricultural use, comprising the device 200. The vehicle 300 may include a controller for controlling the vehicle 300. Reference symbol 100 Computer-implemented method for determining a road gradient 110 Determining the road gradient using a Kalman filter 120 Providing a variety of maps 130 Recording at least a first parameter value 140 Determining at least one associated first characteristic map 150 Output of an adjusted process error and / or measurement error 200 Device for determining a road gradient 210 storage 220 Control 300 vehicles

Claims

[1] Computer-implemented method (100) for determining a road gradient of a roadway traveled by a vehicle (300), comprising the steps: Determining (110) the road gradient by means of a Kalman filter based on an acceleration of the vehicle (300), wherein in a prediction step of the Kalman filter the road gradient is estimated by means of a mathematical system model of the vehicle (300) and an associated process error and in a subsequent correction step of the Kalman filter the estimated road gradient is corrected by one or more parameter values of the vehicle (300) and associated measurement errors; Providing (120) a plurality of characteristic maps for various parameters detectable by the vehicle (300), each configured to output an adjusted process error and / or measurement error based on a parameter value detected for the vehicle (300); Detecting (130) at least one first parameter value of a first detectable parameter of the detectable parameters and determining (140) at least one associated first characteristic map of the plurality of characteristic maps; Outputting (150) an adjusted process error and / or measurement error to the Kalman filter based on the first parameter value and the first map. [2] Method (100) according to claim 1, wherein the plurality of characteristic maps are indicative of respective driving situations of the vehicle (300). [3] Method (100) according to claim 1 or 2, wherein the first characteristic map characterizes predetermined adapted process errors and / or measurement errors for predetermined parameter values, wherein the first parameter value is a parameter value different from the predetermined parameter values of the first characteristic map, the method (100) further comprising: Interpolating or extrapolating the predetermined parameter values and / or the associated adjusted process errors and / or measurement errors of the first characteristic map such that an interpolated or extrapolated adjusted process error and / or an interpolated or extrapolated adjusted measurement error can be assigned to the first parameter value; Output the adjusted interpolated or extrapolated process error and / or measurement error to the Kalman filter. [4] Method (100) according to one of the preceding claims, wherein the at least one first parameter characterizes one of the following: one or more differential locks of the vehicle (300) engaged; a rotational speed of a wheel, axle or transmission drive of the vehicle (300); a speed of the vehicle (300) determined by means of an external source; an acceleration of a vehicle axle of the vehicle (300); a reversing operation of the vehicle (300); a working process of the vehicle (300); a position of the vehicle (300); a course of a steering angle of the vehicle (300); or a subsurface condition of the road surface. [5] Method (100) according to one of the preceding claims, wherein the process error and / or the measurement error are adjusted by means of the plurality of characteristic maps by a process error value and / or measurement error value, wherein the process error value and / or the measurement error value are calculated based on the detected first parameter value. [6] Method (100) according to one of the preceding claims, wherein the process error and / or the measurement error are adjusted by means of the plurality of characteristic maps such that the adjusted process error and / or the adjusted measurement error is adjusted sub-proportionately, proportionally or disproportionately to the first parameter value. [7] Method (100) according to one of the preceding claims, wherein the characteristic maps of the plurality of characteristic maps are one- and / or multi-dimensional characteristic maps, where a parameter value usable for the multidimensional characteristic maps is a multidimensional parameter vector with respective entered parameter values. [8] Method (100) according to one of the preceding claims, further comprising: Detecting a second parameter value of a second detected parameter of the detectable parameters and determining an associated second characteristic map of the plurality of characteristic maps; Weighting the process errors and / or measurement errors output by means of the first and second maps based on predetermined weight proportions of the first and second maps to determine a weighted adjusted process error and / or measurement error; Output the weighted adjusted process error and / or measurement error to the Kalman filter. [9] Method (100) according to one of the preceding claims, further comprising: Providing a plurality of at least partially different Kalman filters, wherein each of the plurality of Kalman filters is assigned at least one characteristic map of the plurality of characteristic maps, wherein the adjusted process error and / or measurement error is output to the Kalman filter assigned to the characteristic map used. [10] Device (200) for determining a road gradient of a roadway traveled by a vehicle (300), comprising means (210; 220) for carrying out the method (100) according to one of claims 1 to 9. [11] A computer program product comprising instructions which, when executed by a processor, cause the processor to carry out the method (100) according to any one of claims 1 to 9. [12] Vehicle (300) comprising the device (200) according to claim 10 and / or a memory for storing the computer program product according to claim 11 and a processor for executing the computer program product.

Images