WO2007060134A1 - Procede et dispositif de determination d'un parametre d'un modele de vehicule de reference - Google Patents
Procede et dispositif de determination d'un parametre d'un modele de vehicule de reference Download PDFInfo
- Publication number
- WO2007060134A1 WO2007060134A1 PCT/EP2006/068589 EP2006068589W WO2007060134A1 WO 2007060134 A1 WO2007060134 A1 WO 2007060134A1 EP 2006068589 W EP2006068589 W EP 2006068589W WO 2007060134 A1 WO2007060134 A1 WO 2007060134A1
- Authority
- WO
- WIPO (PCT)
- Prior art keywords
- value
- vehicle
- model
- model parameter
- driving state
- Prior art date
Links
- 238000000034 method Methods 0.000 title claims abstract description 71
- 238000013528 artificial neural network Methods 0.000 claims abstract description 33
- 230000006870 function Effects 0.000 claims abstract description 33
- 230000006978 adaptation Effects 0.000 claims description 13
- 238000004422 calculation algorithm Methods 0.000 claims description 3
- 238000013459 approach Methods 0.000 claims description 2
- 238000004590 computer program Methods 0.000 claims description 2
- 230000004044 response Effects 0.000 claims 2
- 210000002569 neuron Anatomy 0.000 description 41
- 230000004913 activation Effects 0.000 description 16
- 230000001133 acceleration Effects 0.000 description 10
- 230000006399 behavior Effects 0.000 description 9
- 230000008859 change Effects 0.000 description 9
- 238000005259 measurement Methods 0.000 description 8
- 238000004364 calculation method Methods 0.000 description 7
- 238000010586 diagram Methods 0.000 description 7
- 230000008569 process Effects 0.000 description 6
- 230000008901 benefit Effects 0.000 description 5
- 238000011161 development Methods 0.000 description 5
- 230000018109 developmental process Effects 0.000 description 5
- 238000012549 training Methods 0.000 description 4
- 230000005484 gravity Effects 0.000 description 3
- 238000004519 manufacturing process Methods 0.000 description 3
- 230000003321 amplification Effects 0.000 description 2
- 238000003199 nucleic acid amplification method Methods 0.000 description 2
- 238000012360 testing method Methods 0.000 description 2
- 230000032683 aging Effects 0.000 description 1
- 230000015572 biosynthetic process Effects 0.000 description 1
- 238000013527 convolutional neural network Methods 0.000 description 1
- 230000001419 dependent effect Effects 0.000 description 1
- 238000013461 design Methods 0.000 description 1
- 238000009826 distribution Methods 0.000 description 1
- 230000007717 exclusion Effects 0.000 description 1
- 238000011478 gradient descent method Methods 0.000 description 1
- 210000002364 input neuron Anatomy 0.000 description 1
- 210000004205 output neuron Anatomy 0.000 description 1
- 230000001105 regulatory effect Effects 0.000 description 1
- 239000003381 stabilizer Substances 0.000 description 1
- 238000003860 storage Methods 0.000 description 1
- 238000005728 strengthening Methods 0.000 description 1
Classifications
-
- G—PHYSICS
- G05—CONTROLLING; REGULATING
- G05B—CONTROL OR REGULATING SYSTEMS IN GENERAL; FUNCTIONAL ELEMENTS OF SUCH SYSTEMS; MONITORING OR TESTING ARRANGEMENTS FOR SUCH SYSTEMS OR ELEMENTS
- G05B13/00—Adaptive control systems, i.e. systems automatically adjusting themselves to have a performance which is optimum according to some preassigned criterion
- G05B13/02—Adaptive control systems, i.e. systems automatically adjusting themselves to have a performance which is optimum according to some preassigned criterion electric
- G05B13/0265—Adaptive control systems, i.e. systems automatically adjusting themselves to have a performance which is optimum according to some preassigned criterion electric the criterion being a learning criterion
- G05B13/027—Adaptive control systems, i.e. systems automatically adjusting themselves to have a performance which is optimum according to some preassigned criterion electric the criterion being a learning criterion using neural networks only
-
- B—PERFORMING OPERATIONS; TRANSPORTING
- B60—VEHICLES IN GENERAL
- B60W—CONJOINT CONTROL OF VEHICLE SUB-UNITS OF DIFFERENT TYPE OR DIFFERENT FUNCTION; CONTROL SYSTEMS SPECIALLY ADAPTED FOR HYBRID VEHICLES; ROAD VEHICLE DRIVE CONTROL SYSTEMS FOR PURPOSES NOT RELATED TO THE CONTROL OF A PARTICULAR SUB-UNIT
- B60W30/00—Purposes of road vehicle drive control systems not related to the control of a particular sub-unit, e.g. of systems using conjoint control of vehicle sub-units
- B60W30/02—Control of vehicle driving stability
-
- B—PERFORMING OPERATIONS; TRANSPORTING
- B60—VEHICLES IN GENERAL
- B60W—CONJOINT CONTROL OF VEHICLE SUB-UNITS OF DIFFERENT TYPE OR DIFFERENT FUNCTION; CONTROL SYSTEMS SPECIALLY ADAPTED FOR HYBRID VEHICLES; ROAD VEHICLE DRIVE CONTROL SYSTEMS FOR PURPOSES NOT RELATED TO THE CONTROL OF A PARTICULAR SUB-UNIT
- B60W40/00—Estimation or calculation of non-directly measurable driving parameters for road vehicle drive control systems not related to the control of a particular sub unit, e.g. by using mathematical models
- B60W40/12—Estimation or calculation of non-directly measurable driving parameters for road vehicle drive control systems not related to the control of a particular sub unit, e.g. by using mathematical models related to parameters of the vehicle itself, e.g. tyre models
-
- G—PHYSICS
- G05—CONTROLLING; REGULATING
- G05B—CONTROL OR REGULATING SYSTEMS IN GENERAL; FUNCTIONAL ELEMENTS OF SUCH SYSTEMS; MONITORING OR TESTING ARRANGEMENTS FOR SUCH SYSTEMS OR ELEMENTS
- G05B17/00—Systems involving the use of models or simulators of said systems
- G05B17/02—Systems involving the use of models or simulators of said systems electric
-
- B—PERFORMING OPERATIONS; TRANSPORTING
- B60—VEHICLES IN GENERAL
- B60W—CONJOINT CONTROL OF VEHICLE SUB-UNITS OF DIFFERENT TYPE OR DIFFERENT FUNCTION; CONTROL SYSTEMS SPECIALLY ADAPTED FOR HYBRID VEHICLES; ROAD VEHICLE DRIVE CONTROL SYSTEMS FOR PURPOSES NOT RELATED TO THE CONTROL OF A PARTICULAR SUB-UNIT
- B60W50/00—Details of control systems for road vehicle drive control not related to the control of a particular sub-unit, e.g. process diagnostic or vehicle driver interfaces
- B60W2050/0001—Details of the control system
- B60W2050/0002—Automatic control, details of type of controller or control system architecture
- B60W2050/0014—Adaptive controllers
-
- B—PERFORMING OPERATIONS; TRANSPORTING
- B60—VEHICLES IN GENERAL
- B60W—CONJOINT CONTROL OF VEHICLE SUB-UNITS OF DIFFERENT TYPE OR DIFFERENT FUNCTION; CONTROL SYSTEMS SPECIALLY ADAPTED FOR HYBRID VEHICLES; ROAD VEHICLE DRIVE CONTROL SYSTEMS FOR PURPOSES NOT RELATED TO THE CONTROL OF A PARTICULAR SUB-UNIT
- B60W50/00—Details of control systems for road vehicle drive control not related to the control of a particular sub-unit, e.g. process diagnostic or vehicle driver interfaces
- B60W2050/0001—Details of the control system
- B60W2050/0019—Control system elements or transfer functions
- B60W2050/0028—Mathematical models, e.g. for simulation
-
- B—PERFORMING OPERATIONS; TRANSPORTING
- B60—VEHICLES IN GENERAL
- B60W—CONJOINT CONTROL OF VEHICLE SUB-UNITS OF DIFFERENT TYPE OR DIFFERENT FUNCTION; CONTROL SYSTEMS SPECIALLY ADAPTED FOR HYBRID VEHICLES; ROAD VEHICLE DRIVE CONTROL SYSTEMS FOR PURPOSES NOT RELATED TO THE CONTROL OF A PARTICULAR SUB-UNIT
- B60W50/00—Details of control systems for road vehicle drive control not related to the control of a particular sub-unit, e.g. process diagnostic or vehicle driver interfaces
- B60W2050/0001—Details of the control system
- B60W2050/0043—Signal treatments, identification of variables or parameters, parameter estimation or state estimation
- B60W2050/0057—Frequency analysis, spectral techniques or transforms
Definitions
- the invention relates to a method for determining a value of a model parameter of a reference vehicle model.
- the invention further relates to a device which is suitable for carrying out the method.
- Vehicle dynamics control systems such as the known ESP (Electronic Stability Program) stabilize the vehicle by taking a control deviation between an actual value of a state variable of the vehicle, which is measured by means of a vehicle sensor, and a reference value of the driving state variable by influencing the driving behavior by means of an actuator is regulated.
- the reference value of the state variable is usually determined on the basis of a vehicle model.
- the deployable vehicle models usually contain a number of parameters that have to be adapted to a specific vehicle type so that the model correctly reproduces the reference behavior of a specific vehicle.
- an extensive driving program with predetermined driving maneuvers is usually carried out, in which measurement data are recorded, which are used for the offline identification of the parameters after completion of the driving program.
- this object is achieved by a method having the features of patent claim 1 and by a device having the features of patent claim 15.
- a method of the aforementioned type in which an estimated value of the model parameter is determined several times as a function of at least one driving state variable and / or a driver predetermined variable by means of an artificial neural network, wherein the artificial neural network before a repeated determination of The estimated value is adjusted on the basis of a learning method such that the estimated value of the model parameter approximates the actual value of the model parameter. After a repeated determination, the estimated value is stored as the value of the model parameter.
- a device for determining a value of a reference model of a vehicle comprises an artificial neural network which is designed to determine an estimated value of the model parameter in dependence on at least one second driving state variable and / or a driver predetermined variable several times, wherein the artificial neural network before a repeated determination of the estimated value is adaptable such that the estimated value of the model parameter is the actual Value of the model parameter approximates.
- the estimated value can be stored as a value of the model parameter in a memory.
- an artificial neural network is used to estimate the parameters of a vehicle reference model, which is trained to determine the parameters as realistically as possible.
- the artificial neural network is adapted by means of a learning method such that the estimated value of the model parameter approaches the actual value of the parameter in a repeated calculation.
- Artificial neural networks are capable of solving and solving complex problems that are influenced by a variety of factors. An advantage of using an artificial neural network for estimating the parameters of a reference model is therefore that the values of the parameters can be determined without having to carry out a driving program with defined driving maneuvers for determining the model parameters, which determine the relevant factors in a specific way , This makes it easier to determine the parameters and in a shorter time.
- the estimated value of the model parameter is the value of a parameter in another vehicle model whereby the value of at least one third driving state variable is determined by means of the further vehicle model and wherein the adaptation of the artificial neural network is carried out as a function of the result of a comparison between the calculated value and a value of the third driving state variable determined with the aid of vehicle sensors ,
- a direct measurement of the model parameter is usually not possible, as shown at the beginning. Therefore, the actual value of the parameter can not be used for the training of the artificial neural network.
- the adaptation takes place on the basis of a comparison between the value of a driving state variable calculated on the basis of the estimated parameter value and a value determined by means of sensors.
- the reference vehicle model and the further vehicle model may generally be different from one another, wherein the further vehicle model is selected such that it includes the parameters of the reference vehicle model. Since all parameters of the further vehicle model must be determined by means of the artificial neural network in order to be able to carry out the learning process, it is advantageous if the further vehicle model has as small a parameter excess as possible with respect to the reference vehicle model.
- a further embodiment of the method and the device therefore provides that the reference vehicle model and the further vehicle model are identical. This avoids having to determine parameters of the further vehicle model that are not contained in the reference model with the artificial neural network, so that the number of parameters to be determined by means of the artificial neural network is minimized.
- a development of the method and the device includes that the learning method is a method for supervised learning.
- the supervised learning has the advantage that the training of the artificial neural network is very targeted and realizes very fast realistic estimates for the model parameters.
- An embodiment of the method and the device is characterized in that the learning method is a backpropagation method.
- Vehicle models usually contain approximations that are valid only in a certain driving state area or lead to a realistic description of the vehicle behavior.
- the training of the artificial neural network can therefore be carried out particularly effectively in the scope of the further vehicle model. In an area in which the vehicle model is not valid, it could come because of the then inaccuracy of the model to erroneous adjustments of the artificial neural network.
- An embodiment of the method and the device is therefore characterized in that the adaptation of the artificial lent neural network is made by means of the learning process only when the determined by means of the vehicle sensors value of the third driving state variable is within a predetermined limited range.
- a further embodiment of the method and the device is characterized in that the restricted area corresponds to a validity range of the further vehicle model.
- the range of validity of the further vehicle model is understood to mean the driving state area in which the model realistically describes the driving behavior of the vehicle.
- a development of the method and the device provides that deviations of the estimated value of the model parameter from a predefined initial value are determined by means of the artificial neural network.
- the estimated value of the model parameter is advantageously not calculated directly by means of the artificial neural network, but rather the artificial neural network determines deviations of the parameter value from a predefined initial value.
- the estimated value of the model Rameters is determined during normal operation of the motor vehicle.
- Another embodiment of the method and the device is characterized in that the value of the parameter changes over time.
- values of parameters that change during operation of the vehicle can always be updated.
- model parameters are parameters whose value changes as the load state of the vehicle changes or due to wear.
- An embodiment of the method and apparatus includes where the model parameter is included in the group, comprising a mass of the vehicle, moments of inertia of the vehicle, and wheel slip stiffnesses of the vehicle.
- the values of said parameters generally change during operation of the vehicle, for example due to a change in the state of loading of the vehicle or wear of the tires or in the event of a tire change.
- the changed values can be determined during normal operation of the vehicle.
- a further development of the method and the device provides that the estimated value is stored in a non-volatile memory of the vehicle and that the stored value is updated when the estimated value changes compared to the stored value.
- the use of a non-volatile memory has the advantage that the stored values are preserved when the vehicle or its ignition is switched off. When the ignition is restarted, the last stored values can be read out again. For example, if the value of a parameter changes from the stored value due to a changed load state of the vehicle, the stored value is advantageously updated.
- a method for influencing the driving state of a motor vehicle in which the driving state is influenced as a function of a deviation between an actual value of a first driving state variable and a reference value of the first driving state variable.
- the reference value is determined in a reference vehicle model containing at least one model parameter, and a value of the model parameter is determined by means of a method of the previously described type.
- a system for influencing the driving state of a motor vehicle which comprises a control device which is designed to determine a manipulated variable for driving an actuator influencing the driving state as a function of a deviation between an actual value of a first driving state variable and a reference value of the first driving state variable ,
- the reference value can be determined in a reference vehicle model containing at least one model parameter, the system comprising a device of the previously described type for determining a value of the model parameter.
- 1 is a schematic block diagram of a vehicle dynamics control system
- Fig. 2 is an illustration interconnected
- FIG. 3 is a schematic block diagram illustrating the basic structure of a neuron
- 4a is a diagram illustrating an activation function of a neuron in an embodiment
- 4b is a diagram illustrating an activation function of a neuron in another embodiment
- 5 shows a schematic representation of an artificial neural network in a first embodiment
- 6 shows a schematic representation of an artificial neural network in a second embodiment
- FIG. 7 shows a schematic representation of an artificial neural network in a third embodiment
- FIG. 8 is a schematic block diagram of a device for estimating model parameters of a vehicle model.
- FIG. 1 schematically shows a basic structure of a vehicle dynamics control for a vehicle 101 on the basis of a block diagram of the control loop.
- the vehicle may be, for example, a car or a truck.
- the controlled variable Y is generally a suitable driving state variable.
- the yaw rate ⁇ and / or the slip angle ⁇ is used as the controlled variable.
- the actual value of the control variable Y ⁇ st Y is either measured directly using a sensor of the vehicle 101 or derived from the measured values of one or more sensors.
- control deviation AY Y ref -Y ⁇ st is calculated.
- the control deviation ⁇ 7 represents the input variable of a control device 102, which calculates an output signal as a function of the control deviation.
- the control device 102 is usually activated when the control deviation .DELTA.7 and, if appropriate, further large exceed predetermined regulation entry thresholds.
- the output signals of the control device 102 correspond to an actuating request, according to the proviso of which at least one actuator 103 is controlled, with which the driving behavior of the vehicle 101 can be influenced.
- the output signals are, for example, a yaw moment requirement which is converted by means of the actuator 103.
- the actuator 103 can be, for example, a brake actuator known to the person skilled in the art, with which wheel-specific brake pressures in the wheel brakes of the vehicle 101 can be specifically set up.
- a steering actuator can be used, with which a steering torque in the steering line of the vehicle 101 can be controlled or with the driver independently a wheel steering angle to steerable wheels of the vehicle 101 can be changed.
- the steering actuator for example, be designed as a so-called superposition steering.
- to influence the driving behavior in the drive motor of the vehicle 101 or in the drive train can be intervened.
- the skilled person further actuators
- the vehicle dynamics control system may, for example, have a distribution device which determines a plurality of partial requirements from the setting request of the control device 102, according to which an actuator 103 is actuated in each case.
- the calculation of the reference value Y ref of the controlled variable Y takes place in the reference value calculation device 104 on the basis of a reference model of the vehicle 101 on the basis of variables E which indicate the driving state of the vehicle 101 desired by the driver.
- various models of the vehicle 101 such as single-track models or two-track models in a linear or non-linear embodiment may be used.
- the quantities E are, for example, the wheel steering angle set by the driver on the steerable wheels, which can be detected by means of a steering angle sensor, and the vehicle speed set by the driver, which can be determined, for example, by means of wheel speed sensors.
- the illustrated vehicle dynamics control system is known to the person skilled in the art. With regard to a yaw rate control and control interventions in the brake system and the engine control such a vehicle dynamics control system is described for example in the published patent application DE 195 15 059 Al.
- the reference vehicle model describes the behavior of the vehicle 101 based on parameters that have to be adapted to this vehicle 101. These are, in particular, geometric parameters that can be determined in a simple manner and are essentially immutable. These parameters can be determined on a prototype of the vehicle 101 and stored in a non-volatile memory 105 of the vehicle dynamics control system, which is connected to the reference value calculation device 104. In addition, however, the reference vehicle model usually also contains parameters that can only be determined in driving tests. During operation of the vehicle 101, The values of these parameters also change. Examples of such parameters are the mass of the vehicle 101, moments of inertia of the vehicle 101, or tire skews of the vehicle 101.
- KNN artificial neural network
- An ANN 500 consists of several neurons 201 and directed connections between the neurons 201. As illustrated by way of example in Figure 2 for seven neurons 201 in two layers, each connection between two neurons 20I 1 and 201 : of an ANN 500 is assigned an amplification factor a y , The gain factor a y can also be calculated as the function value of a gain function. The connection is used to weight data with the gain factor a from a neuron 20I 1 to a neuron
- the indices of the neurons 201 are also indicated in FIG.
- Each neuron 20I 1 is associated with a propagation function which relates the inputs U 1 , ..., CL N of the neuron 20I 1 to each other (block 301). With the aid of the propagation function, an input value HeI 1 is determined for a neuron 201i, which is also referred to as network input.
- the network input net L corresponds to the weighted sum of the inputs of the neuron 20I 1 .
- every neuron 20I 1 has an activation function f actl , which is applied to the network input and with which the current activation level a i of the neuron 20I 1 is determined (block 302).
- Examples of possible activation functions f act ⁇ are a step function, as shown in FIG. 4 a , and a function having a linear region and adjoining constant regions, as shown in FIG. 4 b .
- Further examples of activation functions are sigmoid functions, hyperbolic tangent functions or logistic functions.
- the position at which the activation function f act ⁇ of the neuron 20I 1 has the greatest slope is referred to as the threshold value of the neuron 20I 1 and describes the point at which the neuron 20I 1 is particularly sensitive.
- each neuron 20I 1 is assigned an output function f out ⁇ which is applied to the activation level a t to determine the output O 1 of the neuron 20I 1 (block 303).
- the output function f out ⁇ be recognized as an identity function, since the output O 1 by means of the activation function f act ⁇ can already be set sufficiently well.
- the topology of KNN 500 provides multiple layers of neurons 201, which are in particular an input layer 501 and an output layer 502.
- the neurons 201 of the input layer 501 accept the input signals / KNN 500 and relay them to the output layer 502.
- the outputs of the neurons 201 of the output layer 502 correspond to the output signals O of the KNN 500.
- An exemplary topology of an ANN 500 consisting of an input layer 501 and an output layer 502 is exemplified in FIG.
- the KNN 500 may also have one or more intermediate layers 601
- An intermediate layer 601 is also called a hidden layer and the intermediate neurons are also called hidden neurons.
- FIG. 6 shows, by way of example, a topology of the KNN 500 with two intermediate layers 601i and 60I2.
- FIG. 5 and 6 are so-called feed-forward networks in which the connections between the neurons 201 run exclusively from one layer into a subsequent layer in the direction of the output layer 502.
- FIG. 7 shows an ANN 500 with a feedback drawn by way of example.
- so-called lateral feedback can exist in which connections exist between the neurons of a layer.
- a neuron 201 can also be fed back to itself.
- a neuron 201 in a particular layer does not have to be connected exclusively to neurons 201 of a neighboring layer, but rather can also be connected to neurons 201 of other layers, as is likewise exemplified in FIG.
- the input signals / CNN 500 used to determine the estimates of the model parameters are driving state quantities measured using sensors of the vehicle 101 or determined from the measurements of vehicle sensors and / or one or more of the quantities E set by the driver of the vehicle 101.
- the driving state variables used as input signals / may be, for example, the Yaw rate, yaw acceleration, lateral acceleration, slip angle, slip angle velocity, and / or slip angles of the vehicle 101, this listing being meant to be exemplary and in no way limiting.
- the yaw rate can be measured by means of a yaw rate sensor, from whose signals the yaw acceleration can also be calculated.
- the lateral acceleration can be measured by means of a lateral acceleration sensor.
- Silt angle, slip angle velocity and slip angle can be derived from the signals from sensors such as the yaw rate sensor and / or the lateral acceleration sensor.
- an input neuron is preferably present in the input layer 501 of the KNN 500 used, which receives the corresponding input signal.
- the output signals O of the KNN 500 include the estimated values of the model parameters of the reference vehicle model, which is used within the reference value calculation device 104 of the vehicle dynamics control system of the vehicle 101 to determine the reference values Y ref of the control variable Y.
- an output neuron is preferably provided in the output layer 502, the output of which corresponds to the estimated value of the parameter.
- the KNN 500 used can basically be configured as desired.
- the propagation, activation and output functions of the neurons 201 can basically be chosen arbitrarily.
- the topology of the KNN 500 can basically be designed as desired.
- the KNN 500 used is a Elman or Jordan network known per se to the person skilled in the art.
- the KNN 500 used is part of a device for estimating the model parameters, which is illustrated schematically in FIG. 8 by means of a block diagram.
- the driver influences the driving state of the vehicle 101 by adjusting certain magnitudes E, such as the wheel steering angle of the steerable wheels of the vehicle 101 or the vehicle speed. This sets a driving state, which can be described by driving state variables.
- the values X 1 of the driving state variables as well as the variables E predefined by the driver are supplied to the KNN 500 as input signals O.
- the vehicle model may be the reference vehicle model used in the reference value calculation unit 104 or another vehicle model. In each case, however, the vehicle model should be selected such that it contains the parameters of the reference model to be determined.
- the estimated values p ⁇ determined in the KNN 500 are limited in a limiting device 802 to physically meaningful or possible values. As a result, estimates p are obtained which are physically plausible.
- Estimated values X 1 (P 1 ) of the driving state variables are then calculated in the computing device 803 on the basis of the selected vehicle model, for which purpose the parameters 80 1 are preferably also supplied to the quantities E set by the driver.
- the parameters of the vehicle model are based on the limited estimates p of the calculation.
- the estimated values X 1 (P j ) of the driving state variables thus calculated are then compared in the comparator 804 with the values X 1 of the driving state variables obtained from the measurements.
- the differences Ax 1 certainly. In other embodiments, however, the comparison can also be based on other mathematical relationships, such as quotient formation.
- the result of the comparison is supplied to an adaptation device 805, which adapts the ANN 500 as a function of the result of the comparison by means of a learning method.
- the training of KNN 500 on the basis of the learning process takes place in successive cycles, wherein in each cycle first estimates p j of the parameter p ⁇ are calculated and then an adjustment of the KNN 500 is made.
- a learning method can be used for strengthening or supervised learning.
- the adaptation of the KNN 500 is dependent on whether substantially correct estimates of the model parameters have been determined or not. The determination of essentially correct estimated values can be determined, for example, if the differences Ax 1 are smaller than predefined threshold values.
- the adaptation of the KNN 500 depends on the quality of the estimates. The quality of the estimated values is determined on the basis of a predetermined quality criterion or a quality function.
- the quality of the estimate is determined on the basis of a quality criterion based on the result of the comparison between the values determined using the sensors Values X 1 and the calculated in the calculator 804 values X 1 (P j ) of the used driving state variables is applied.
- the quality criterion can basically be chosen arbitrarily. However, it should be noted that the quality of the estimation depends on the quality criterion used.
- the amplification factors of the connection between the neurons 201, the activation, propagation or output function of one or more neurons 201 or the threshold value of one or more neurons 201 can be adapted.
- the possible learning methods there is basically no restriction.
- any learning method that is applicable to the selected KNN 500 can be used.
- a back propagation method known per se to the person skilled in the art can be used as the learning method, in which the weights between the neurons 201 of the KNN 500 are adapted by means of a gradient descent method.
- Such a method is based on an error function to be minimized which, in the case of the backpropagation method, represents the aforementioned quality criterion of the backpropagation method.
- an error function for example, the sum of the squared differences Ax 1 can be used.
- Vehicle models usually contain approximations that are only valid in a certain driving state area.
- the adaptation of the KNN 500 is therefore carried out in one embodiment only if such a driving condition exists. This will be in the activation device 806 of the system shown in FIG. 8 is determined on the basis of an activation logic, which checks in particular whether some or all of the values X 1 of the driving state variables are within a predetermined range which corresponds to the validity range of the vehicle model. If this is the case, the activation device 806 sends an activation signal to the adaptation device 805, which then makes adjustments to the KNN 500.
- the KNN 500 it is preferable to calculate new estimated values p 1 of the parameters in cycles, which become increasingly realistic due to the adaptation of the KNN 800.
- the KNN 500 learns the determination of realistic estimates p ⁇ of the model parameters. If the deviation between the values X 1 obtained from the measured values and the values x t (p) of the driving state variables calculated by the vehicle model is sufficiently small, in particular if the difference Ax 1 is smaller than a threshold value, the estimated values p become in the memory 105 of the vehicle dynamics control system stored. The determination of the estimated values P J can be continued after the storage in order to further improve the values. The stored values are updated regularly.
- the previously described method may be performed with a prototype of the vehicle 101 to determine the values of the parameters included in the selected reference vehicle model prior to the start of series production of the vehicle 101.
- the optimized estimates P J of the parameters in the production of the vehicle 101 are stored in the memory 105 of the vehicle dynamics control system.
- the parameter estimation can also be done during normal operation of the Vehicle 101 continued.
- the vehicle 101 is equipped with the illustrated parameter estimation device, which performs the parameter estimation process during the operation of the vehicle 101.
- the estimated values p ⁇ are stored in the memory 105 or those already stored Values updated. Since it is a non-volatile memory 105, such as an EEPROM, whose data is not erased when the power supply is interrupted, the last stored estimates p ⁇ can be adopted after a restart of the ignition.
- estimated values p ⁇ for the parameters of the reference vehicle model can be determined directly. However, it is also possible to determine deviations from predefined initial values. This is particularly advantageous during normal operation of the vehicle 101, since in this way the parameter values stored in the memory 105 during production can be adapted to changed conditions. As long as changed parameter values result from the parameter estimation process, the output values can be used.
- the KNN 500 the estimated values p can be determined and subsequently the deviations from the output values can be determined.
- the KNN 500 can also be configured to provide estimates of the deviation.
- a percentage exclusion can be achieved. expressed relative deviation from the initial values can be determined.
- ⁇ denotes the yaw rate of the vehicle 101, ⁇ its slip angle, v the vehicle speed and ⁇ the wheel steering angle of the steerable wheels of the vehicle 101.
- the model includes the following parameters:
- J z moment of inertia of the vehicle with respect to its vertical axis (yaw axis)
- the vehicle speed v is simplified in the single-track model also simplified as a parameter, although it is a state variable. Therefore, the value of the vehicle speed v is not estimated.
- the other parameters of the model mentioned above are of interest and therefore appreciated. It can be assumed that the center of gravity of the vehicle 101 does not change significantly, so that the two parameters l v and l h can be assumed to be constant. They therefore do not necessarily have to be estimated.
- the remaining vehicle model parameters m, J z , c v and c h can not be determined in a simple manner and can be changed by a change in the loading state of the vehicle 101, by wear of the tires or after a tire change and are therefore estimated by means of the KNN 500 ,
- the linear one-track model realistically depicts vehicle behavior in a linear driving range. If the linear one-track model is also used in the arithmetic unit 804, the adaptation of the KNN 500 is therefore carried out in this linear driving range.
- the range is characterized in that the lateral acceleration is less than about 0.4 g or 4 m / s 2 (g denotes the gravitational acceleration). Therefore, it is checked in the activation means 806 whether the value of the lateral acceleration measured by means of a lateral acceleration sensor is smaller than this value. If this is the case, the adaptation device 805 is activated.
- the parameters are always optimally adapted by the parameter estimation to any parameter changes which may be caused, for example, by a change in the loading state or by aging or wear.
- the optimum parameters are available in the vehicle dynamics control system over the entire service life of the vehicle 101.
Landscapes
- Engineering & Computer Science (AREA)
- Automation & Control Theory (AREA)
- Physics & Mathematics (AREA)
- Artificial Intelligence (AREA)
- Mechanical Engineering (AREA)
- Transportation (AREA)
- Evolutionary Computation (AREA)
- General Physics & Mathematics (AREA)
- Mathematical Physics (AREA)
- Health & Medical Sciences (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Medical Informatics (AREA)
- Software Systems (AREA)
- Control Of Driving Devices And Active Controlling Of Vehicle (AREA)
- Feedback Control In General (AREA)
Abstract
L'invention concerne un procédé de détermination d'une valeur d'un paramètre d'un modèle de référence de véhicule, avec lequel une valeur de référence d'une première grandeur d'état de roulage peut être déterminée. Le procédé est caractérisé en ce qu'une valeur estimée (formule (1)) du paramètre du modèle est déterminée au moyen d'un réseau neuronal artificiel (500) en fonction d'au moins une deuxième grandeur d'état de roulage et/ou d'une grandeur (E) prédéterminée par un conducteur. Le réseau neuronal artificiel (500) est adapté pour déterminer de manière répétée la valeur estimée (formule (1)) à l'aide d'un procédé d'apprentissage de telle sorte que la valeur estimée (formule (1)) du paramètre du modèle s'approche de la valeur effective du paramètre du modèle. Après la détermination répétée, la valeur estimée (formule (1)) est conservée en mémoire comme valeur du paramètre du modèle. L'invention concerne en outre un dispositif qui permet de mettre en oevre le procédé.
Applications Claiming Priority (4)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
DE102005055902.6 | 2005-11-22 | ||
DE102005055902 | 2005-11-22 | ||
DE102006054425.0 | 2006-11-16 | ||
DE102006054425A DE102006054425A1 (de) | 2005-11-22 | 2006-11-16 | Verfahren und Vorrichtung zum Ermitteln eines Modellparameters eines Referenzfahrzeugmodells |
Publications (1)
Publication Number | Publication Date |
---|---|
WO2007060134A1 true WO2007060134A1 (fr) | 2007-05-31 |
Family
ID=38037949
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
PCT/EP2006/068589 WO2007060134A1 (fr) | 2005-11-22 | 2006-11-16 | Procede et dispositif de determination d'un parametre d'un modele de vehicule de reference |
Country Status (2)
Country | Link |
---|---|
DE (1) | DE102006054425A1 (fr) |
WO (1) | WO2007060134A1 (fr) |
Cited By (6)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
DE102008030667A1 (de) | 2007-07-03 | 2009-01-08 | Continental Teves Ag & Co. Ohg | Verfahren und Vorrichtung zum Schätzen von Parametern |
DE102019115967A1 (de) * | 2019-06-12 | 2020-12-17 | Bayerische Motoren Werke Aktiengesellschaft | Hybrides Fahrzeugmodell basierend auf einem Einspurmodell und einem neuronalen Netz |
WO2021058223A1 (fr) | 2019-09-27 | 2021-04-01 | Bayerische Motoren Werke Aktiengesellschaft | Procédé d'application de fonctions de conduite automatisée de manière efficace et simulée |
CN113454544A (zh) * | 2019-02-19 | 2021-09-28 | 西门子股份公司 | 为至少一个机器提供模型的方法、训练***、模拟机器运行的方法及模拟*** |
CN113799772A (zh) * | 2020-09-18 | 2021-12-17 | 北京京东乾石科技有限公司 | 车辆的控制方法、装置以及控制*** |
DE102022113831A1 (de) | 2022-06-01 | 2023-12-07 | Bayerische Motoren Werke Aktiengesellschaft | Verfahren und Vorrichtung zur Parametrierung einer Fahrzeug-Komponente mittels Inferenz |
Families Citing this family (13)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
DE102008042924B4 (de) | 2008-10-17 | 2024-02-22 | Robert Bosch Gmbh | Verfahren zur Regleroptimierung in Fahrzeugregelsystemen |
SE537144C2 (sv) * | 2012-12-04 | 2015-02-17 | Scania Cv Ab | Skattning av en tröghet för ett tillstånd i ett fordon |
DE102016225772A1 (de) * | 2016-12-21 | 2018-06-21 | Audi Ag | Prädiktion von Verkehrssituationen |
CN108128309A (zh) * | 2017-09-01 | 2018-06-08 | 特百佳动力科技有限公司 | 一种车辆工况实时预测的方法 |
DE102017218703A1 (de) * | 2017-10-19 | 2019-04-25 | Continental Teves Ag & Co. Ohg | Verfahren zur Wertbestimmung von Parametern |
DE102017219673A1 (de) * | 2017-11-06 | 2019-05-09 | Robert Bosch Gmbh | Verfahren, Vorrichtung und Computerprogram zur Detektion eines Objektes |
DE102019006933A1 (de) * | 2019-10-04 | 2021-04-08 | Man Truck & Bus Se | Technik zur Modellparameteranpassung eines Dynamikmodells zur Quer- und Längsführung eines Kraftfahrzeugs |
DE102019127906A1 (de) * | 2019-10-16 | 2021-04-22 | Bayerische Motoren Werke Aktiengesellschaft | Verfahren und Vorrichtung zur Ermittlung eines Wertes eines Fahrzeugparameters |
DE102019128115B4 (de) * | 2019-10-17 | 2024-07-18 | Bayerische Motoren Werke Aktiengesellschaft | Fahrzeugmodell für Längsdynamik |
EP4040240A1 (fr) | 2021-02-08 | 2022-08-10 | Siemens Aktiengesellschaft | Modèle intégrant pour un système technique et son procédé de fourniture |
CN112949187B (zh) * | 2021-03-05 | 2022-06-14 | 株洲齿轮有限责任公司 | 整车质量计算方法 |
DE102021107458A1 (de) | 2021-03-25 | 2022-09-29 | Dr. Ing. H.C. F. Porsche Aktiengesellschaft | Steuervorrichtung und Verfahren |
DE102021206880A1 (de) | 2021-06-30 | 2023-01-05 | Robert Bosch Gesellschaft mit beschränkter Haftung | Verfahren und Vorrichtung zur optimalen Parametrisierung eines Fahrdynamikregelungssystems für Fahrzeuge |
Citations (7)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
DE19515059A1 (de) | 1994-11-25 | 1996-05-30 | Teves Gmbh Alfred | Fahrstabilitätsregler mit reibwertabhängiger Begrenzung der Referenzgierrate |
US5532929A (en) * | 1992-12-16 | 1996-07-02 | Toyota Jidosha Kabushiki Kaisha | Apparatus for controlling vehicle driving power |
EP0877309A1 (fr) * | 1997-05-07 | 1998-11-11 | Ford Global Technologies, Inc. | Capteurs virtuels pour vehicules basés sur des réseaux neuronaux entrainés à partir de données générées par des modèles de simulation |
JP2000213395A (ja) * | 1999-01-25 | 2000-08-02 | Matsushita Electric Ind Co Ltd | 空燃比制御装置 |
WO2002049900A1 (fr) * | 2000-12-20 | 2002-06-27 | Central Queensland University | Systeme et procede de prediction de la dynamique de vehicules |
EP1449743A2 (fr) * | 2003-02-19 | 2004-08-25 | Fuji Jukogyo Kabushiki Kaisha | Dispositif et procédé de génération de modèle de mouvement de véhicule |
JP2005263100A (ja) * | 2004-03-19 | 2005-09-29 | Mitsubishi Fuso Truck & Bus Corp | 車両制御装置 |
-
2006
- 2006-11-16 DE DE102006054425A patent/DE102006054425A1/de active Pending
- 2006-11-16 WO PCT/EP2006/068589 patent/WO2007060134A1/fr active Application Filing
Patent Citations (7)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US5532929A (en) * | 1992-12-16 | 1996-07-02 | Toyota Jidosha Kabushiki Kaisha | Apparatus for controlling vehicle driving power |
DE19515059A1 (de) | 1994-11-25 | 1996-05-30 | Teves Gmbh Alfred | Fahrstabilitätsregler mit reibwertabhängiger Begrenzung der Referenzgierrate |
EP0877309A1 (fr) * | 1997-05-07 | 1998-11-11 | Ford Global Technologies, Inc. | Capteurs virtuels pour vehicules basés sur des réseaux neuronaux entrainés à partir de données générées par des modèles de simulation |
JP2000213395A (ja) * | 1999-01-25 | 2000-08-02 | Matsushita Electric Ind Co Ltd | 空燃比制御装置 |
WO2002049900A1 (fr) * | 2000-12-20 | 2002-06-27 | Central Queensland University | Systeme et procede de prediction de la dynamique de vehicules |
EP1449743A2 (fr) * | 2003-02-19 | 2004-08-25 | Fuji Jukogyo Kabushiki Kaisha | Dispositif et procédé de génération de modèle de mouvement de véhicule |
JP2005263100A (ja) * | 2004-03-19 | 2005-09-29 | Mitsubishi Fuso Truck & Bus Corp | 車両制御装置 |
Cited By (7)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
DE102008030667A1 (de) | 2007-07-03 | 2009-01-08 | Continental Teves Ag & Co. Ohg | Verfahren und Vorrichtung zum Schätzen von Parametern |
CN113454544A (zh) * | 2019-02-19 | 2021-09-28 | 西门子股份公司 | 为至少一个机器提供模型的方法、训练***、模拟机器运行的方法及模拟*** |
DE102019115967A1 (de) * | 2019-06-12 | 2020-12-17 | Bayerische Motoren Werke Aktiengesellschaft | Hybrides Fahrzeugmodell basierend auf einem Einspurmodell und einem neuronalen Netz |
WO2021058223A1 (fr) | 2019-09-27 | 2021-04-01 | Bayerische Motoren Werke Aktiengesellschaft | Procédé d'application de fonctions de conduite automatisée de manière efficace et simulée |
CN113799772A (zh) * | 2020-09-18 | 2021-12-17 | 北京京东乾石科技有限公司 | 车辆的控制方法、装置以及控制*** |
CN113799772B (zh) * | 2020-09-18 | 2024-03-01 | 北京京东乾石科技有限公司 | 车辆的控制方法、装置以及控制*** |
DE102022113831A1 (de) | 2022-06-01 | 2023-12-07 | Bayerische Motoren Werke Aktiengesellschaft | Verfahren und Vorrichtung zur Parametrierung einer Fahrzeug-Komponente mittels Inferenz |
Also Published As
Publication number | Publication date |
---|---|
DE102006054425A1 (de) | 2007-05-31 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
WO2007060134A1 (fr) | Procede et dispositif de determination d'un parametre d'un modele de vehicule de reference | |
DE102005013807B4 (de) | Verfahren zum Regeln eines Fahrzeugs und integriertes Fahrzeugregelungssystem | |
DE102016218863B4 (de) | Vorrichtung und Verfahren zum Steuern eines motorgetriebenen Servolenksystems | |
EP1089901B1 (fr) | Circuit de regulation pour reguler la stabilite de roulement d'un vehicule sur la base d'un modele de reference de vehicule | |
DE102007039176B4 (de) | Verfahren zur Ermittlung des Reibwertes eines Fahrzeugreifens gegenüber dem Untergrund | |
EP2790983B1 (fr) | Dispositif de commande pour véhicule automobile, véhicule automobile et procédé de configuration du dispositif de commande | |
DE3446016A1 (de) | Verfahren zur ermittlung eines optimalen schlupfwerts | |
WO2012004331A1 (fr) | Procédé de détermination de la force exercée par la crémaillère d'un dispositif de direction d'un véhicule | |
DE102019211052A1 (de) | Verfahren zum Trainieren eines neuronalen Netzes für eine Schätzung eines Reibwertes | |
DE102008030667A1 (de) | Verfahren und Vorrichtung zum Schätzen von Parametern | |
EP4028294B1 (fr) | Dispositif et procédé d'estimation d'un coefficient de frottement d'une route | |
DE102019111041A1 (de) | Verfahren und Vorrichtung zur Ermittlung eines Schätzwertes des Schwimmwinkels eines Fahrzeugs | |
DE10304769A1 (de) | Vorhersagender Steueralgorithmus für ein Anti-Blockier-Bremssystem eines Kraftfahrzeugs | |
DE102021206880A1 (de) | Verfahren und Vorrichtung zur optimalen Parametrisierung eines Fahrdynamikregelungssystems für Fahrzeuge | |
DE102011083266B4 (de) | Untersteuerungs-Einschätzung für Fahrzeuge | |
EP2210798B1 (fr) | Procédé de réglage actif de l'inclinaison d'une roue de véhicule automobile, notamment de la chute et de la trace, et dispositif correspondant | |
DE19919180A1 (de) | Regelschaltung zum Regeln der Fahrstabilität eines Fahrzeugs | |
DE102019127906A1 (de) | Verfahren und Vorrichtung zur Ermittlung eines Wertes eines Fahrzeugparameters | |
DE102019127908A1 (de) | Verfahren und Vorrichtung zur Parametrierung einer Fahrzeugfunktion | |
DE102019128459A1 (de) | Steuervorrichtung für ein Fahrzeug | |
EP1845011B1 (fr) | Procédé destiné à l'optimisation d'un modèle à une voie | |
DE112016005072T5 (de) | Verfahren und System zur Erleichterung des Lenkens eines Fahrzeugs beim Fahren entlang einer Strasse | |
DE102020204434A1 (de) | Verfahren und Vorrichtung zum Bestimmen einer Reibungsangabe für ein Rad eines Kraftfahrzeugs | |
DE102021205561A1 (de) | Verfahren und System zur Steuerung eines Fahrzeugs | |
DE10329278A1 (de) | Stabilisierungsvorrichtung, damit ausgestattetes Fahrzeug und Stabilisierungsverfahren |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
121 | Ep: the epo has been informed by wipo that ep was designated in this application | ||
122 | Ep: pct application non-entry in european phase |
Ref document number: 06819564 Country of ref document: EP Kind code of ref document: A1 |