Method for determining total mass of motor vehicle

Abstract

The invention relates to a method (100) for determining the total mass of a motor vehicle having at least one wheel during operation of the motor vehicle using vehicle longitudinal force (10) and longitudinal acceleration (20)-related signals on a vehicle bus (1), in particular on a vehicle controller area network bus (CAN), on the basis of the Newton's second law of motion, the longitudinal force (10) and the longitudinal acceleration (20) are frequency-filtered, in particular band-pass frequency-filtered, in particular for removing force components and acceleration components caused by aerodynamic drag and / or slowly changing forces or effects of uphill and / or downhill or the like, the mass calculation being carried out only during a predetermined movement of the vehicle, especially during longitudinal or linear or other suitable movements.

Description

Technical Field

[0001] The present invention relates to a method for determining the total mass of a motor vehicle having at least one wheel during operation. Background Technology

[0002] Today, motor vehicles are equipped with a large number of monitoring and information systems for monitoring the condition of the vehicle and its individual components. These systems, for example, transmit information between different components or provide vehicle information or warning information to the driver.

[0003] One such system is the Tire Information System (TIS). Initially used only to provide tire pressure information, TIS is evolving to offer more comprehensive tire information, such as tread depth. Decreased tread depth is caused by wear between the tire and the road surface. One factor contributing to wear is the vertical force acting on the tire, which is related to vehicle mass.

[0004] Therefore, a method for determining the current gross vehicle weight is needed in relation to determining tread depth or monitoring tire information systems. Furthermore, the determined weight can also be used by other vehicle systems, such as vehicle dynamics control systems and systems for improving driving comfort. By understanding the vehicle's current gross vehicle weight, these systems can improve their control algorithms, further enhancing the safety and / or comfort of the vehicle. Summary of the Invention

[0005] Therefore, one objective of this invention is to provide a method for determining the total mass of a motor vehicle in a cost-effective, simple, and especially computer-resource-saving manner, which is particularly necessary in tire information system scenarios.

[0006] The above-described task is solved by the method according to claim 1. Other embodiments of the invention and beneficial further improvements are described in the following description and dependent claims.

[0007] This invention specifically applies Newton's second law of motion to the longitudinal motion of a vehicle, thus allowing the vehicle's mass to be primarily calculated by dividing the longitudinal force by the longitudinal acceleration. If the drive axle is the front axle, vehicle motion is caused by traction; if the drive axle is the rear axle, vehicle motion is caused by propulsion; for four-wheel drive vehicles, vehicle motion can be calculated from both traction and propulsion. In the following text, to avoid limiting the scope of this invention, only a single drive axle and a single axle torque are considered; in multi-axle drive scenarios, the torques need to be accumulated. Traction and propulsion are not distinguished separately but together constitute the longitudinal force.

[0008] When a vehicle is in motion, an effective longitudinal force is generated at the point of contact between the tires and the road. During linear motion, the force is longitudinal; if wheel slippage is very small, the longitudinal force can be calculated from the wheel torque or drive axle torque and wheel radius. Wheel torque is generated by the vehicle's motor and transmission. On some vehicles, the axle torque value can be obtained directly via the vehicle bus, which may in particular be a Controller Area Network (CAN) bus. On other vehicles, only the motor torque value is provided; the axle torque can be calculated from the gear ratio, which can be obtained directly via the vehicle bus, or from the motor speed (revolutions per minute (RPM)) and wheel speed (RPM) values, which are typically available via the vehicle bus. On some vehicles, the braking torque value is also provided via the vehicle bus and can be incorporated into the longitudinal force calculation.

[0009] The longitudinal forces of a vehicle result in longitudinal acceleration, which is measured in particular by the vehicle's inertial measurement unit and provided via the vehicle bus. The longitudinal forces and corresponding longitudinal accelerations constitute the fundamental information used in this invention.

[0010] Auxiliary information can also typically be obtained from the vehicle bus and used in this invention, such as signals related to driving conditions, including vehicle speed, steering angle, lateral acceleration, yaw rate (yaw rate), and wheel speed. These signals are used to consider only linear longitudinal motion. Other auxiliary information that can be used in this invention is provided by transmission-related signals, typically including signals such as motor speed and wheel speed, while other signals such as gear ratio, gear position, clutch pedal position, and brake pedal position depend on the specific vehicle type and model.

[0011] According to a preferred embodiment, the longitudinal force of the vehicle is calculated using motor torque information and wheel or tire radii available on the vehicle bus, wherein the longitudinal acceleration is preferably obtained from the vehicle bus. Frequency filtering of the longitudinal force and longitudinal acceleration is performed, in particular, before the total mass calculation based on Newton's second law.

[0012] According to another preferred embodiment, both longitudinal force and longitudinal acceleration need to be filtered by low-pass frequency to remove noise, and then filtered by band-pass frequency to remove slowly changing forces such as air resistance, uphill or downhill and their corresponding slowly changing acceleration components.

[0013] According to another preferred embodiment, this processing is then handled by a normal condition filter, which selects a predetermined or appropriate driving scenario and a corresponding time interval involving only linear longitudinal movement. The normal condition filter requires the aforementioned auxiliary information and, using predetermined, and especially appropriately selected, thresholds, performs predetermined, and especially predetermined small turning or lateral movements, such as those provided by lateral acceleration and yaw rate signals, and / or predetermined, and especially predetermined small wheel slips, such as those provided by the relative difference in front and rear wheel speeds, and / or by a constant gear or transmission ratio, and / or a predetermined, and especially predetermined, appropriate vehicle speed and / or longitudinal acceleration range. If the longitudinal force does not include braking force, the normal condition filter also excludes braking situations.

[0014] According to another preferred embodiment, the auxiliary signal, like the longitudinal force and longitudinal acceleration, is filtered using the same low-pass frequency.

[0015] According to another preferred embodiment, if the motor vehicle is equipped with an automatic transmission, and if the instantaneous axle torque of the motor vehicle is available on the vehicle bus, the longitudinal force is calculated by dividing the axle torque by the wheel radius. If the axle torque is not available on the vehicle bus, it is calculated by multiplying the instantaneous motor torque by the gear ratio, which is provided by the vehicle bus or calculated as the ratio of the motor speed to the wheel speed. If braking torque is also available, it can be included in the longitudinal force.

[0016] According to another preferred embodiment, if the motor vehicle is equipped with a manual transmission and no shaft torque is provided on the vehicle bus, the longitudinal force and its associated longitudinal acceleration are calculated and / or used only when the longitudinal force is determined to be valid. Preferably, the longitudinal force is determined to be valid if the vehicle's transmission chain is closed and stable. Particularly preferably, the longitudinal force is determined to be valid if the vehicle's clutch pedal is not engaged and / or the gear is engaged and not in neutral and / or the gear ratio is not affected by transients. If no braking torque is available, the brake pedal not being engaged can serve as an additional verification condition for the validity of the longitudinal force.

[0017] According to another preferred embodiment, the total mass is calculated as the ratio of the root mean square (RMS) value of the longitudinal force to the root mean square (RMS) value of the longitudinal acceleration, wherein both RMS values ​​are calculated within an effective time interval derived by a normal condition filter, which starts from a preset, particularly appropriately selected, initial time and continues to the current time, and may be values ​​obtained through a bandpass filtering step, or values ​​obtained by exponentially smoothing the results of the bandpass filtering step.

[0018] According to another preferred embodiment, the original instantaneous total mass calculation at the current time point is performed within an effective time interval provided by the normal condition filter by dividing the original instantaneous root mean square (RMS) value of the longitudinal force, especially after bandpass filtering, by the original instantaneous root mean square (RMS) value of the longitudinal acceleration, especially after bandpass filtering.

[0019] According to another preferred embodiment, the original instantaneous total mass at the current time point is obtained by linear regression calculation of the paired values ​​of longitudinal force and longitudinal acceleration within the effective time interval provided by the normal condition filter.

[0020] According to another preferred embodiment, within the current valid time interval provided by the normal condition filter, a first raw instantaneous total mass value is calculated and used as the slope of a linear regression of longitudinal force, particularly bandpass-filtered, and longitudinal acceleration, particularly bandpass-filtered. A second raw instantaneous total mass value is calculated as the inverse slope of the linear regression between bandpass-filtered longitudinal acceleration and bandpass-filtered longitudinal force. Preferably, the first and second raw instantaneous total mass values ​​are used to evaluate noise in the longitudinal acceleration and longitudinal force data and / or noise in the first and second raw instantaneous total mass values. Particularly preferably, if the relevant noise exceeds a predetermined threshold, the second raw instantaneous total mass value is discarded. In principle, longitudinal acceleration can be considered as the independent variable, and longitudinal force as the dependent variable. The regression slope of the longitudinal force, as an acceleration function with a mass dimension, is considered as a calculated value of the vehicle mass. Ordinary linear regression assumes that the independent variable values ​​are noise-free and minimizes noise in the dependent variable. Preferably, the longitudinal force is calculated by the vehicle's motor control unit and is relatively noise-free, while the longitudinal acceleration is measured by an accelerometer and is affected by road irregularities and noise, etc. Therefore, it is beneficial to treat longitudinal force as the independent variable and longitudinal acceleration as the dependent variable, and to perform a linear regression on longitudinal acceleration as a function of longitudinal force. Thus, the regression slope has a dimension of the inverse of mass, and the instantaneous total mass of the vehicle is given by a second original instantaneous total mass value, which is calculated as the inverse of the longitudinal acceleration regression slope, as a function of longitudinal force over a normal time interval under the considered conditions. The difference between the first and second original instantaneous total mass values ​​can be used to assess noise in the data; if their relative difference exceeds a predetermined, particularly appropriately chosen, threshold, the current second original instantaneous total mass value is discarded.

[0021] According to another preferred embodiment, the normal time interval needs to meet additional constraints, wherein the constraints are a minimum range of longitudinal acceleration and / or a minimum number of longitudinal force and longitudinal acceleration samples within the relevant time interval. In this way, the total mass can be calculated using the most reliable data.

[0022] According to another preferred embodiment, factors affecting the calculation of the vehicle's total mass, such as reduced motor torque, are incorporated into the longitudinal force calculation. These factors can be considered, for example, gear efficiency or transmission efficiency. Depending on the availability of vehicle bus signals, transmission efficiency can be considered a function of the selected gear or gear ratio. Vehicles equipped with manual transmissions can have engaged gears, which can be easily transmitted directly on the vehicle bus. In vehicles equipped with automatic transmissions, the gear ratio can be transmitted. Alternatively, whether the vehicle is equipped with a manual or automatic transmission, the gear or gear ratio can be determined as the ratio of motor speed to wheel speed, both of which are typically available on the vehicle bus.

[0023] The simplest calculation method assumes that efficiency, depending on the gear or transmission rate, is constant. A first possible implementation related to constant transmission efficiency is to learn the efficiency from scratch without prior knowledge. A second possible implementation related to transmission efficiency is to continue learning with previously learned values ​​and initial values ​​provided; this is called the ES version (derived from Exponential Smoothing, also known as Exponential Moving Average). In a third possible implementation related to transmitter efficiency, the transmitter efficiency is considered to depend on the motor speed and is approximated as a linear or quadratic polynomial.

[0024] According to another preferred embodiment, the power loss on the vehicle transmission chain, depending on the gear ratio or the engaged gear, is taken into account as a subunit efficiency factor (an efficiency factor less than 1). The corrected instantaneous total mass value is calculated by multiplying the original instantaneous total mass value by the subunit efficiency factor (an efficiency factor less than 1). The calculation of transmission efficiency / gear efficiency is preferably based on the gear or gear ratio corresponding to the effective time interval provided by the normal condition filter. Particularly preferably, the gear or gear ratio input to these calculations, depending on the gear efficiency, has a delay time that is the same as the delay time generated by the low-pass filter.

[0025] According to another preferred embodiment, the transmission efficiency / gear efficiency is calculated and learned in a predetermined, especially appropriately performed, corrective driving cycle, particularly when the vehicle has been weighed and its total mass is known, using the original instantaneous total mass calculation value and the known total mass. Preferably, after the transmission efficiency learning is completed, the corrected instantaneous total mass value is calculated during normal vehicle driving operations by multiplying the original instantaneous total mass value by the learned transmission efficiency.

[0026] According to another preferred embodiment, the transmission efficiency / gear efficiency is considered to depend on the vehicle motor speed and approximate as a linear or quadratic polynomial. Polynomial regression can be iteratively calculated for each gear or gear ratio, where data obtained during the training or learning driving cycle is processed and ultimately stored using the current motor speed and the original gear ratio, employing least mean squares or recursive least squares methods. After learning is complete, for each normal time interval, the efficiency value can be retrieved from the current gear and current motor speed (revolutions per minute (RPM)) and used for vehicle gross weight correction.

[0027] According to another preferred embodiment, the dependence of transmission chain efficiency on motor coolant temperature can be similarly incorporated into the correction of longitudinal force calculations and original mass calculations. This dependence can be provided by the vehicle manufacturer for a specific motor oil, or it can be learned during a learning phase, for example, by performing multiple driving cycles using a known vehicle mass at different ambient temperatures, starting from a cold start of the motor, similar to the process of learning transmission / gear efficiency.

[0028] According to another preferred embodiment, the power loss depending on the motor coolant temperature is also taken into consideration as a subunit efficiency factor (an efficiency factor less than 1), wherein the corrected instantaneous total mass value of the vehicle is calculated by multiplying the original instantaneous total mass value by the subunit efficiency factor (an efficiency factor less than 1), wherein preferably, the power loss caused by the motor coolant temperature is learned using a weighed vehicle in several predetermined correction driving cycles, particularly preferably starting from different low temperatures.

[0029] According to another preferred embodiment, instantaneous mass calculation is performed using a statistical method, which involves calculating the average value and the standard error of the average value from the start time point to the current time point. The average value provides a current vehicle mass estimate. The standard error of the average value is a measure of the estimation error; if the relative error, i.e., the ratio between the standard error and the average value, is less than a predetermined, particularly appropriately chosen, threshold, such as 5%, correlation convergence can be considered achieved, and the vehicle mass estimate can be transmitted as the final result of the correlation method.

[0030] According to another preferred embodiment, a statistical method is applied to calculate the total mass at the current time point by means of the original instantaneous total mass value or the corrected instantaneous total mass value, especially the correction for the efficiency of the transmission device, preferably from a predetermined, especially appropriately selected, start time point up to the current time point, especially the end time point of the calculation; or based on the convergence time, in which case the relative error of the total mass calculation (especially assessed as the ratio of the noise measurement value in each temporary mass calculation from the start time point to the current time point—e.g., the standard deviation of the original instantaneous total mass value—to the total mass calculated at the current time) is less than a predetermined, especially appropriately selected, threshold, especially 5%.

[0031] According to another preferred embodiment, the method is used as a function of a vehicle tire information system, preferably to provide useful information for tread depth monitoring or other functions of the tire information system. In particular, it determines the tread depth of the wheel tires based on a determined total mass.

[0032] According to another preferred embodiment, in the first step, a low-pass filter is applied to the longitudinal force and longitudinal acceleration, particularly for noise reduction; in the second step, further frequency filtering, particularly band-pass frequency filtering, is applied to the longitudinal force and longitudinal acceleration, especially to remove force and acceleration components caused by slowly changing forces (such as aerodynamic drag and / or uphill and / or downhill); in the third step, by using auxiliary signals, particularly lateral acceleration and / or yaw rate and / or wheel speed, only linear longitudinal driving scenarios and corresponding quality conditions provided by a condition-normal filter of the auxiliary signals are selected. The calculation time interval is as follows: In the fourth step, the original vehicle mass is calculated by using the root mean square (RMS) values ​​of longitudinal force and longitudinal acceleration or by performing linear regression, especially with longitudinal force as the independent variable and longitudinal acceleration as the dependent variable; In the fifth step, the calculated original mass value is corrected by multiplying the calculated original mass value by an efficiency factor related to the gear or transmission ratio, especially learned in a learning or calibration driving cycle with known vehicle mass; In the sixth step, statistical methods are applied to the corrected original mass value to obtain the total mass of the vehicle, especially the exit time. Attached Figure Description

[0033] The accompanying drawings described herein are provided to better understand the invention and are an integral part of it. The exemplary embodiments and descriptions of the invention are intended to explain the invention and do not constitute an undue limitation thereof. In the drawings:

[0034] Figure 1a A flowchart illustrating a method for determining the total mass of a motor vehicle according to a basic embodiment of the present invention is provided. This method is applicable to vehicles equipped with automatic transmissions and involves the simplest calculations.

[0035] Figure 1b The diagram to d shows the longitudinal force calculation variation based on a specific available Controller Area Network (CAN) bus signal;

[0036] Figure 2 A flowchart illustrating an alternative embodiment of the invention is provided, applicable to vehicles equipped with automatic and manual transmissions, wherein the axle torque signal can be acquired on the controller area network bus (CAN).

[0037] Figure 3A flowchart illustrating another alternative embodiment of the method according to the invention, applicable to manual transmission vehicles, wherein the shaft torque signal cannot be obtained on the controller area network bus (CAN);

[0038] Figure 4a , 4b The speed curves and motor speeds during a specific test drive are displayed in graphical form, with the entire range of motor speeds (revolutions per minute (RPM)) used sequentially across all gears as a training or correction driving cycle.

[0039] Figure 5 Presented in chart form according to Figure 2 The local root mean square (RMS) method and its basis Figure 3 The instantaneous or local quality calculation results of two local regression methods are presented, along with the following: Figure 4a , 4b The relevant gears in the driving cycle shown;

[0040] Figure 6a , 6b Presented in chart form according to Figure 4a , 4b The transmission efficiency learning process in the training or correction driving cycle employs a learning method without prior knowledge and a method that uses the learned efficiency as an initial value and continues to learn using exponential smoothing.

[0041] Figure 7a , 7b Presented in chart form Figure 6a and 6b Calculated values ​​of the original vehicle mass and the efficiency-corrected vehicle mass in both cases;

[0042] Figure 8a , 8b Presented in chart form according to Figure 6a and Figure 6b The two scenarios shown illustrate the estimated vehicle mass and its 95% confidence interval obtained by averaging the efficiency-corrected mass values ​​from the start to the current time, while also displaying the 5% exit time and performance values; and

[0043] Figure 9a The results obtained in another driving cycle are presented in a chart format, as shown in section c, which is compared with those obtained in the following driving cycle. Figure 4a , 4b The training or correction driving cycle shown is conducted under drastically different conditions, in which the transmission efficiency has been fixed. Detailed Implementation

[0044] The present invention will now be described in detail with reference to the accompanying drawings and specific embodiments. It should be noted that the features in the embodiments can be combined with each other. In all figures, corresponding components are always provided with the same reference numerals.

[0045] Figure 1a The diagram illustrates a flowchart of a method 100 for determining the total mass of a motor vehicle according to the present invention, which is applied in its simplest basic form to vehicles equipped with automatic transmissions. During vehicle operation, the vehicle controller local area network (CAN) bus 1 transmits a signal 5 containing instantaneous information related to the traction / propulsion torque generated by the vehicle's motor and its transmission chain. Based on the available signals, step 110 calculates the longitudinal force 10. Simultaneously, the vehicle controller local area network (CAN) bus 1 transmits the instantaneous longitudinal acceleration 20 of the vehicle.

[0046] Subsequently, the longitudinal force 10 and longitudinal acceleration 20 are processed by a noise reduction low-pass filtering step 120a. Then, the low-pass filtered force and acceleration, collectively referred to as 30, are processed by a bandpass processing step 130, the purpose of which is to remove slowly varying force and corresponding slowly varying acceleration components such as air resistance and uphill / downhill gradient.

[0047] Subsequently, the bandpass-filtered force and acceleration, collectively referred to as 40, are transmitted as input to the vehicle mass calculation step 150. This calculation is enabled and / or verified by a condition-normal "filtering" step 140, which uses auxiliary signals, collectively referred to as 50, which are also available via the Controller Area Network (CAN) bus 1 and low-pass filtered in step 120b. The condition-normal "filter" 140 selects appropriate driving conditions and time intervals involving only linear longitudinal motion: it forces smaller cornering or lateral movement provided by lateral acceleration and yaw rate signals, smaller wheel slippage provided by the relative difference in front and rear wheel speeds, a constant gear ratio or transmission rate, and keeps vehicle speed and longitudinal acceleration within appropriate ranges by using predetermined, appropriately selected thresholds. Braking scenarios are also excluded if braking force is not included in the longitudinal force. A corresponding low-pass filter 60 is actually used to prevent erroneous outputs from the sometimes rapidly changing auxiliary signals 50. The condition "filter" 140 consists of a comparator for each condition and an AND gate, and its operation involves no delay. Since its output 70 is used to enable and / or verify the vehicle mass calculation performed in step 150, low-pass filters 120a and 120b are actually the same low-pass filter 120 to ensure that the force and acceleration signals 40, after being filtered by the basic bandpass filter, are synchronized with the verification signal 70. In the simplest form of vehicle mass estimation, the effective values ​​or root mean square (RMS) values ​​of longitudinal force and longitudinal acceleration are calculated over the time interval verified by the condition-normal filter (up to the current time point). The estimated vehicle mass at the current time point is the ratio of the RMS value of longitudinal force to the RMS value of longitudinal acceleration, calculated as described above.

[0048] Figure 1b Figure d shows Figure 1a Variations of longitudinal force calculations 110b to 110d performed in general step 110. Figure 1b In the middle, the axle traction torque / propulsion torque 11 is available, and the longitudinal force 10 is calculated by dividing it by the wheel radius or tire radius. Figure 1c In a preferred embodiment, brake shaft torque 12 can also be used and added. Figure 1d In this case, the shaft torque cannot be directly obtained; it must be calculated by multiplying the motor torque 13 by the gear ratio (which can be obtained directly via the Controller Area Network (CAN) bus 1 or by dividing the motor speed (RPM) by the wheel speed). Besides the example above, other possibilities exist, and other specific CAN signal sets can be easily adapted by those skilled in the art to provide the longitudinal force 10.

[0049] Figure 2The diagram illustrates a flowchart of a preferred embodiment of a method 100 according to the present invention, as an alternative. The method is applicable to vehicles equipped with automatic or manual transmissions, wherein axle torque 11 and possibly braking torque 12 are available. Figure 1a Compared to the basic implementation shown in d, this implementation has some improvements.

[0050] The first improvement is that, in step 151, the instantaneous or local root mean square (RMS) of the bandpass-filtered force and acceleration is calculated within the normal start-to-normal end time interval provided by the normal condition filter 140. Subsequently, the vehicle mass is calculated by dividing the root mean square of the force (RMS) by the root mean square of the acceleration (RMS), both of which are considered at the end of the normal time interval.

[0051] The second improvement is the introduction of transmission efficiency / gear efficiency and its corresponding mass correction in step 160. Using transmission efficiency requires either a gear position signal or a transmission rate signal 15, depending on the available auxiliary signal 50. The gear position signal can be obtained directly from the Controller Area Network (CAN) bus 1; alternatively, the gear ratio can be calculated by dividing the motor speed by the wheel speed. To ensure consistency between the calculated mass / efficiency and the gear position / gear ratio signal 15, step 170 delays the latter by the same time as the delay produced in the low-pass filtering step 120, resulting in a delayed gear position / gear ratio signal 16. The calculated vehicle mass value is then corrected using the gear-related transmission rate as described below.

[0052] Vehicle (original) mass M Raw (M 原始 In principle, it is calculated by the ratio of the longitudinal force to the longitudinal acceleration of the vehicle under this force:

[0053] M Raw = (Longitudinal force) / (Longitudinal acceleration).

[0054] Due to transmission losses, the actual force causing acceleration is relatively small. These losses can be quantitatively characterized by the subunit transmission efficiency factor (a transmission efficiency factor less than 1), Eff. Therefore, the calculated vehicle mass M, including the transmission efficiency factor, is... Eff for:

[0055] M Eff = (Eff * (longitudinal force)) / (longitudinal acceleration) = Eff * M Raw .

[0056] Therefore: Eff = M Eff / M RawThe efficiency factor can be determined under calibration conditions, in which case a scale can be used to measure the mass to be estimated, which is taken as the vehicle's measured mass (VMM). Therefore,

[0057] Eff = VMM / M Raw .

[0058] If multiple measurements are performed, and these measurements are affected by noise (assuming the noise has a mean of zero), then M Raw The average value M calculated from the original mass Raw,Avg replace:

[0059] Eff = VMM / M Raw,Avg .

[0060] Consider the available data index n. An algorithm is needed to calculate M when a new original quality is calculated. Raw,n When available for use, the estimated Eff is Eff. n Therefore, iterative estimation is preferred. Using Eff n Estimate the mass M Eff , n Calculated as

[0061] M Eff、n =Eff n *M Raw,n .

[0062] Because the transmission wear may differ in each gear, it is necessary to calculate the wear for each gear individually. When n is large, Eff... n It should converge to Eff, while M Eff , n It should converge to the vehicle's measured mass VMM.

[0063] To learn transmission / gear efficiency, a measured mass of the vehicle is required. This mass should be provided over one or more driving cycles, which should be set up to realistically cover almost all driving scenarios, gears, speeds, and accelerations, serving as training / learning data for subsequent transmission efficiency learning algorithms. Once learning is complete, the efficiency value is fixed / stored, and the vehicle mass is calculated based on the stored value.

[0064] The simplest case and calculations assume that the efficiency, depending on the gear ratio / transmission rate, is constant. Two implementations of efficiency learning and quality calculation in the relevant case are detailed here; the iteration (new quality calculation) n performed at the end of the effective time interval provided by the condition-normal filter includes the following steps. Local root mean square (RMS) force and acceleration are used in this example, but the original quality calculation results provided by the current local regression can also be used.

[0065] According to the first possible implementation related to transmission efficiency, efficiency is learned from scratch without any prior knowledge: this is called the AN version (derived from the average value—no initialization):

[0066] TE_AN.1: The current raw mass calculation is as follows:

[0067] The average raw mass to date is:

[0068]

[0069] As is well known, the average value can be obtained by iteratively calculating each n.

[0070] TE_AN.2: Original efficiency and corrected current efficiency:

[0071] TE_AN.3: Calculation of current vehicle mass after efficiency correction: M Corr,N =Eff Corr,N ·M Raw,Avg,N .

[0072] When prior learning has been performed and initial values ​​have been provided, a second implementation related to transmission efficiency can be obtained, and the learning process continues—this is called the ES version (derived from exponential smoothing, also known as exponential moving average):

[0073] TE_ES.1: Original current efficiency and its reciprocal: For all N, including 1, the relevant efficiency is calculated iteratively using exponential smoothing, where α = 0.8 or another appropriate value:

[0074]

[0075] For n=1, Eff n-1 =Eff0, which is the initial efficiency.

[0076] Taking inverse efficiency into account, thus for larger N, Eff N Converging to VMM / M Raw,Avg,N (Same as the previous version).

[0077] TE_ES.2: Calculation of current vehicle mass after efficiency correction: M Corr,N =Eff N ·M Raw,N .

[0078] Figure 3 A flowchart of the method 100 according to the invention is shown as an alternative preferred embodiment, applicable to vehicles equipped with manual transmissions, wherein the torque 11 of the traction axle / drive axle cannot be directly obtained and must be obtained as follows: Figure 1d As shown, this is calculated from the motor torque 13. A key feature of this implementation is that the motor torque value does not generate a valid longitudinal force 10, and is activated via a valid step 190 when multiple conditions of certain auxiliary signals 50 are met: the clutch pedal is not depressed, the gear is engaged and not in neutral, and the gear ratio is stable. If braking torque is available, it can be incorporated into the longitudinal force; otherwise, the case where the brake pedal is not operated must also be included.

[0079] like Figure 3 As shown, one advantage is that, within the current valid time interval provided by the normal condition filter 140, instantaneous mass calculation or local mass calculation is performed using local regression step 152. The first raw instantaneous total mass value (MFA) can be calculated by the slope of a linear regression of the bandpass-filtered longitudinal force as a function of the bandpass-filtered longitudinal acceleration (F(A)). This slope has a mass dimension and can be used as an estimate of the vehicle's mass. However, considering the characteristic of linear regression to account for noise in the dependent variable, it is more advantageous to analyze the longitudinal acceleration as the dependent variable and the longitudinal force as the independent variable, since the signal or data of longitudinal acceleration contains more random noise than the data of longitudinal force. Therefore, in step 152, a linear regression of the longitudinal acceleration as a function of the longitudinal force (A(F)) is performed. The resulting regression coefficients are the reciprocals of the mass; by inverting them, each second raw instantaneous total mass value (MAF) can be calculated. Noise in the data and raw quality values ​​can be assessed by comparing the difference between the first raw instantaneous total mass value (MFA) and the second raw instantaneous total mass value (MAF). If the relevant noise exceeds a preset, especially appropriately selected, threshold, the MAF can be discarded.

[0080] It can be noted that step 151 is as follows: Figure 2 As shown, the mass is calculated by dividing by the local root mean square (RMS) of the force and acceleration. Step 152 is as follows: Figure 3 As shown, using local regression to calculate quality, the two steps can be completed in [the following text is incomplete and requires further context]. Figure 2 and Figure 3 The various implementations are interchangeable. By using local time data, they can better handle situations where vehicle load may change dynamically. They can also perform gear-related mass correction, i.e., step 160, as described above. Figure 2 As shown in the associated detailed explanation.

[0081] Whether Figure 2 still Figure 3 After the local quality calculation, a cumulative statistical step 180 is followed. This step is related to... Figure 8a , 8b The relevant explanations provide a detailed description.

[0082] Figure 4a and 4bThe demonstration features speed curves, motor speed (revolutions per minute (RPM)), and gear selection for vehicles equipped with manual transmissions, primarily used in training / correction driving cycles conducted on highways. Figure 4a The diagram shows the speed curve and gear position changes over time. Rectangle 201 in the diagram represents the legend of speed and gear position. Figure 4b This diagram illustrates the changes in motor speed (revolutions per minute (RPM)) and gear position over time. Rectangle 202 in the diagram represents the legend for motor speed (RPM) and gear position. The data is presented in a format covering common motor speed (RPM) ranges across all gear positions, with a maximum speed of 120 km / h.

[0083] Figure 5 The demonstration shows the use of according to Figure 2 The local root mean square (RMS) method and based on Figure 3 The results of instantaneous or local quality calculations using two local regression methods, and based on... Figure 4a , 4b The diagram shows the gear positions during the driving cycle, where rectangle 203 in the figure represents the corresponding legend, and the solid circle and dashed line represent the (reverse / inverse) regression M =

[0084] [F(A)] -1 Hollow circles and dotted lines represent (direct / forward) regression M = A(F), and x and dashed lines represent the local root mean square (RMS) splits of F and A. The quality calculation results of inverse regression are consistently lower than those of forward regression, while the results of RMS division are usually, but not always, somewhere in between. In some cases, the data showing differences between forward and inverse regression results may contain more noise, but these cases are not excluded.

[0085] Figure 6a and Figure 6b As shown in the diagram Figure 4a , 4b As shown, in the training or correction driving cycle, the learning process of transmission efficiency can be categorized into two methods: one is learning entirely based on no prior knowledge, and the other is using the already learned efficiency as an initial value and continuing to learn using exponential smoothing. Specifically, based on... Figure 2 The instantaneous root mean square (RMS) / local root mean square (RMS) quality calculation step 151 is shown or based on Figure 3 After performing a new local quality calculation in the local regression step 152 shown, according to Figure 2 and Figure 3In efficiency learning step 160, the original or initial transmission efficiency (represented by a square) and the adjusted transmission efficiency (represented by a circle) are displayed according to gear position, from gear 1 to 6 and -1 (brake gear). Figure 6a The demonstration shows the situation without any initialization (according to...). Figure 2 The learning efficiency of the TE_AN version is shown. Figure 6b This demonstrates the initialization efficiency of the final value calculated using the first scenario, and uses exponential smoothing (based on...). Figure 2 The update process for the TE_ES version is shown, where α = 0.95. The initial raw efficiency is represented by hollow or solid squares; the final adjusted efficiency is represented by solid circles; the learning process differs, but the results are consistent. Figure 9b The exponential smoothing learning with α = 0.1 can be observed; this α value effectively fixes the efficiency of the transmission device at its initial value, which is equivalent to no learning.

[0086] Figure 7a and Figure 7b Displayed in diagram form Figure 6a and Figure 6b The original vehicle mass calculation values ​​and efficiency-corrected mass calculation values ​​for the two learning scenarios, as well as the gear at the mass calculation time point. Figure 7a For version TE_AN, Figure 7b For version TE_ES, rectangles 204a and 204b in the figure represent the corresponding legend. According to the legend, the original mass calculation value M is... Raw The efficiency-corrected mass calculation value M is represented by a solid square. Corr Represented by solid circles, they are respectively based on Figure 2 and Figure 3 The inputs and outputs (data and outputs) of step 160 are shown below. Figure 2 Correspondingly, the gear delay of the mass calculation time point is represented by a hollow square, which is another input for step 160. The figure also shows the vehicle's measured mass VMM and its value by a horizontal dashed line, to which the corrected mass value will converge.

[0087] Figure 8a and Figure 8b Presented in a diagrammatic way according to Figure 2 and Figure 3 The results of performing cumulative statistics step 180 are shown (data corresponding to...). Figure 2 The estimated vehicle mass and its 95% confidence interval were obtained by analyzing data from... Figure 6a and 6bThe efficiency for both scenarios shown is obtained by averaging the corrected mass values ​​from the start to the current time and 5% of the exit time. The figure also shows the vehicle's measured mass (VMM) for performance evaluation; this scenario represents the learning phase. Figure 8a Indicates the TE_AN version. Figure 8b This indicates the TE_ES version. Rectangles 205a and 205b in the figure illustrate the corresponding legend. The calculation is the average of the efficiency-corrected quality values ​​from the start of the driving cycle to the current time point, as shown in the previous corresponding figure, i.e., μ(t), and the standard error of the average value to the current time point σ(t); displayed as an error bar, i.e., μ(t) ± 2σ(t) — considering a 95% confidence level. When the relative error calculated as the ratio 2σ(t) / μ(t) falls below an appropriately selected threshold (5% in this case), an exit signal is triggered, and the quality estimation can be terminated. This is represented in the figure as t Exit This indicates that, in any case, in Figure 8a and Figure 8b During the calculation, the computation continues until the data is exhausted. For both exit and termination, the performance is displayed as the vehicle mass estimate (VME) at that point in time. Exit =μ(t) Exit )±2σ(t Exit ), and the error VME relative to the known vehicle mass. Exit – For VMM, the error calculation at the “end” is similar. The latter error is expected to be small because this is the learning phase and the VMM is known.

[0088] Figures 9a to 9c Presented graphically, in accordance with Figure 4a , Figure 4b The data or results obtained from another driving cycle under completely different conditions of the training or correction driving cycle shown, wherein the transmission efficiency has been fixed, i.e.: city driving, significantly improved vehicle quality and no learning process. Figure 9a The graph displays speed curves and gear positions over time, with rectangle 206a representing a legend for speed and gear position. Although the vehicle's measured mass VMM is known, the learning function is disabled, and the transmission efficiency is the previously learned value. Disabling the learning function (α=1) is available in [the following text is incomplete and requires further context]. Figure 9b As seen in the image, the diagram is consistent with... Figure 6a , Figure 6b Containing the same information, the transmission efficiency (circle) is no longer adjusted, but remains constant, especially the initial value (solid square). Figure 9c and Figure 8a , Figure 8b Containing the same information (rectangle 206b in the diagram shows the corresponding legend), the difference lies in the fact that in practical applications, the learned vehicle mass is unknown. Here, it is used for performance evaluation: although at the exit time point, VME ExitApproaching VMM, but VME End –VMM greater than Figure 8a , Figure 8b The value in the figure is normal and can be mainly explained by the simple model used; for the requirements of a tire information system, this trade-off between simplicity and performance is reasonable.

Claims

1. A method (100) for determining the total mass of a motor vehicle having at least one wheel during operation, wherein, When using the relevant signals of longitudinal force (10) and longitudinal acceleration (20) present on the vehicle bus (1), especially the vehicle's CAN bus, the instantaneous longitudinal force (10) and its associated instantaneous longitudinal acceleration (20) of the vehicle are determined in the form of associated value pairs, wherein the total mass is calculated using the associated value pairs based on Newton's second law, wherein frequency filtering, especially bandpass frequency filtering, is applied to the longitudinal force and longitudinal acceleration, especially to remove the force and acceleration components caused by aerodynamic drag and / or slowly changing forces or effects of uphill and / or downhill, and wherein the mass calculation is performed only during a predetermined, especially longitudinal or linear or other suitable vehicle movement.

2. The method (100) according to claim 1, wherein, The quality calculation time interval is provided by a condition normal filter, which uses the available auxiliary signals (50) on the vehicle bus (1), especially on the vehicle CAN bus, and uses predetermined, especially appropriately selected thresholds to perform predetermined, especially predetermined small, turning or lateral movements indicated by lateral acceleration and yaw rate and / or predetermined, especially predetermined small, wheel slippage and / or constant gear or transmission ratios indicated by the relative difference in front and rear wheel speeds and / or predetermined, especially predetermined appropriate, vehicle speed and / or longitudinal acceleration (20) ranges.

3. The method (100) according to claim 2, wherein, The auxiliary signal (50) used by the normal condition filter is the same as the frequency filter used for longitudinal force (10) and longitudinal acceleration (20), especially the same low-pass filter.

4. The method (100) according to any of the preceding claims, wherein, The motor vehicle is equipped with a manual transmission, wherein the longitudinal force (10), in particular the longitudinal traction or propulsion force of the vehicle, and its associated longitudinal acceleration (20) are determined and / or used only when the longitudinal force (10) is determined to be effective, wherein the longitudinal force (10) is preferably determined to be effective when the vehicle clutch pedal is not depressed and / or is engaged and not in neutral and / or the gear ratio is stable.

5. The method (100) according to any of the preceding claims, wherein, The total mass at the current time point is calculated as the ratio of the root mean square (RMS) value of the longitudinal force to the root mean square (RMS) value of the longitudinal acceleration, wherein, preferably, these two RMS values ​​are calculated within an effective time interval generated by the normal condition filter, which is from a predetermined, particularly appropriately selected, initial time point up to the current time point, and uses values ​​derived from the bandpass filtering step (130) or values ​​obtained by exponentially smoothing the values ​​generated by the bandpass filtering step (130).

6. The method (100) according to any one of claims 2 to 5, wherein, The original instantaneous total mass value is calculated as the ratio of the original instantaneous root mean square (RMS) value of the longitudinal force (10) after bandpass filtering to the original instantaneous root mean square (RMS) value of the longitudinal acceleration (20) after bandpass filtering, wherein both original instantaneous root mean square (RMS) values ​​are calculated within each effective time interval provided by the normal condition filter.

7. The method (100) according to any one of claims 2 to 6, wherein, Within the current valid time interval provided by the normal condition filter, a first raw instantaneous total mass value (MFA) is calculated, which is in particular the linear regression slope of the bandpass-filtered longitudinal force (10) as a function of the bandpass-filtered longitudinal acceleration (20); wherein, a second raw instantaneous total mass value (MAF) is calculated, which is in particular the inverse slope of the linear regression of the bandpass-filtered longitudinal acceleration (20) as a function of the bandpass-filtered longitudinal force (10), wherein, preferably, the first raw instantaneous total mass value (MFA) and the second raw instantaneous total mass value (MAF) are used to evaluate the noise in the data of longitudinal acceleration (20) and longitudinal force (10) and / or the noise in the first raw instantaneous total mass value (MFA) and the second raw instantaneous total mass value (MAF), wherein, particularly preferably, the second raw instantaneous total mass value (MAF) is discarded if the noise exceeds a predetermined threshold.

8. The method (100) according to any one of claims 2 to 7, wherein, The normal effective interval is subject to additional constraints, which are: the minimum range of longitudinal acceleration (20) and / or the minimum number of longitudinal force (10) and longitudinal acceleration (20) samples within the interval.

9. The method (100) according to any one of claims 6 to 8, wherein, The power loss in the vehicle transmission chain associated with the gear ratio or the engaged gear is considered as a subunit efficiency factor, wherein the corrected instantaneous total mass value is calculated by multiplying the original instantaneous total mass value by the subunit efficiency factor. Preferably, the transmission efficiency is calculated based on the gear or gear ratio corresponding to the effective interval provided by the normal condition filter. Particularly preferred is that the gear or gear ratio input to these calculations for gear-related efficiency has the same delay time as that generated by the low-pass filter, in particular.

10. The method (100) according to claim 9, wherein, The transmission efficiency is calculated and learned during a predetermined, especially properly implemented, corrective driving cycle, particularly when the vehicle has been weighed and its total mass is known, using the original instantaneous total mass calculation value and the known total mass. Preferably, after the transmission efficiency learning is completed, the corrected instantaneous total mass value is calculated by multiplying the original instantaneous total mass value by the learned transmission efficiency during normal vehicle driving operation.

11. The method (100) according to claim 9 or 10, wherein, The efficiency of the transmission is considered to depend on the speed of the vehicle's motor and is approximated as a linear or quadratic polynomial.

12. The method (100) according to any one of claims 6 to 11, wherein, The power loss depending on the motor coolant temperature is also taken into consideration as a subunit efficiency factor, wherein the corrected instantaneous total mass of the vehicle is calculated by multiplying the original instantaneous total mass by the subunit efficiency factor, wherein, preferably, the power loss depending on the motor coolant temperature is learned using a loaded vehicle in multiple predetermined correction driving cycles, and particularly preferably, the correction driving cycles are executed starting from different low temperatures.

13. The method (100) according to any one of claims 5 to 12, wherein, The total mass at the current time point is calculated using statistical methods from the original instantaneous total mass value or the corrected instantaneous total mass value, especially the total mass value corrected for the efficiency of the transmission device. This method preferably starts from a predetermined, especially appropriately selected, start time point up to the current time point, especially the end time point of the calculation, or is based on the convergence time. In this case, the relative error of the total mass calculation, especially as a measure of noise in each temporary mass calculation from the start time point to the current time point, such as the ratio of the standard deviation of the original instantaneous total mass value to the total mass calculated at the current time point, is less than a preset, especially appropriately selected, threshold, especially 5%.

14. The method (100) according to any one of the preceding claims, wherein, The method (100) is used as a function in the vehicle tire information system, preferably providing useful information through the tread depth monitoring function or other functions of the tire information system.

Images