WO2015128318A2

WO2015128318A2 - Method and device for measuring vital data of a driver of a motor vehicle, and steering wheel for a motor vehicle

Info

Publication number: WO2015128318A2
Application number: PCT/EP2015/053810
Authority: WO
Inventors: Koshan Mahdi; Jens Binder; David Maas; Valentina Djokic
Original assignee: Takata AG
Priority date: 2014-02-27
Filing date: 2015-02-24
Publication date: 2015-09-03
Also published as: WO2015128318A3; DE102014211406A1

Abstract

The invention relates to a method and to a device for measuring vital data of a driver of a motor vehicle. According to the method, an electrodermal activity of the driver is determined by using at least one first electrode signal (111) and/or an electrocardiogram of the driver is registered by using at least one second electrode signal (121). According to the invention, movement artifacts in the first and/or the second electrode signal (111, 121) or in measured values derived from the first and/or the second electrode signal (111, 121) are identified and reduced (115) by means of a statistical method. The invention further relates to a steering wheel for a motor vehicle.

Description

Method and device for measuring vital data of a driver of a motor vehicle and steering wheel for a motor vehicle

description

The invention relates to a method for measuring vital data of a driver of a motor vehicle according to the preamble of claim 1, a device for measuring vital data of a driver of a motor vehicle according to the preamble of claim 1 and a steering wheel for a motor vehicle according to the preamble of claim 12.

Driver assistance systems are known from the prior art, which engage correctively and optimally in the driving behavior of a vehicle and / or act on a driver of the vehicle. In order for a driver assistance system to intervene to the correct extent and also at the desired time, the best possible knowledge of the current mental and / or physical condition of the driver (such as his current stress level) and, for example, other vehicle occupants is desirable. For detecting the mental and / or physical condition of the driver, different sensor systems for determining biometric data (vital data) of the driver of a motor vehicle are known, wherein information about the mental and / or physical condition of the driver is to be obtained from the biometric data. For example, sensor systems are used to detect electrodermal activity with respect to a driver's body region and / or to register an electrocardiogram. For example, US 201 1/0245643 A1 discloses a steering wheel for a motor vehicle with a sensor device for detecting biometric parameters of the motor vehicle driver. However, the recorded vital data are often subject to error, so that a reliable statement about the mental and / or physical condition of the driver is difficult. The problem to be solved by the invention is to improve the quality of the driver's acquired biometric data. The problem is solved by means of the method having the features of claim 1, the device having the features of claim 11 and by the steering wheel having the features of claim 12. Further developments of the invention are specified in the dependent claims. Thereafter, a method for measuring vital data of a driver of a motor vehicle is provided, wherein an electrodermal activity (EDA) of the driver using a first electrode signal determined and / or an electrocardiogram (ECG) of the driver with the aid of a second electrode signal is registered. According to the invention, identification and reduction (in particular elimination) of movement artifacts in the first and / or the second electrode signal or in measured values derived from the first and / or the second electrode signal takes place with the aid of a statistical method.

In particular, the first electrode signal is galvanically isolated from the second electrode signal, wherein the first and the second electrode signal are generated in accordance with galvanically separated from each other electrodes, which will be explained below. It is also conceivable to use a plurality of first electrode signals (i.e., multiple EDA signals) and / or a plurality of second electrode signals (i.e., multiple ECG signals), i. Signals from multiple EDA electrodes and / or signals from multiple ECG electrodes. In particular, the respective EDA and ECG signals are respectively evaluated together and e.g. jointly subjected to statistical motion artifact reduction.

With the aid of the electrodermal activity determined from the first electrode signal and / or the electrocardiogram obtained from the second electrode signal, biometric data such as the skin conductance, non-specific fluctuations (NSF) of the skin conductance, the heart rate and / or the respiratory rate can be determined in turn allow conclusions about the mental condition (in particular the stress level) of the driver. Processing of signals by means of statistical method improves in particular the reliability of the biometric data and thus the statements regarding the mental condition of the driver.

The determination of biometric data from the first and the second electrode signal is effected in particular by evaluating at least one electrical parameter (in particular a voltage or a current intensity) of the electrode signals, i. below the first and second electrode signals, e.g. An electrical voltage and / or current that can be tapped off at an electrode is understood. For example, the electrode signal or a measurement signal derived from the electrode signal is continuously detected and digitized, so that at least one measured value of a parameter (such as the mentioned voltage or current) of the electrode signal or a quantity derived therefrom (such as a conductance) is assigned to discrete time points. Thus, the first and the second electrode signal may comprise a set of measured values, or a set of measured values may be derived from the first or second electrode signal, the measured values being accessible to a statistical analysis for improving the measured value quality.

The improvement in the quality of the measured value is achieved at least by identifying those measured values which are greatly disturbed by a movement of the driver or which have only been caused by the movement (so-called "movement artifacts") and thus contain no information about the driver's mental or physical condition For example, when using electrodes located on a steering wheel of the vehicle, motion artifacts may arise from a movement of the driver's hand.The motion artifacts are reduced (ie, eliminated), ie, the number and / or amplitude of the measurements identified as motion artifact reduced.

For example, a set of measured values (in particular in the form of skin conductance values) are derived from the first electrode signal, the coefficient of variation being determined with respect to the measured values and, depending on the coefficient of variation, a maximum value or a minimum value being determined for identifying and reducing movement artifacts , whereby measured values whose amount exceeds the maximum value or measured values whose amount is less than the minimum value are eliminated. The coefficient of variation VarK of the measured values X is the quotient of the standard deviation of the measured values X and the expected value of the measured values X:

Standard deviation (X)

Expected value (X) expected value

/ Var (X) is the variance of the measurements.

It is conceivable, in particular, that the elimination of the measured values with the aid of the coefficient of variation and the threshold value by interpolation takes place temporally with respect to the measured values of adjacent measured values to be eliminated. By way of example, a measured value which is heavily erroneous due to a movement of the driver is replaced by a measured value which was determined by interpolation of the measured values adjacent to the faulty measured value.

According to another embodiment of the method according to the invention, an independence analysis is carried out for identifying and reducing movement artifacts in the second electrode signal (the ECG signal) with respect to measured values of at least one parameter (for example a voltage and / or a current intensity) of the second electrode signal. The independence analysis is used in particular for calculating independent components of the second electrode signal. However, methods for carrying out an independence analysis are known per se, so that it need not be discussed further here. It is also conceivable that prior to the independence analysis, a principal component analysis is carried out to structure the measured values. However, methods for carrying out such a principal component analysis are also known per se, so that a detailed representation is unnecessary. In addition, it is possible that a bandpass filtering of the first and / or the second electrode signal is performed. By way of example, low-pass filtering of the first electrode signal, in particular of the analog first electrode signal, takes place before digitization and correspondingly before the statistical one Reduction of motion artifacts. It is also possible for the first and / or the second electrode signal or measured values derived from the first and / or the second electrode signal to be subjected to further signal processing methods, eg methods of factor graph analysis and / or wavelet transformation.

As already mentioned above, a skin conductance or a measure equivalent to a skin conductance can be determined on the basis of the first electrode signal. According to one embodiment of this variant of the invention, an inverse convolution of the skin conductance or the measured variable is performed with a model function in order to determine the tonic portion of the skin conductance or the measured variable.

The model function f (t) is, for example, a bi-exponential function, for example in the following form:

Here are g, χι, τ ₂ pre-definable coefficients.

It is possible that after performing the inverse convolution by means of the model function, a smoothing of the resulting values takes place in order to reduce a statistical error of the values.

By subtracting the tonic portion of the skin conductance determined by the inverse folding, the phasic portion of the skin conductance can be determined. In turn, maxima can be determined in the phasic component and the number and / or the amplitude of the maxima can be determined as a parameter for a mental state (for example, a stress state) of the driver. In other words, the maxima in the phasic portion are interpreted as "nonspecific fluctuation", that is, as fluctuations not caused by external stimuli, and thus allow conclusions to be drawn, in particular, about the driver's mental condition.

According to another embodiment of the invention, QRS complexes are determined in the second electrode signal (the ECG signal) and the time interval of the QRS complexes is determined as a parameter for a mental state of the driver. about determining the time interval of the QRS complexes, in particular the distance of the heartbeats (interbeat interval - IBI) is determined.

The invention also relates to a device for measuring vital data of a driver of a motor vehicle, in particular for carrying out a method as described above, having an evaluation unit for determining an electrodermal activity of the driver using a first electrode signal and / or registering an electrocardiogram of the driver using a second electrode signal, wherein the evaluation unit is designed to identify and reduce movement artifacts in the first and / or the second electrode signal or in measured values derived from the first and / or the second electrode signal by means of a statistical method.

In particular, the device is part of a steering wheel of a motor vehicle, wherein the evaluation unit is integrated, for example, in the steering wheel or arranged on the steering wheel. For example, the evaluation unit is designed in the form of a suitably programmed microcontroller.

In a further aspect, the invention also relates to a steering wheel for a motor vehicle, in particular for carrying out a method as described above, with two electrode areas, the two electrode areas each having at least two electrodes for determining an electrodermal activity of the driver and at least one of these electrodes Have electrode for registering an electrocardiogram.

It is conceivable that the electrodes for determining the electrodermal activity and / or the electrode for registering the electrocardiogram extend along the circumference of a steering wheel rim of the steering wheel, i. run around a hub of the steering wheel. Of course, the electrodes do not need to extend exclusively along the steering wheel rim, but may well have sections that extend obliquely or transversely to the steering wheel rim.

According to one embodiment of this variant of the invention, at least one of the electrodes for determining the electrodermal activity and the electrode for registering the Electrocardiogram each have a portion extending along a parallel to a radius of the steering wheel oriented circumference of the steering wheel rim of the steering wheel. For example, the electrodes each have a meandering or rectangular course along the steering wheel rim, so that their main extension runs along the steering wheel rim, but they also have sections which are oriented parallel to a radius of the steering wheel.

It is also conceivable, in particular, that the two electrode areas each have at least two electrodes for registering the electrocardiogram, i. the steering wheel has in each of its electrode areas in each case two electrodes for determining the electrodermal activity and two electrodes for registering the electrocardiogram.

The invention is explained in more detail below with reference to embodiments with reference to the figures. Show it:

FIG. 1 shows a block diagram of a method for measuring vital data of a

Driver of a motor vehicle according to an embodiment of the invention; FIG. 2 shows a block diagram of a partial step of the method from FIG. 1;

FIG. 3 shows a bi-exponential model function for determining the tonic and phasic component of a skin conductance determined using the method according to the invention;

FIG. 4 shows a block diagram for calculating non-specific fluctuations of the skin conductance;

FIG. 5 shows a circuit for determining the skin conductance;

FIG. 6 shows a block diagram of a method for reducing

Motion artifacts in an electrocardiogram;

Figure 7 is a block diagram of a de-noise method for an ECG signal; 8 shows schematically a steering wheel with an electrode device according to an embodiment of the invention; Figure 9 shows a variant of an electrode surface of the steering wheel of Figure 8; and

FIG. 10 shows an alternative variant of an electrode surface of the steering wheel from FIG. 8.

In the example of the inventive method according to Figure 1, a steering wheel 1 of a motor vehicle has a first and a second electrode surface 1 1, 12, each of electrodes for generating a first electrode signal 1 1 1, with which determines an electrodermal activity (EDA) of a driver of the motor vehicle and electrodes for generating a second electrode signal 121, with which an electrocardiogram (ECG) of the driver can be recorded.

The inventive method according to the embodiment of FIG. 1 therefore includes both a determination and evaluation of the EDA and a determination and evaluation of an ECG of the driver. The processing of the EDA and ECG signals via galvanically separated signal paths (left or right path in Fig. 1), in order to minimize mutual interference between the two electrode signals 1 1 1, 121 or completely avoided.

The EDA signal (the first electrode signal 1 1 1) is first tapped by means of a voltage divider circuit (eg, as shown in FIG. 5) on the EDA electrodes of the first and / or the second electrode surface 1 1, 12, and a measurement signal 1 1 1 a, which corresponds to a skin conductance of the driver (one hand of the driver) (signal acquisition 1 12). The measurement signal 1 1 1 a is filtered by means of a low-pass filter 1 13 and digitized in an analog-to-digital converter 1 14. The digitized measurement signal 1 1 1 a comprises a set of measured values of the skin conductance recorded at different times, which are subjected to a statistical motion artifact correction.

The movement artifact correction 15 serves to identify and reduce movement artifacts in the measurement signal 1 1 1 a, ie for identifying and eliminating at least some erroneous measured values, which are at least essentially only a movement (in particular a hand or both hands of the driver), using a statistical method, as will be explained in more detail below. Subsequent to the movement artifact correction 15, a downsampling (ie, a step down by reducing the time support points) and smoothing of the signal (step 16) and, according to step 1, a calculation of the phasic portion of the skin conductance (ie, the measurement signal 1 1 1 a). The calculation of the phasic fraction is carried out in particular by an inverse convolution using a model function, as will also be explained below.

In the phasic portion of the measuring signal 1 1 1 a 1 18 maxima are determined with a peak detection method and finally determines the number of maxima per unit time and / or the amplitude of the maxima as a measure of a mental state of the driver of the motor vehicle (Determination of Non-Specific Fluctuations - NSF 1 19).

The ECG signal is tapped, amplified (step 122) and filtered as the second electrical signal 121 at the ECG electrodes of the electrode surfaces 11, 12 (step 123). The amplified and filtered second electrical signal 121 is digitized in an analog-to-digital converter (step 124) and subjected to filtering (50 Hz filtering 125). This is followed by de-icing 126 of the second electrical signal 121, which will be explained in more detail below. In the noisy signal, motion artifacts are identified and eliminated by a statistical method (step 127). Although this artifact correction is like the artifact correction 1 15 of the measurement signal 1 1 1 a based on a statistical method. However, the statistical method used with respect to the second electrical signal 121 differs from the artifact correction 15 with respect to the EDA measurement signal 11 1 1 a. In particular, for the motion artifact correction 127 with respect to the second electrical signal 121, principal component and independence analysis methods are applied, as will be explained in conjunction with FIG. In the second electrical signal 121 processed with the aid of the movement artifact correction 127, a QRS detection 128 is carried out, from which, in particular, the heart rate 129 and by determining the time interval of the QRS complexes InterBeatintervalle (IBI calculation 130) are determined. It is also conceivable that an EDR signal 131 (EDR: ECG derived respiration - ECG-based respiratory signal) is determined from the ECG signal 121. In turn, the respiratory rate 132 may be calculated from the EDR signal 131. Furthermore, medical anomalies can be extracted from the ECG signal, for example by a form analysis and corresponding classification methods. By evaluating, on the one hand, the corrected EDA measurement signal 1 1 1 a and the corrected ECG signal 121, in particular, information regarding nonspecific fluctuations - NSF of the skin conductance (from the EDA signal), the heart rate and the respiratory rate (from the ECG signal ) be won. This information, in turn, as mentioned, conclusions about the mental condition of the driver of the motor vehicle, in particular its stress level to. It is conceivable, for example, that the analysis of the EDA and ECG signal, ie in particular also a classification of the driver's mental state based on the measured vital data, takes place in an evaluation unit (eg steering wheel-integrated) and the analysis results in particular via a vehicle data bus in-vehicle Units (such as a driver assistance system) and / or external devices are provided.

It should be noted that the signal processing shown in Fig. 1, of course, is only exemplary. The invention does not necessarily require the execution of all listed there steps. Rather, it is conceivable that some of the signal processing steps explained with reference to FIG. 1 are dispensed with or that these are carried out in a modified form. For example, it is quite possible to dispense with the downsampling step 1 17. It is also possible that at least some of the steps are executed in a different order.

FIG. 2 shows in detail the sequence of the movement artifact correction 1 15 shown in FIG. 1 with respect to the measurement signal 1 1 1 a (ie the skin conductance value). As already mentioned above, the digitized measurement signal 1 1 1 a consists of a set of measured values which in each case indicate a skin conductance measured at a specific point in time or to which a skin conductance can be assigned. These measured values are subjected to a statistical method, wherein in a first step the coefficient of variation is calculated with respect to a plurality of temporally successive measured values (step 201). The calculation of the coefficient of variation has already been explained above.

This is followed by a new determination of the standard deviation (step 202) and then a threshold value formation 203 as a function of the results of steps 201, 202. The two runs of determining the standard deviation take place with different window sizes (in particular different numbers of the underlying measured values) , Thus, with the first calculation of the standard deviation (as part of step 201), very many small-scale artifacts are identified, which are combined into an artifact in the second pass (step 202). The second pass provides a kind of low-pass filtering of potential artifacts found in the first pass.

The threshold value serves to identify measurement artefacts (step 204) and to remove them from the measurement signal 1 1 1 a (the amount of the measured values). For this, e.g. First, measured values whose magnitude exceeds the specified threshold value are identified and eliminated from the measurement signal by interpolation of measured values which are to be eliminated in time. The interpolation comprises extracting time vectors associated with the motion artifacts (step 205) and interpolating a respective measurement value for points in time associated with a motion artifact (interpolation 206) and respectively replacing the erroneous measurement value with the interpolated skin conductance (correction 207). The measured value quantity corrected in this way (i.e., the corrected measuring signal 1 1 1 a) is finally subjected to the steps 1 16 - 1 19 explained with reference to FIG.

In particular, therefore, a calculation of the phasic portion of the corrected skin conductance values (according to step 17 in FIG. 1) takes place. For this purpose, the signal is subjected to a deployment operation, wherein the tonic portion of the skin conductance (the measurement signal 1 1 1 a) is determined by inverse folding of the skin conductance with a model function. An example of such a model function, which in particular models a reaction of skin pores to an external current flow, is the function already mentioned above

An example of the time course of this function is shown in FIG. Referring to Fig. 4, the tonic portion 270 of the skin conductance is determined by the inverse convolution 250 using f (t) and a smoothing 260 of the unfolded signal. Subtraction 280 of the tonic portion 270 of the skin conductance results in the phasic portion of the signal. In the phasic component, the peak value recognition 18 shown in FIG. 1 is used to identify the non-specific fluctuations 19 as a measure of the driver's stress level.

FIG. 5 shows a possible circuit for determining the skin conductance from the first electrode signal 1 1 1. Two EDA electrodes 100, 200, in particular of a steering wheel, are controlled by the hand of a driver (represented by the variable resistor 51, ie the reciprocal of the skin conductance). connected with each other. To determine the skin conductance (of the resistor 51), the circuit has a voltage divider 50 with a resistor R1. The voltage divider 50 is not designed to supply the electrodes 100, 200 with a constant current or a constant voltage. Rather, the resistor R1 is designed so that the circuit provides the greatest possible gain and thus the best possible waveform. The voltage drop across the hand (resistor 51) connected to each other electrode surfaces 100, 200 voltage is amplified in an amplifier 52 and depending on this voltage, the skin conductance or the skin conductance equivalent measure (ie, the measurement signal 1 1 1 a) determined. FIG. 6 shows the sequence of step 127 from FIG. 1, according to which motion artifacts are identified and eliminated in the ECG signal (the second electrode signal 121). The digitized ECG signal represents a set (at different times) of measured values of a parameter (in particular of a voltage) of the electrode signal 121. The reduction of the motion artifacts is carried out as well as with respect to the EDA measured values by means of statistical methods which are based on these measured values be applied. First, the ECG measurement values are subjected to bandpass filtering 301 and to a principal component analysis 302 in order to obtain statistically uncorrelated measured values. After the principal component analysis 302, an independence analysis of the statistically independent components of the measured values takes place according to the principle of blind source separation (step 303). The blind source separation 303 provides for separation of the signal into various sources 304, ie, into the actual ECG signal and separately into motion artifacts (ie, signals due solely to movement of the driver). The elimination of motion artifacts described in FIG. 6 may be preceded by de-noise according to step 126 of FIG. This Entrauschung is shown in Figure 7 in detail. The digitized first electrode signal (ECG signal) 121 is first transformed into the frequency domain (step 31 1). In the frequency domain, filtering 312 is performed and the signal is transformed back to the time domain (step 313). In the signal thus filtered, recognition of QRS complexes 314 takes place using, for example, known recognition algorithms (such as a Pan-Tompkins algorithm). An IBI calculation 315 and an analysis 316 of the noise profile in the area between adjacent QRS complexes (in particular between adjacent R waves) are performed. To determine the noise profile between adjacent QRS complexes (ie the characteristic of the noise in this range), in particular both a histogram and the power of the noise between the respective QRS complexes are determined. These two values allow determination of the noise profile. Furthermore, a further transformation into the frequency range 317 and a renewed filtering take place in analogy to the filtering 312.

Figure 8 shows schematically a steering wheel 1 according to an embodiment of the invention. The steering wheel 1 is similar to the already indicated in Figure 1 steering wheel, wherein it has correspondingly a first and second electrode surface 1 1, 12 has. The electrode surfaces 1 1, 12 each have at least two electrodes (EDA electrodes) for determining an electrodermal activity of the driver and at least one of the EDA electrodes galvanically isolated electrodes (ECG electrodes) for registering an electrocardiogram. Possible configurations of the electrodes of the two electrode surfaces 11, 12 are shown in FIGS. 9 and 10, respectively, where one of the two electrode surfaces 11, 12 is shown in each case.

According to FIG. 9, at least one of the electrode surfaces 11, 12 comprises a first and a second EDA electrode 100, 200, which are designed in particular as dry electrodes, and an ECG electrode 300, which is arranged galvanically isolated from the EDA electrodes 100, 200 is. The electrodes 100, 200, 300 are fixed in particular on a surface of the steering wheel rim. The two EDA electrodes 100, 200 are used in particular for a double EDA measurement, i. Each of the electrode surfaces 11, 12 has the two EDA electrodes 100, 200 and accordingly supplies an EDA measurement signal, which increases the measurement accuracy. The ECG electrodes 300 of the electrode surfaces 1 1, 12 together serve in particular a capacitive ECG measurement.

The electrodes 100, 200, 300 each extend mainly along a circumference (of a hub of the steering wheel 1) of the steering wheel rim of the steering wheel 1, ie the dimensions of the electrodes 100, 200, 300 are greater along the circumference of the steering wheel rim than in a direction perpendicular thereto , In particular, the first EDA electrode 100 extends in the manner of a rectangular function, having portions 110 extending in a radial direction with respect to the steering wheel. The radial sections 10 are connected to one another via connecting sections 120 running along the circumference of the steering wheel rim. The second EDA electrode 200 includes a side portion 210 extending along the rim of the steering wheel, from which a plurality of spaced-apart radially extending lands 220 protrude so as to each extend between two of the radial portions 110 of the first EDA electrode 100. The ECG electrode 300 is analogous to the second EDA electrode 200, that is to say it also has a lateral section 310 which, however, viewed in the radial direction, lies on another side of the first EDA electrode 100. From the lateral portion 310 protrude radial webs 320 (opposite to the webs 220), namely such that they too are each flanked by two radial sections 110 of the first EDA electrode 100.

FIG. 10 shows a further possible embodiment of at least one of the electrode surfaces 11, 12 of the steering wheel 1. Thereafter, the electrode surface 11, 12 has both two EDA electrodes 100, 200 and two ECG electrodes 300, 400. Thus, not only a double EDA measurement, but also a double ECG measurement can be realized in order to improve the accuracy of the ECG measurement, in particular the reduction of motion artifacts. The electrodes 100, 200, 300, 400 are galvanically isolated from each other, i. The two ECG signal paths are also duplicated and galvanically isolated from each other. The galvanic separation of the electrodes 100, 200, 300, 400 from each other is realized in particular by a distance between the electrodes 100, 200, 300, 400. The electrodes 100, 200, 300, 400 each have a sinusoidally extending side 165, 265, 365, 465 and a straight side 155, 255, 355, 455 extending at least substantially along the circumference of the steering wheel rim, the straight sides 255, 455 of the second EDA electrode 200 and the second ECG electrode 400 face each other. Correspondingly, the straight sides 155, 355 of the first EDA electrode 100 and the first ECG electrode 300 face away from each other, so that in each case a belly of the sinusoidal side 165 of the first EDA electrode 100 merges into a valley of the sinusoidal side 465 of the second ECG Electrode 400 and a belly of the sinusoidal side 365 of the second EDA electrode 300 extends into a valley of the sinusoidal side 265 of the first ECG electrode 200.

It should be noted that the electrode surfaces 1 1, 12 may be configured identically. However, this is not mandatory, but it is quite conceivable, for example, that the shape of the electrode surfaces 1 1, 12 and in particular the shape and configuration of the electrodes of the electrode surfaces 1 1, 12 is different. LIST OF REFERENCE NUMBERS

1 steering wheel

1 1 first electrode surface

12 second electrode surface

50 voltage divider

51 hand

52 amplifiers

100, 200 EDA electrode

1 10 radial section

1 1 1 first electrode signal

1 1 1 a Measurement signal

1 12 signal acquisition

1 13 low-pass filter

1 14, 124 analog-to-digital converter

1 15 Motion artifact correction

1 16-1 17 Process step

1 18 Peak detection

1 19 non-specific fluctuation

120 connecting section

121 first electrode signal

122 reinforcement

125 filtering

126 detoxification

127 Motion artifact correction

128 QRS detection

129 heart rate

130 Interbeat interval

131 EDR signal

132 respiratory rate

155, 255, 355, 455 straight side electrode

165, 265, 365, 465 sinusoidal side electrode 201 calculation coefficient of variation Calculation standard deviation

thresholding

artifact identification

Zeitvektorextrahierung

interpolation

correction

, 310 lateral section

, 320 footbridge

inverse folding

smoothing

tonic portion

subtraction

, 400 ECG electrode

Bandpass filtering

Principal component analysis blind source separation

source

, 317 Transformation in frequency domain

filtering

Transformation in time domain

Detection of QRS complexes

IBI calculation

Noise profile analysis

Claims

Patent claims

1 . Method for measuring vital data of a driver of a motor vehicle, wherein an electrodermal activity of the driver is determined using at least a first electrode signal (1 1 1 ) and / or an electrocardiogram of the driver is registered using at least a second electrode signal (121), characterized by

Identifying and reducing (1 15) movement artifacts in the first and/or the second electrode signal (1 1 1, 121) or in measured values derived from the first and/or the second electrode signal (1 1 1, 121) using a statistical method .

2. The method according to claim 1, characterized in that a set of measured values is derived from the first electrode signal (1 1 1), wherein for

Identifying and reducing motion artifacts in these measured values determines the coefficient of variation in relation to the measured values and, depending on the coefficient of variation, a maximum value or a minimum value is set, with measured values whose amount exceeds the maximum value or measured values whose amount falls below the minimum value being eliminated .

3. The method according to claim 2, characterized in that the elimination of the measured values is carried out by interpolation of measured values that are temporally adjacent to the measured values to be eliminated.

4. The method according to any one of the preceding claims, characterized in that an independence analysis is carried out to identify and reduce movement artifacts in the second electrode signal (121) with respect to measured values of at least one parameter of the second electrode signal.

Method according to claim 4, characterized in that before the

Independence analysis, a principal component analysis of the measured values is carried out.

Method according to one of the preceding claims, characterized in that a bandpass filtering of the first and/or the second electrode signal (1 1 1, 121) is carried out.

Method according to one of the preceding claims, characterized in that a skin conductance or a measurement variable (1 1 1 a) equivalent to a skin conductance is determined based on the first electrode signal (1 1 1, 121) and an inverse fold (250) of the skin conductance or the measurement variable (1 1 1 a) is carried out with a model function in order to determine the tonic component of the skin conductance or the measured variable (1 1 1 a).

Method according to claim 7, characterized in that the phasic component of the skin conductance or the measured variable (1 1 1 a) is determined by means of the tonic component.

Method according to claim 8, characterized in that maxima are determined in the phasic portion of the skin conductance or the measured variable (1 1 1 a) and the number and/or the amplitude of the maxima are determined as a characteristic value for a mental state of the driver.

0. Method according to one of the preceding claims, characterized in that QRS complexes are determined in the second electrode signal (121) and the time interval of the QRS complexes is determined as a characteristic value for a mental state of the driver.

1 . Device for measuring vital data of a driver of a motor vehicle, in particular for carrying out a method according to one of the preceding claims, with an evaluation unit for determining an electrodermal activity of the driver using at least a first electrode signal (1 1 1) and / or for registering an electrocardiogram of the Driver using at least a second electrode signal (121), characterized in that the evaluation unit is designed to identify and reduce movement artifacts in the first and/or the second electrode signal (1 1 1 , 121) or in measured values derived from the first and/or the second electrode signal (1 1 1 , 121) using a statistical method is.

12. Steering wheel for a motor vehicle, in particular for carrying out the method according to one of claims 1 to 10

- two electrode areas (1 1, 12), characterized in that the two electrode areas (1 1, 12) each have at least two electrodes (100, 200) for determining an electrodermal activity of the driver and at least one electrode (300) which is galvanically isolated from these electrodes ) to record an electrocardiogram.

13. Steering wheel according to claim 12, characterized in that the electrodes (100, 200) for determining the electrodermal activity and / or the electrode (300) for registering the electrocardiogram extend at least in sections along the circumference of a steering wheel rim of the steering wheel (1).

14. Steering wheel according to claim 12 or 13, characterized in that at least one of the electrodes (100, 200) for determining the electrodermal activity and the electrode (300) for registering the electrocardiogram each have a section (1 10, 220, 320), which extends along a parallel to a radius of the

Steering wheel (1) oriented circumference of the steering wheel rim of the steering wheel (1).

15. Steering wheel according to one of claims 12 to 14, characterized in that the two electrode areas (1 1, 12) each have at least two electrodes (300, 400) for registering the electrocardiogram.