METHOD AND SIGNAL PROCESSING UNIT FOR DETERMINING THE RESPIRATORY ACTIVITY OF A PATIENT

Abstract

Process/unit for determining intrinsic breathing activity of a ventilated patient. The process/unit carries out a first ventilating operation, in which a ventilator parameter at a first setting. The process/unit generates a first set of signal values as a function of measured values, which were measured at the first setting. A first breathing activity value is derived using a predefined lung mechanical model and the first set of signal values. The process/unit calculates a value for the reliability that the first breathing activity value agrees with the corresponding actual breathing activity value. Depending on this reliability assessment, the process/unit checks whether a predefined triggering criterion is met. If this criterion is met, then the process/unit triggers a change step, in which the ventilator parameter is set at a second setting. It carries out an additional ventilating operation, in which the ventilator parameter is set at the second setting.

Claims

1. A computer-implemented process for determining a breathing activity indicator, which indicator correlates with intrinsic breathing activity of a patient, wherein the process comprises the steps of: providing a ventilator configured to mechanically ventilate the patient at least temporarily and being operable depending on a first variable ventilator parameter, wherein the first ventilator parameter has an effect on control of a flow of a gas to the patient and/or from the patient and/or of a pressure of this gas; providing a predefined lung mechanical model, which model describes at least one relationship between the breathing activity indicator and at least one measurable signal providing a signal processing unit configured to carry out a first and a second ventilating operation, while the first ventilator parameter is set to a respective set point, wherein each one of the ventilating operations at the respective set point comprises the steps that the signal processing unit receives a measured value, per measurable signal occurring in the lung mechanical model, wherein the value is measured while the first ventilator parameter is set to the respective set point, generates at least one set of signal values with a respective signal value per measurable signal of the lung mechanical model using values measured at the respective set point, derives at least one breathing activity value for the breathing activity indicator, uses for deriving the breathing activity value the lung mechanical model and the set of signal values being generated with values measured at the respective set point and, controls the ventilator with a control goal that the ventilator assists the intrinsic breathing activity of the patient, wherein the first ventilator parameter is set to the set point, the method comprising the further steps that the signal processing unit carries out the first ventilating operation, in which the first ventilator parameter is set to a first set point, derives a first breathing activity value during the first ventilating operation, calculates a reliability assessment for a reliability that the first breathing activity value agrees with a corresponding actual value of the breathing activity indicator of the patient, and checks whether a predefined triggering criterion is met, wherein the triggering criterion depends on the calculated reliability assessment for the step of deriving the first breathing activity value, and wherein the triggering criterion is met at least if the calculated reliability assessment is below a predefined first reliability threshold for the derivation of the first breathing activity value, and as a response to the detection that the triggering criterion is met, the signal processing unit triggers a change step, in which the first ventilator parameter is set to a second set point, which differs from the first set point, and carries out the second ventilating operation, in which the first ventilator parameter is set to the second set point instead of to the first set point.

2. A process in accordance with claim 1, wherein: the signal processing unit derives a second breathing activity value during the second ventilating operation, which second operation is carried out at the second set point, the signal processing unit uses at least one second set of signal values, which has been generated using measured values which have been measured at the second set point for the deriving the second breathing activity value, and additionally uses the lung mechanical model for the derivation.

3. A process in accordance with claim 2, wherein: during the second ventilating operation carried out with the second set point the signal processing unit uses for the derivation of the second breathing activity value the set of signal values generated by using values measured at the first set point and before the change step in addition to using the set of signal values, which has been generated using values measured at the second set point.

4. A process in accordance with claim 1, wherein: a parameter for the feed of gas to the patient is used as the first ventilator parameter, and the signal processing unit triggers the step of reducing or increasing the feed of gas to the patient, during the change step, and then triggers an additional change step in order to increase the feed of gas to the patient again or in order to reduce it again.

5. A process in accordance with claim 1, wherein: the step that the signal processing unit controls the ventilator during the first ventilating operation comprises the step that the signal processing unit lets the first ventilator parameter be set at the first set point during the first ventilating operation as long as the triggering criterion is not met and controls the ventilator as a function of the first breathing activity value derived at the first set point wherein the control goal is to assist the intrinsic breathing activity of the patient.

6. A process in accordance with claim 1, wherein: the step that the signal processing unit controls the ventilator comprises the step that the signal processing unit controls the ventilator after the change step as a function of a signal for the flow rate and/or for the pressure in a circuit of gas between the ventilator and the patient wherein the control goal is to assist the intrinsic breathing activity of the patient wherein the control as a function of the signal is performed at least if the calculated reliability assessment is below the first reliability threshold or below a second, lower reliability threshold.

7. A process in accordance with claim 1, wherein: the signal processing unit controls the ventilator with a control goal that the flow of gas to the patient and/or from the patient, which flow is brought about by the ventilator, is synchronized with intrinsic breathing activity of the patient, the signal processing unit repeatedly carries out a ventilating operation during the control in order to achieve the control goal, and the signal processing unit carries out the steps of calculating the respective reliability assessment, of triggering a change step for the first ventilator parameter if the calculated reliability assessment is below the first reliability threshold and afterwards of carrying out an additional ventilating operation with the changed set point.

8. A process in accordance with claim 1, wherein: if the triggering criterion is not met, the signal processing unit carries out at least one additional ventilating operation, in which the first ventilator parameter remains at the first set point, the signal processing unit generates an additional set of signal values with one value per signal occurring in the lung mechanical model, and the signal processing unit derives an additional breathing activity value using the additional set of signal values and calculates an assessment for the reliability of the derivation thereof.

9. A process in accordance with claim 1, wherein: the first ventilating operation comprises the steps that the signal processing unit receives for each signal occurring in the lung mechanical model at least two respective measured values, which values have been measured at the first set point, using the received measured values, generates a plurality of sets of signal values wherein every value set comprises a respective signal value per measurable signal, derives the first breathing activity value using at least two of the plurality of sets of signal values generated up to now at the first set point, and calculates the reliability assessment for deriving the breathing activity value depending on the sets of signal values used for the derivation.

10. A process in accordance with claim 1, wherein: if the reliability assessment for deriving the first breathing activity value meets the triggering criterion, the signal processing unit triggers the change step such that the second set point depends on the calculated reliability assessment.

11. A process in accordance with claim 1, wherein: the predefined lung mechanical model has a first model parameter being variable over time, wherein in the step of deriving a breathing activity value, the signal processing unit by using the set of signal values, which has been generated at the respective set point, derives a value for the first model parameter of the predefined lung mechanical model and derives the breathing activity value using the first model parameter value and the lung mechanical model, wherein the signal processing unit, in the step of calculating the reliability assessment for deriving the breathing activity value calculates an assessment for the reliability of the derivation of the first model parameter value.

12. A process in accordance with claim 11, wherein: the lung mechanical model has a first model parameter and a second model parameter (E, k.sub.eff, P0), wherein the signal processing unit, calculates a reliability assessment for the derivation of the first model parameter value as a first reliability assessment and a reliability assessment for the derivation of the second model parameter value as a second reliability assessment, triggers a first change step if the first reliability assessment meets the triggering criterion, and triggers a second change step when the second reliability value meets the triggering criterion, wherein the first change step pertains to the first ventilator parameter and the second step process pertains to a different ventilator parameter, and/or wherein the first change step leads to a different set point than the second change step.

13. A process in accordance with claim 1, wherein: for deriving the respective breathing activity value, the signal processing unit applies in at least one ventilating operation the lung mechanical model to at least one first set of signal values and to at least one second set of signal values, wherein the measured values of the first set of signal values have been measured at the first set point of the first ventilator parameter, wherein the measured values of the second set of signal values have been generated at the second set point or at an additional set point which differs from the first set point, wherein the signal processing unit calculates a respective weighting factor for each set of signal values used for the derivation, and wherein the signal processing unit uses the weighing factors for deriving the breathing activity value.

14. A process in accordance with claim 13, wherein: the signal processing unit calculates the weighting factors such that the smaller the number of sets of signal values, which are used for the derivation of the breathing activity value have been generated at a respective set point, the higher is the weighting factor for the set of signal values generated at the respective set point, and/or the sum of the weighting factors for the sets of signal values, which have been measured at a respective set point, is equal to a predefined share value, and/or each weighting factor depends on in which phase in the course of a sequence comprising at least one breath of the patient this set of signal values has been generated.

15. A process in accordance with claim 13, wherein: in at least one ventilating operation, the signal processing unit calculates for at least two different set points used up to now a respective breathing activity single value and uses for this calculation at least one set of signal values which has been generated using measured values which have been measured at the respective set point used during this ventilating operation and combines the breathing activity single values using the weighting factors into a breathing activity value.

16. A process in accordance with claim 1, wherein: the breathing activity indicator can be measured at the second set point, wherein the signal processing unit determines the second breathing activity value at the second set point and for determining the value, generates a signal value for the breathing activity indicator by at least one measurement at the second set point, and wherein in the step of calculating the assessment for the reliability of the derivation of the first breathing activity value, the signal processing unit compares the derived first breathing activity value with the determined second breathing activity value.

17. A signal processing unit for determination of a breathing activity indicator, which indicator correlates with an intrinsic breathing activity of a patient, wherein the signal processing unit is connected to or configured to be connected to a ventilator at least temporarily, wherein the ventilator is configured to mechanically ventilate the patient at least temporarily and to be operated depending on a first variable ventilator parameter, the first ventilator parameter having an effect on control of a flow of a gas to the patient and/or from the patient and/or of a pressure of the gas, the signal processing unit is configured to have reading access to a memory at least temporarily, the memory storing or being configured to store a lung mechanical model, which model describes at least one relationship between the breathing activity indicator and at least one measurable signal, and wherein the signal processing unit is configured to carry out a first and a second ventilating operation, while the first ventilator parameter is set to a respective set point, wherein in each ventilating operations the signal processing unit is configured to receive a measured value, per measurable signal occurring in the lung mechanical model, wherein the value is measured while the first ventilator parameter is set to the respective set point, and to generate at least one set of signal values with a respective signal value per measurable signal of the lung mechanical model using values measured at the respective set point, to derive at least one breathing activity value for the breathing activity indicator value for deriving the breathing activity value to use the lung mechanical model and the set of signal values generated at the respective set point, and to control the ventilator with a control goal that the ventilator assists the intrinsic breathing activity of the patient, wherein the first ventilator parameter is set at the respective set point, wherein the signal processing unit is configured, to carry out the first ventilating operation, in which the first ventilator parameter is set to a first set point, to derive a first breathing activity value during the first ventilating operation and to calculate a reliability assessment for the reliability that the derived first breathing activity value agrees with the corresponding actual value of the breathing activity indicator of the patient, and wherein the signal processing unit is configured to check whether a predefined triggering criterion is met, which depends on the calculated reliability assessment for the derivation of the first breathing activity value, wherein the triggering criterion is met at least when the calculated reliability assessment for the derivation of the first breathing activity value is below a predefined first reliability threshold, and wherein the signal processing unit is configured, as a response to the detection that the triggering criterion is met, to trigger a change step, in which the first ventilator parameter is set to a second set point, which differs from the first set point, and to carry out the second ventilating operation, in which the first ventilator parameter is set to the second set point instead of to the first set point.

18. A process in accordance with claim 1, wherein a plurality of the steps of claim 1 are performed by a computer program, which program can be executed on the signal processing unit, the execution causing the signal processing unit to execute the plurality of the steps.

19. A process in accordance with claim 1, wherein a plurality of the steps of claim 1 are provided by a signal sequence, comprising commands, which can be executed on the signal processing unit, the execution causing the signal processing unit to execute the plurality of the steps.

20. A process for determining a breathing activity indicator which is representative of an intrinsic breathing activity of a patient, the process comprising the steps of: providing a ventilator configured to mechanically ventilate the patient and to operate as a function of a variable ventilator parameter, the ventilator parameter being configured to effect control of a flow of gas to, or from, the patient, and/or of a pressure of the gas; providing a predefined lung mechanical model which model describes a relationship between the breathing activity indicator and a measurable signal; providing a signal processing unit configured to operate the ventilator to perform a ventilating operation with the ventilator parameter being set to a set point, wherein the ventilating operation at the set point comprises that the signal processing unit: receives one measured value which has been measured for the measurable signal while the ventilator parameter is set to the set point, derives at least one breathing activity value for the breathing indicator, uses the lung mechanical model and the set of signal values at the set point for the derivation of the breathing activity value for the breathing activity indicator, controls the ventilator to have the ventilator assist intrinsic breathing activity of the patient, when the ventilator parameter is set to the set point, wherein the signal processing unit, carries out at least one first ventilating operation, in which the ventilator parameter is set to a first set point, derives a first breathing activity value during the first ventilating operation, and calculates a value for the reliability that the first breathing activity value agrees with the corresponding actual value of the breathing activity of the patient, and wherein the process comprises the additional steps that the signal processing unit checks whether a predefined triggering criterion is met, wherein the triggering criterion depends on the calculated reliability assessment for the derivation of the first breathing activity value, and wherein the triggering criterion is met at least when the calculated reliability assessment is below a predefined first reliability threshold for the derivation of the first breathing activity value, and as a response to the detection that the triggering criterion is met, the signal processing unit, triggers a change step, in which the ventilator parameter is set to a second set point which is different from the first set point, and carries out a second ventilating operation, in which the ventilator parameter is set to the second set point instead of at the first set point.

Description

BRIEF DESCRIPTION OF THE DRAWINGS

[0139] In the drawings:

[0140] FIG. 1 schematically shows which sensors measure which different signals, which are used for the derivation of the intrinsic breathing activity of the patient;

[0141] FIG. 2 shows which signals are derived from the measured values of which sensors;

[0142] FIG. 3 shows an exemplary weighting function, with which a plurality of sets of signal values are weighted;

[0143] FIG. 4 shows an exemplary weighting of sets of signal values based on the frequency of signal values;

[0144] FIG. 5 shows a first part of a flow chart: Derivation of a breathing activity value and decision whether the predefined triggering criterion is met;

[0145] FIG. 6 shows a second part of the flow chart: Regular operation with sufficiently reliable breathing activity value;

[0146] FIG. 7 shows a third part of the flow chart: Carrying out of an easier maneuver;

[0147] FIG. 8 shows a fourth part of the flow chart: Carrying out of a more serious maneuver;

[0148] FIG. 9 shows a fifth part of the flow chart: Derivation of model parameter values based on sets of signal values, which have been generated during a maneuver;

[0149] FIG. 10 shows a sixth part of the flow chart: Derivation of a breathing activity value during the maneuver, calculation of the reliability of the derivation thereof; and

[0150] FIG. 11 shows a seventh part of the flow chart: Decision on how the mechanical ventilation will be continued after a maneuver.

DESCRIPTION OF PREFERRED EMBODIMENTS

[0151] Referring to the drawings, in the exemplary embodiment, a patient P is ventilated mechanically by a ventilator 1 at least from time to time. The mechanical ventilation shall be carried out in a manner synchronized with the intrinsic breathing activity of the patient P. The ventilator 1 is regulated as a function of the intrinsic breathing activity of the patient P.

[0152] The ventilator 1 in one embodiment operates in a pressure-controlled manner. In the control, the reference variable is, in this case, a required time curve of the pneumatic pressure of the breathing, preferably in the airway of the patient P. The manipulated (controlled) variable is then the pneumatic pressure, which the mechanical ventilation achieves. This desired curve of the pressure shall be synchronized with the pressure which is variable over time and which the intrinsic breathing activity of the patient P achieves, and the desired curve therefore depends on the intrinsic breathing activity. In another embodiment, the reference variable in the control is a required time curve of the volume, i.e., of the fill level of the lungs of the patient P. The manipulated variable is the flow of breathing air into the lungs and out of the lungs, which is achieved by the mechanical ventilation. In this embodiment as well, the desired curve of the volume is to be synchronized with the intrinsic breathing activity of the patient P.

[0153] For synchronizing, it is necessary in both types of control to determine a preferably pneumatic value for the intrinsic breathing activity of the patient P, for example, the Pressure indicator P.sub.mus, which correlates with the pressure that the respiratory muscles of the patient P generate. The breathing activity pressure indicator P.sub.mus, which is variable over time and is preferably pneumatic, cannot be measured directly during the mechanical ventilation, but rather is determined at each scanning time t.sub.i by [0154] a plurality of variable values appearing in the ventilation circuit being measured, [0155] a set of signal values being generated from a respective measured value per measurable signal, and [0156] a value for the preferably pneumatic breathing activity indicator P.sub.mus, i.e., an estimated breathing activity value P.sub.mus,est(t.sub.i), being repeatedly derived from at least one generated set of signal values, preferably from a plurality of sets of signal values.

[0157] In case of a proportional control of the ventilator 1, the pressure P.sub.art(t.sub.i) generated by the ventilator 1 is ideally proportional to the estimated breathing activity value P.sub.mus,est(t.sub.i) at each scanning time t.sub.i, i.e.,

P.sub.art(t.sub.i)=x*P.sub.mus,est(t.sub.i),  (1)

wherein P.sub.mus,est(t.sub.i) is an estimated breathing activity value and x is a predefined proportionality to factor. This proportionality factor x is also designated as degree of assistance by the ventilator 1. In an ideal synchronization, P.sub.art(t.sub.i)=x*P.sub.mus,est(t.sub.i).

[0158] A data-processing signal processing unit carries out the just described control at an upper level, for example, the pressure-controlled or the volume-controlled regulation, and uses for this estimated values P.sub.mus,est(t.sub.i) for the breathing activity value, wherein the values P.sub.mus,est(t.sub.i) are derived using sets of signal values. The signal processing unit calculates in the upper-level control values for the pressure and/or volume flow currently to be generated by the ventilator. The signal processing unit carries out, furthermore, a control at a lower level in order to derive from the required values for the pressure to be generated actuating actions for adjusting elements of the ventilator 1, and these adjusting elements bring about the mechanical ventilation of the patient P.

[0159] FIG. 1 schematically shows which sensors measure the intrinsic breathing activity and the mechanical ventilation of the patient P. Shown are [0160] the patient P, [0161] the esophagus Sp and the diaphragm Zw of the patient P, [0162] a ventilator 1, which mechanically ventilates the patient P at least from time to time and which comprises a data-processing signal processing unit 5, [0163] a memory 9, to which the signal processing unit 5 has reading access at least from time to time and in which a computer-available lung mechanical model 20 is stored, [0164] four sets 2.1.1 through 2.2.2 of sensors with at least one respective measuring electrode, wherein the sets of measuring electrodes 2.1.1 and 2.1.2 are arranged parallel to the sternum and the sets of measuring electrodes 2.2.1 and 2.2.2 are arranged at the costal arch, [0165] a pneumatic sensor 3, which measures the airway pressure P.sub.aw in front of the mouth of the patient P as well as the volume flow Vol′ of breathing air into the lungs and out of the lungs of the patient P, [0166] an optional optical sensor 4, which comprises an image recording device and an image analysis unit and is directed towards the chest area of the patient P, and [0167] an optional pneumatic sensor 6 in the form of a probe or of a balloon in the esophagus Sp and close to the diaphragm Zw of the patient P, which measures a pressure P.sub.es in the esophagus Sp.

[0168] The measuring electrodes 2.1.1 through 2.2.2 as well as an electrode, not shown, for electrical ground make possible a non-invasive electromyography measurement (EMG measurement). It is also possible to position sensors at the body of the patient P and as close as possible to the signal source, which make possible a mechanomyogram measurement (MMG measurement).

[0169] FIG. 2 shows which signals are derived from the measured values of which sensors. These signals and possible sources for measurement errors are explained below.

[0170] The four sets of measuring electrodes 2.1.1 through 2.2.2 of measuring electrodes and the electrode for ground yield measured values. These measured values are processed, and the processing yields overall at least one electrical signal, which correlates with electrical pulses, which are generated in the body of the patient P. Some of these electrical pulses bring about that the respiratory muscles of the patient P contract and consequently bring about the movement of breathing air into the lungs and out of the lungs. The electrically stimulated respiratory muscles bring about a pressure, which correlates with the sought pneumatic indicator P.sub.mus for the intrinsic breathing activity. Electrical pulses in addition to these electrical pulses bring about that the heart of the patient P beats.

[0171] The measured values of the four sets of measuring electrodes 2.1.1 through 2.2.2 are thus processed and yield an overall electrical signal, which results from a superimposition of a respiratory and of a cardiogenic signal, after processing. The respiratory signal is sought. The effect of the cardiogenic signal on the overall electrical signal is compensated by calculation as far as possible, for example, by applying a method, which is described in DE 10 2015 015 296 A1, in DE 10 2007 062 214 B3 or in M. Ungureanu and W. M. Wolf: “Basic Aspects Concerning the Event-Synchronous Interference Canceller,” IEEE Transactions on Biomedical Engineering, Vol. 53, No. 11 (2006), pp. 2240-2247. This compensation by calculation yields an electrical respiratory signal Sig, which varies with time. This electrical respiratory signal Sig has been obtained close to the signal source, i.e., close to the respiratory muscles, and correlates with the electrical pulses, which move the respiratory muscles of the patient P, and thus with the pneumatic breathing activity indicator P.sub.mus.

[0172] Even after the processing and compensation by calculation, the electrical respiratory signal Sig can still be superimposed by unwanted signals that are caused, for example, by electrochemical effects on the contact surface between the skin of the patient P and a measuring electrode 2.1.1 through 2.2.2. In addition, the patient P may change his posture during the measurement, and the effect of the cardiogenic signal will not be able to be compensated completely or not correctly by calculation.

[0173] The pneumatic sensor 3 measures measured values, which are caused by a superimposition of the intrinsic breathing activity of the patient P and the mechanical ventilation. These measured values are caused exclusively by the intrinsic breathing activity only when the mechanical ventilation is interrupted. The airway pressure P.sub.aw and the volume flow Vol′, i.e., the flow per time unit of breathing air into the lungs and out of the lungs of the patient P, are derived from these measured values.

[0174] The intrinsic breathing activity of the patient P is affected by lung mechanical parameters. Values for the lung mechanical parameters and the volume flow cannot be determined approximately at the same time solely by a signal pneumatic sensor. In addition, the pneumatic sensor 3 is not arranged directly or at all in the mouth of the patient P, but rather is arranged at a spaced location from the patient P in the ventilator or at the ventilator 1, especially in order to meet hygienic requirements in a hospital. Therefore, a transmission channel occurs between the airway of the patient P and the pneumatic sensor 3, which transmission channel especially comprises the hose between the patient P and the ventilator 1 as well as the mouthpiece in the mouth of the patient P. Hence, a time delay occurs between the generation of a pressure in the body of the patient P and the time of a measured value of the pneumatic sensor 3, which measured value was caused by this pressure. For these two reasons, namely lack of observability and time delay, the mechanical ventilation cannot, as a rule, be ideally synchronized with the intrinsic breathing activity of the patient P solely on the basis of measured values of the pneumatic sensor 3.

[0175] The optical sensor 4 is capable of determining the geometry of the body of the patient P by means of image processing, and this determined body geometry correlates with the current filling level Vol of the lungs, but also depends on additional parameters. Therefore, the optical sensor 4 can alone, as a rule, measure the lung fill level only approximately and with greater uncertainty.

[0176] The optional pneumatic sensor 6 measures the pressure P.sub.es in the esophagus Sp of the patient P. In many cases, however, it is not desired to insert a pneumatic sensor 6 into the esophagus Sp of the patient, especially because the insertion and removal of the sensor takes a relatively long time and this would stress the patient P in some cases. In addition, a sensor 6 in the esophagus Sp measures the pneumatic indicator P.sub.mus for the breathing activity as well only with a time delay and superimposed by unwanted signals.

[0177] For the reasons stated above, it is desirable, on the one hand, to carry out the mechanical ventilation of the patient P as a function of a pneumatic indicator P.sub.mus for this intrinsic breathing activity, wherein the estimated values P.sub.mus,est(t.sub.i) are derived by means of measured values of sensors close to the signal source, here measured values of the measuring electrodes 2.1.1 through 2.2.2. On the other hand, the current intrinsic breathing activity indicator P.sub.mus shall be derived with sufficiently high reliability, so that the mechanical ventilation is synchronized with the intrinsic breathing of the patient P in a sufficiently reliable manner. Hence, in the exemplary embodiment, the mechanical ventilation is regulated on the basis of measured values of the measuring electrodes 2.1.1 through 2.2.2 as well as on the basis of measured values of the pneumatic sensor and optional measured values of additional sensors 4 and/or 6.

[0178] In one embodiment, a signal value Vol′(t.sub.i) for the volume flow Vol′, which is variable over time, is generated at each scanning time t.sub.i, and a signal value Vol(t.sub.i) for the current volume Vol, i.e., the current fill level of the lungs, is derived from this by means of numerical integration. In addition or instead of this, it is also possible to derive the signal value Vol(t.sub.i) for the current volume from the measured values of the optional sensor 4. Note: The scanning time t.sub.i is the time, to which a signal value or value for the breathing activity indicator P.sub.mus pertains. The value itself will be able to be calculated later.

[0179] According to the present invention, a lung mechanical model 20 is predefined and stored in the memory 9 in a computer-accessible form. This lung mechanical model 20 comprises at least one relationship, especially a model equation. The relationship or at least one relationship of the lung mechanical model 20 describes a connection between a breathing activity indicator P.sub.mus, which correlates with the intrinsic breathing activity of the patient P, and a plurality of measurable signals, especially at least some of the following signals: [0180] the airway pressure (pressure in airway, P.sub.aw), obtained from measured values of the sensor 3, the esophageal pressure (pressure in esophagus, P.sub.es), obtained from measured values of the sensor 6, [0181] the airway flow (flow, Vol′), likewise obtained from measured values of the sensor 3, [0182] the lung volume (Vol), derived from the airway flow Vol′ or obtained from measured values of the sensor 4, and/or [0183] the content of carbon dioxide (CO.sub.2) in the exhaled breathing air.

[0184] In one embodiment, the following two linear model equations are predefined as the lung mechanical model 20:

P.sub.aw(t)=R*Vol′(t)+E*Vol(t)+P.sub.mus(t)+P0 and  (2)

P.sub.mus(t)=k.sub.eff*Sig(t).  (3)

[0185] Herein [0186] P.sub.mus(t) is the breathing activity indicator, which is being sought and is variable over time and which correlates with the pneumatic pressure generated by the respiratory muscles of the patient P at the time t, [0187] P.sub.aw(t) is the airway pressure measured in the patient circuit, preferably as pressure difference in relation to the ambient pressure, wherein the airway pressure P.sub.aw is used as a measurable signal and during the mechanical ventilation results from a superimposition of the intrinsic breathing activity of the patient P and the ventilation by the ventilator 1, and otherwise exclusively from the intrinsic breathing activity, [0188] R is a lung mechanical factor, which describes the breathing resistance, which the airway of the patient P sets against the volume flow Vol′, [0189] E is a lung mechanical factor for the elasticity of the lungs of the patient, [0190] P0 is a lung mechanical constant, which is, for example, a pneumatic value for the effect of an incomplete exhalation (iPEEP) of the patient P, [0191] Sig(t) is the above-described electrical respiratory signal (EMG signal), or else, a mechanomyographic signal (MMG signal), which is determined by analysis of measured values of the measuring electrodes 2.1.1 through 2.2, or von MMG sensors, and [0192] k.sub.eff is a proportionality factor between the pneumatic pressure P.sub.mus and the electrical respiratory signal Sig of the measuring electrodes 2.1.1 through 2.2.2 or the mechanical respiratory signal, wherein the factor k.sub.eff describes the so-called electromechanical efficiency, i.e., how well electrical pulses are converted into muscle activity in the body of the patient P.

[0193] The introduction of (3) into (2) yields the following model equation:

P.sub.aw(t)=R*Vol′(t)+E*Vol(t)+k.sub.eff*Sig(t)+P0.  (4)

[0194] This model (4) is only approximately true. It has four model parameters, i.e., the lung mechanical factors R, E and k.sub.eff as well as the summand P0. The values of these model parameters are, as a rule, not known in advance and vary from patient to patient, and also in the same patient P with time. The values of the model parameters are therefore derived approximately from sets of signal values, which is described farther below.

[0195] In many cases, the summand P0 can be assumed to be constant over time. This model equation is in a preferred embodiment then differentiated in advance once after time, and the summand P0, which is assumed to be constant, disappears, as a result. The differentiation yields the following model equation:

P.sub.aw′(t)=R*Vol″(t)+E*Vol′(t)+k.sub.eff*Sig′(t).  (5)

[0196] Only three model parameter values are still to be estimated. The values of these signals Vol′, Vol, Sig can, in turn, subsequently be calculated by means of numerical integration.

[0197] In another embodiment, a model equation is predefined with additional summands and additional lung mechanical parameters, for example, the following model equation:

P.sub.aw(t)=R*Vol′(t)+E*Vol(t)+I*Vol″(t)+Q*Abs[Vol′(t)]*Vol′(t)+S*Vol.sup.2(t)+P.sub.mus(t)+P0.

Herein

[0198] Q describes the resistance to the air flow which the turbulent flow generates in the hose between the ventilator 1 and the patient P, [0199] S describes the change in the compliance of the lungs and/or of the chest as a function of the volume Vol, and [0200] I describes the resistance to the acceleration of the breathing air, wherein this resistance I is negligibly low in case of sufficiently low acceleration.

[0201] The same model equations (3) and (4) with possibly different model parameter values are used in another embodiment once for the inhalation (inspiration, subscript ins) and once for the exhalation (expiration, subscript exp), so that the following two model equations are used:

P.sub.aw,ins(t.sub.i)=R.sub.ins*Vol′(t)+E.sub.ins*Vol(t)+k.sub.eff,ins*Sig(t.sub.i)+P0.sub.ins  (7)

and

P.sub.aw,exp(t.sub.i)=R.sub.exp*Vol′(t)+E.sub.exp*Vol(t.sub.i)+k.sub.eff,exp*Sig(t)+P0.sub.exp  (8)

with a respective set of model parameters for inhalation and for exhalation.

[0202] It is also possible to calculate a respective value R.sub.ins or E.sub.ins, which is valid for the inhalation, and a respective value R.sub.exp or E.sub.exp, which is valid for the exhalation, only for the model parameters R and E. A respective single value, which is valid both for inhalation and for exhalation, is calculated for the other model parameters.

[0203] Sets of signal values, which have been generated from measured values measured during inhalation, are used exclusively to derive estimated values for the model parameters of the model equation (7). Correspondingly, the model parameter values of the model equation (8) are estimated exclusively using sets of signal values, which have been generated during exhalation.

[0204] In another embodiment, the following linear relationships are predefined as model equations:

P.sub.es(t)=E.sub.cw*Vol(t)−P.sub.mus(t)+P0 and  (9)

P.sub.mus(t)=k.sub.eff*Sig(t).  (3)

P.sub.es(t) is the esophageal pressure, which is measured, for example, by the pneumatic sensor 6 in the esophagus Sp. The factor E.sub.CW describes the elasticity based on the chest wall (chestwall) of the patient P.

[0205] The introduction of (3) into (9) yields the following model equation:

P.sub.es(t)=E.sub.cw*Vol(t)−k.sub.eff*Sig(t)+P0.  (10)

[0206] The above-mentioned model equations (2) through (10) are only ideally valid. The lung mechanical model 20 specified with at least one model equation describes reality only approximately, and the signals are superimposed by unwanted signals and are affected by measurement errors. Hence, the values of the model parameters also can only be derived approximately, and the derivation of the model parameter values and thus also the derivation of a value for the breathing activity are hence inevitably subject to an estimation uncertainty.

[0207] The following description pertains to the model equation

P.sub.aw(t)=R*Vol′(t)+E*Vol(t)+k.sub.eff*Sig(t)+P0.  (4)

[0208] The process described below can correspondingly also be applied to other model equations, which belong to a lung mechanical model 20.

[0209] A respective set of signal values, namely the set of signal values {P.sub.aw(t.sub.i), Vol′(t.sub.i), Vol(t.sub.i), Sig(t.sub.i)}, is generated from measured values at each scanning time t.sub.i.

[0210] Estimated values {R.sub.est(t.sub.i), E.sub.est(t.sub.i), k.sub.eff,est(t.sub.i), P0.sub.est(t.sub.i)} are derived for the model parameters—four in this case—by means of the lung mechanical model 20 and sets of signal values.

[0211] In a preferred embodiment, a regression method is applied to the predefined model equation (4) in order to derive a breathing activity value P.sub.mus,est(t.sub.i). A sum of squares error is especially preferably minimized.

[0212] In one embodiment, the model parameters {R, E, k.sub.eff, P0} in the model equation (4) are considered to be constant over time, and all sets of signal values generated hitherto are used to derive values for the model parameters.

[0213] In another embodiment, the fact that the values of these model parameters can change with time is taken into consideration. In one possible embodiment, a number N of scanning times is predefined. Estimated values {R.sub.est(t.sub.i), E.sub.est(t.sub.i), k.sub.eff,est(t.sub.i), P0.sub.est(t.sub.i)} are derived exclusively using the N sets of signal values, which are last in time, i.e., the last N scanning times up to scanning time t.sub.i (inclusive) form an analysis time window. The number N is, on the one hand, selected to be so high that a sufficiently reliable regression analysis can be carried out, and, on the other hand, so low that the model parameters {R, E, k.sub.eff, P0} can be considered to be constant over time in the analysis time window.

[0214] In one possible embodiment, the chronologically last N sets of signal values are weighted equally, i.e., for example, with a weighting factor α(t.sub.i)=1/N. In another embodiment, the weighting factor α(t.sub.i) of a set of signal values is smaller, the older this set of signal values is.

[0215] In another embodiment, it is determined to which respective time during a breath a set of signal values pertains. A weighting function is predefined, which describes the weighting factor as a function of the measurement time during a single breath. The period of time of a breath is preferably standardized. FIG. 3 shows as an example such a weighting function, wherein the time t is plotted on the x axis and the weighting factor α(t.sub.i) as a function of time is plotted on the y axis. Herein [0216] the interval from 0 to T is designated as the standardized or typical period of time for a single breath, [0217] T_I designates the beginning of inhalation (inspiration), [0218] T_E designates the beginning of exhalation (expiration), and [0219] x1, x2 and x3 designate three predefined weighting factors, wherein, for example, x3=2, x2=1 and x1=0.5.

[0220] In a third embodiment, the sets of signal values are weighted as a function of the respective set point of the first ventilator parameter and/or as a function of frequencies of signal values, preferably as follows: The fewer sets of signal values have been determined at a defined set point and/or the more seldom a signal value occurs in the sets of signal values used for the current estimation, the higher is the weighting factor for a set of signal values in the current estimation.

[0221] One example: During the last N scanning times t.sub.1, . . . , t.sub.N, N1 sets of signal values were determined at the standard set point, N2 sets of signal values were determined at a second set point, which is different from the standard set point, and N3 sets of signal values were determined at a third set point, which is different from the standard set point and from the second set point. Then, N=N1+N2+N3 is valid, the sets of signal values determined at the standard set point receive the weighting factor α(t.sub.i)=1/N1, the sets of signal values determined at the second set point receive the weighting factor α(t.sub.i)=1/N2 and the sets of signal values determined at the third set point receive the weighting factor α(t.sub.i)=1/N3.

[0222] FIG. 4 shows an example of such a weighting as a function of the frequency of set points and signal values. In the period of time T_O, an occlusion was carried out (no mechanical ventilation, and the intrinsic breathing of the patient is stopped), and the sets of signal values generated during the occlusion are especially highly weighted.

[0223] The weighting shown in FIG. 4 depends on the frequency of signal values of the signals P.sub.aw, Vol′, Vol and Sig. Sets of signal values with rarely occurring signal values receive a higher weighting than those with frequently occurring signal values. The weightings of the signal values, which [weightings] depend on the frequency, are combined into an overall weighting of a set of signal values. The time curve a(t.sub.i) of this overall weighting is shown in FIG. 4.

[0224] These embodiments can be combined. For example, each weighting factor a(t.sub.i) is designated as a product

α(t.sub.i)=α.sub.1(t.sub.i)*α.sub.2(t.sub.i)*α.sub.3(t.sub.i),

wherein the first factor α.sub.1(t.sub.i) depends on the age of the set of signal values, the second factor α.sub.2(t.sub.i) depends on the relative time during a single breath, and the third factor α.sub.3(t.sub.i) depends on the number of the sets of signal values determined at this set point and/or the number of signal values, cf. FIG. 4. The weighting factors of the N sets of signal values are preferably standardized, so that their sum is, e.g. equal to 1.

[0225] In a variant of this embodiment, a recursive regression method is applied at the first N scanning times, wherein four model parameter values R(t.sub.i−1, E(t.sub.i−1), k.sub.eff(t.sub.i−1) as well as P0(t.sub.i−1) have been derived before a scanning time namely on the basis of the N last scanning times with t.sub.i−1 as the last scanning time, and wherein four updated model parameter values {R.sub.est(t.sub.i), E.sub.est(t.sub.i), k.sub.eff,est(t.sub.i), P0.sub.est(t.sub.i)} are derived after the scanning time t.sub.i using the previous four model parameter values {R.sub.est(t.sub.i−1), E.sub.est(t.sub.i−1), k.sub.eff,est(t.sub.i−1), P0.sub.est(t.sub.i−1)} and the current set of signal values {P.sub.aw(t.sub.i), Vol′(t.sub.i), Vol(t.sub.i), Sig(t.sub.i)}. The subscript .sub.est shows that these are estimated values. This recursive method saves computing time and can be combined with the use of weighting factors.

[0226] In one embodiment, an estimated value for the pneumatic indicator P.sub.mus is derived at each scanning time t.sub.i as follows, cf. the model equation (3):

P.sub.mus,est(t.sub.i)=k.sub.eff,est(t.sub.i)*Sig(t.sub.i).  (11)

[0227] The derivation of the breathing activity indicator P.sub.mus(t.sub.i) is subject to uncertainty, especially in both factors k.sub.eff,est(t.sub.i) and Sig(t.sub.i). In one embodiment of the present invention, a so-called maneuver is carried out in case of low reliability in order to increase the reliability of the derivation. In this maneuver, a first operating parameter BG of the ventilator 1 is in the exemplary embodiment set from a standard set point EW_Std from time to time to at least one different set point and then back to the standard set point EW_Std. This maneuver is carried out, for example, for single breaths of the patient P. In case of the standard set point EW_Std, the ventilator 1 is regulated such that the mechanical ventilation is synchronized at best with the intrinsic breathing activity of the patient P, for example, such that:

P.sub.art(t.sub.i)=P.sub.mus,est(t.sub.i)is valid.  (12)

[0228] In case of a different set point, the ventilator 1 is regulated as follows, moreover, as a function of the derived breathing activity value P.sub.mus,est(t.sub.i), but deviating from the regular operation, e.g., with at least one of the following deviations from the regular operation: [0229] The proportional control according to (12) is carried out with a low degree of assistance x1<x—or else, with a higher degree of assistance x2>x. [0230] The volume flow Vol′ of breathing air, which flows from the ventilator 1 to the patient P, is limited to a maximum value. [0231] The pneumatic pressure P.sub.art, with which the ventilator 1 mechanically ventilated the patient P, is limited to a maximum value. [0232] The ventilator 1 fills the lungs of the patient P only up to a predefined volume limit. The patient P can achieve an additional increase in the lung volume only by intrinsic breathing activity. [0233] The amplitude and/or the frequency of ventilation strokes, which the ventilator 1 carries out, is reduced and/or limited. [0234] The ventilator 1 is switched over from a pressure-controlled ventilation, which is carried out at the standard set point, into a volume-controlled ventilation, which is carried out at the different set point. [0235] The ventilator 1 is switched over from a volume-controlled ventilation, which is carried out at the standard set point EW_Std, into a pressure-controlled ventilation, which is carried out at the different set point.

[0236] A maneuver may also consist of the ventilator 1 not being regulated at all, but rather being controlled or deactivated, or else, being regulated, but not as a function of the estimated breathing activity value P.sub.mus,est(t.sub.i), but, for example, as follows: [0237] The ventilator 1 is regulated as a function of the airway pressure P.sub.aw(t.sub.i) and/or of the volume flow Vol′(t.sub.i), which the pneumatic sensor 3 measures, and/or as a function of the esophageal pressure P.sub.es(t.sub.i), which the pneumatic sensor 6 measures. As explained above, it is a drawback to regulate the ventilator 1 continuously in this manner. A maneuver, in which the ventilator 1 is regulated for a short time in such a way and then again in a regular manner as described above, is useful in some cases, however. [0238] The ventilator 1 is controlled and is not regulated as a function of the intrinsic breathing activity of the patient P. In the control, the ventilator 1 uses, for example, a predefined desired curve for the pressure P.sub.aw or the volume flow Vol′ to be generated during the mechanical ventilation. [0239] The ventilator 1 completely sets the mechanical ventilation of the patient P (occlusion), and the intrinsic breathing activity of the patient P is stopped, for example, by valves at the ventilator 1 being closed and the patient being prevented from breathing. This occlusion is carried out for at most 5 sec, preferably for at most 1 sec, and is not hazardous for the patient P in case of such a short duration.

[0240] This occlusion is preferably carried out at a predefined relative time during a breath of the patient P, for example, at the end of the inhalation (end-inspiratory occlusion) or at the end of the exhalation (end-expiratory occlusion). The model equation

P.sub.aw(t)=R*Vol′(t)+E*Vol(t)+P.sub.mus(t)+P0  (2)

is also applied during an occlusion in one embodiment. During the occlusion, the volume flow Vol′ is negligibly low, so that Vol′(t)=0. During an occlusion at the end of exhalation, the remaining volume is contained in the summand P0, so that Vol(t)=0 is valid. In this case, therefore

P.sub.aw(t)=P.sub.mus(t)+P0.  (13)

[0241] Hence, P.sub.mus can be easily measured during an occlusion. However, thanks to the present invention, an occlusion only needs to be carried out when this is necessary.

[0242] According to the present invention, a reliability assessment ZM(t.sub.i) is calculated, which is an assessment for how reliable the derivation of the breathing activity value, here, i.e., P.sub.mus,est(t.sub.i), is.

[0243] For example, a sequence of the chronologically last M+1 estimated model parameter values {R.sub.est(t.sub.i−M), E.sub.est(t.sub.i−M), k.sub.eff,est(t.sub.i−M), P0.sub.est(t.sub.i−M)}, {R.sub.est(t.sub.i), E.sub.est(t.sub.i), k.sub.eff,est(t.sub.i), P0.sub.est(t.sub.i)} is used for this calculation.

[0244] In one embodiment, a covariance matrix is calculated from the last model parameter values at each scanning time t.sub.i, namely according to the calculation rule

[00001]$\begin{array}{cc}\mathrm{Cov}\left(t_i\right)=\left(\begin{array}{cccc}\mathrm{Var}\left(R,R\right)\left(t_i\right)& \mathrm{Cov}\left(E,R\right)\left(t_i\right)& \mathrm{Cov}\left(k_{\mathrm{eff}},R\right)\left(t_i\right)& \mathrm{Cov}\left(P0,R\right)\left(t_i\right)\\ \mathrm{Cov}\left(R,E\right)\left(t_i\right)& \mathrm{Var}\left(E,E\right)\left(t_i\right)& \mathrm{Cov}\left(k_{\mathrm{eff}},E\right)\left(t_i\right)& \mathrm{Cov}\left(P0,E\right)\left(t_i\right)\\ \mathrm{Cov}(R,& (\mathrm{Cov}(E,& \mathrm{Var}(k_{\mathrm{eff}},& \mathrm{Cov}(P0,\\ k_{\mathrm{eff}})\left(t_i\right)& k_{\mathrm{eff}})\left(t_i\right)& k_{\mathrm{eff}})\left(t_i\right)& k_{\mathrm{eff}})\left(t_i\right)\\ \mathrm{Cov}(R,& \mathrm{Cov}(E,& \mathrm{Cov}(k_{\mathrm{eff}},& \mathrm{Var}(P0,\\ P0)\left(t_i\right)& \mathrm{P0})\left(t_i\right)& P0)\left(t_i\right)& P0)\left(t_i\right)\end{array}\right).& \left(14\right)\end{array}$

[0245] A high cross correlation between two different model parameters, for example, a greater value for Cov(E,R) between E and R at the scanning time t.sub.i, means that the effect of these two model parameters E and R can be distinguished from one another only poorly based on the sets of signal values present up to now.

[0246] In one embodiment, the value P.sub.mus,est(t.sub.i) of the pneumatic indicator P.sub.mus at the scanning time t.sub.i is calculated according to the model equation (3) by means of the estimated respiratory signal Sig, i.e., according to

P.sub.mus,est(t.sub.i)=k.sub.eff,est(t.sub.i)*Sig(t.sub.i).  (11)

The empirical variance (empirical spread)

Var[P.sub.mus(t.sub.i)]=Var(k.sub.eff,k.sub.eff)(t.sub.i)*Sig(t.sub.i).sup.2  (15)

is preferably calculated as a value for the estimation uncertainty at the scanning time t.sub.i.

[0247] Other values for the estimation uncertainty can likewise be used.

[0248] In one variant, a value for the estimation uncertainty is calculated at the end of a respective breath or after a predefined period of time. If, for example, M scanning times t.sub.i−1, . . . , t.sub.i−M are in the period of time of this breath, then the arithmetic mean, the median or another mean is calculated via the M empirical variances

Var[P.sub.mus(t.sub.i+1)], . . . ,Var[P.sub.mus(t.sub.i+M)]

and used as the value for the estimation uncertainty.

[0249] As just described, the empirical variance

Var[P.sub.mus(t.sub.i)]=Var(k.sub.eff,k.sub.eff)(t.sub.i)*Sig(t.sub.i).sup.2  (15)

is used as a value for the estimation uncertainty in one embodiment, and the arithmetic mean or another mean via the empirical variances Var[P.sub.mus(t.sub.i+1)], Var[P.sub.mus(t.sub.i+M)] is used in another embodiment.

[0250] In another embodiment, the deviations and the measurement errors are combined into one error, which is variable over time,

err(t)=P.sub.aw(t)−R*Vol′(t)−E*Vol(t)−k.sub.eff*Sig(t)−P0.  (16)

If the model equation (4) were to describe reality exactly and no measurement errors were to occur, then err(t) would be equal to 0 at each time. This is not valid in reality, and err(t) varies over time. The signal processing unit 5 calculates a reliability assessment ZM(t.sub.i) and uses for this preferably N sets of signal values for the chronologically last N scanning times as well as the model equation (16) indicated above for the error err(t), which is variable over time. Preferably, the signal processing unit 5 applies a statistical method to calculate the reliability assessment ZM(t.sub.i). The ventilator 1 in the exemplary embodiment is operated with the standard set point EW_Std, after the ventilation was begun and as long as the signal processing unit 5 has not detected that a predefined triggering criterion E1 is met. At the standard set point EW_Std, for example, the pressure P.sub.an(t.sub.t) of the mechanical ventilation generated by the ventilator 1 is equal to x*P.sub.mus,est(t.sub.i), wherein the proportionality factor (degree of assistance) x remains constant. At the standard set point EW_Std, the ventilator 1 is, for example, always operated in a pressure-controlled manner.

[0251] As soon as the triggering criterion or a triggering criterion E1 is met, the signal processing unit 5 triggers a change step. The predefined triggering criterion E1, which triggers the change step, depends on at least one calculated reliability assessment and is met, for example, when at least one of the following events is detected: [0252] The chronologically last calculated reliability assessment ZM(t.sub.i) for the derivation of the breathing activity value, i.e., of a value P.sub.mus,est(t.sub.i) for the pneumatic indicator P.sub.mus, is below a predefined reliability limit. Synonymous with this is the fact that the value for the estimation uncertainty in the derivation of the breathing activity value P.sub.mus,est(t.sub.i) is above a predefined uncertainty limit. [0253] The chronologically last M calculated reliability assessments ZM(t.sub.i), ZM(t.sub.i−1), . . . always become smaller and come close to the reliability limit from above. [0254] At least one last calculated reliability assessment ZM(t.sub.i) is significantly smaller than at least one, preferably a plurality of previously calculated reliability assessments ZM(t.sub.i−n), ZM(t.sub.i−1).

[0255] According to the present invention, the signal processing unit 5 triggers a maneuver, i.e., a change step if it has detected that the triggering criterion E1 is met, especially when the last calculated reliability assessment ZM(t.sub.i) is below the predefined reliability limit or the estimation uncertainty value is above the predefined estimation uncertainty limit. A maneuver comprises the step of the ventilator 1 being operated from time to time with a set point different from the standard set point EW_Std. Examples of a maneuver were indicated above.

[0256] The maneuver is carried out with the goal of deriving values P.sub.mus,est(t.sub.i) for the breathing activity indicator P.sub.mus, which values were estimated with higher reliability during the maneuver and/or after the maneuver. A value P.sub.mus,est(t.sub.i) for the pneumatic indicator P.sub.mus is derived using sets of signal values, which have been generated at the different set point, as well as preferably additionally using sets of signal values, which have been generated before the maneuver, i.e., at the standard set point EW_Std.

[0257] The maneuver is ended as soon as the signal processing unit 5 has detected that the predefined ending criterion E3 is met. This ending criterion E3 is met, for example, when at least one of the following events has occurred: [0258] A predefined time limit has elapsed since the start of the maneuver, e.g., since the start of the occlusion, and the maneuver may no longer be continued. [0259] The chronologically last P calculated reliability assessments are above the predefined reliability limit, i.e., the reason for the maneuver no longer exists. [0260] The maneuver does not bring about an increase in the reliability assessment. A different maneuver is then preferably carried out instead of the maneuver currently being carried out.

[0261] As an example, the triggering and carrying out of maneuvers is explained below.

[0262] In this example, a first estimation uncertainty limit of, e.g., 1 mbar is predefined and a second, higher estimation uncertainty limit of, e.g., 2 mbar is predefined. As long as the estimation uncertainty value is below the first estimation uncertainty limit, the ventilator 1 is operated with the standard set point EW_Std. If the estimation uncertainty value is between the two estimation uncertainty limits, then an easier maneuver is carried out, in which the ventilator 1 is still regulated as a function of the estimated breathing activity value P.sub.mus,est(t.sub.i). An easier maneuver comprises, for example, at least one of the following steps: [0263] The degree of assistance x is reduced abruptly or even in a sliding manner to a smaller value x1<x during the maneuver, i.e., the ventilator 1 is operated according to P.sub.art(t.sub.i)=x1*P.sub.mus,est(t.sub.i). [0264] The assisting pressure P.sub.art is reduced calmly or otherwise below a maximum value for single breaths. [0265] The assisting pressure or the volume flow is limited.

[0266] It)

[0267] If the estimation uncertainty value is actually above the higher estimation uncertainty limit, then a more serious maneuver is carried out, in which the estimated breathing activity value P.sub.mus,est(t.sub.i) is not used, but rather, for example, an occlusion or a closed-loop control or an open-loop is carried out, instead, as a function of P.sub.aw(t.sub.i) and/or von Vol′(t.sub.i). Which more serious maneuver is carried out depends in one embodiment on the estimation uncertainty value, for example, on how far above the higher estimation uncertainty limit it is.

[0268] For example, the mechanical ventilation for a short period of time is completely set and the intrinsic breathing of the patient P is stopped (occlusion). During an occlusion, the airway pressure P.sub.aw, which the sensor 3 measures, depends only on the intrinsic breathing activity of the patient P, for example, P.sub.mus=P.sub.aw. After the end of the occlusion, the current value for the pneumatic indicator P.sub.mus is again derived using the signals P.sub.aw, Vol′ and Vol and the model equation (4), as just described above, wherein the signal values, which were measured during the occlusion are additionally used for the derivation, however.

[0269] In one embodiment, the maneuver, which is carried out in case of an estimation uncertainty value above the higher estimation uncertainty limit, depends on the covariance matrix Cov(t.sub.i) shown according to formula (14) or on a different value for the correlation between different model parameters. If, for example, the cross correlation Cov(R,k.sub.eff)(t.sub.i) between the two estimations R.sub.est and k.sub.eff,est is great, then the flow Vol′ of breathing air caused by the ventilator 1 is reduced during the maneuver. In the model equation

P.sub.aw(t)=R*Vol′(t)+E*Vol(t)+k.sub.eff*Sig(t)+P0  (2)

this reduction has an effect on the summand R*Vol′(t), but a markedly less effect on the summand k.sub.eff*Sig(t). If the cross correlation Cov(E,k.sub.eff)(t.sub.i) between the two estimations E.sub.est(t.sub.i) and k.sub.eff,est(t.sub.i) or the cross correlation Cov(R,E)(t.sub.i) between the two estimations R.sub.est(t.sub.i) and E.sub.est(t.sub.i) is great, then the ventilator 1 is actuated during the maneuver with the goal of keeping the volume Vol, i.e., the fill level of the lungs, constant for a predefined time. In the above model equation, this maneuver has an effect on the summand E*Vol(t), but a markedly less effect on the summand k.sub.eff*Sig(t).

[0270] FIG. 5 through FIG. 11 show a flow chart, which illustrates an exemplary embodiment of the process according to the present invention and of the signal processing unit according to the present invention.

[0271] FIG. 5 illustrates in a first part of the flow chart how an estimated breathing activity value P.sub.mus,est(t.sub.i) is derived and how it is decided whether the triggering criterion E1 is met. FIG. 6 shows in a second part of the flow chart the regular operation of the ventilator 1, i.e., the operation at the standard set point EW_Std. FIG. 7 shows in a third part of the flow chart how an easier maneuver is carried out. FIG. 8 shows how a more serious maneuver is carried out. FIG. 9 shows how sets of signal values are generated during a maneuver and how model parameter values are derived by means of these sets of signal values. FIG. 10 shows how a breathing activity value is derived during a maneuver. FIG. 11 shows how it is checked in a plurality of steps whether and how the mechanical ventilation of the patient P shall be continued.

[0272] The flow chart is explained below.

[0273] At the beginning of the mechanical ventilation, a first ventilator parameter BG is set at a predefined standard set point EW_Std. As long as this setting is maintained, the ventilator 1 is operated in the regular operation. Even after the end of a maneuver, the ventilator 1 is regulated in the regular operation. In this regular operation, the ventilator 1 is preferably regulated as a function of the pneumatic indicator P.sub.mus and a standard assistance factor x. As described above, a respective estimated value P.sub.mus,est(t.sub.i) or P.sub.mus,est.sup.m(t.sub.i) is derived at each scanning time t, and used as the breathing activity value. The superscript .sup.m indicates that the respective value was calculated or derived during a maneuver, which will be described farther below.

[0274] The signal processing unit 5 receives measured values from the sensors 2.1.1 through 2.2.2 and 3 and optionally from the optical sensor 4 and/or from the pneumatic sensor 6 in step S1. The signal processing unit 5 processes these measured values. This processing yields a respective set of signal values {P.sub.aw(t.sub.i), Vol′(t.sub.i), Vol(t.sub.i), Sig(t.sub.i)} for each scanning time

[0275] In step S2, the signal processing unit 5 derives from the sets of signal values a set {R.sub.est(t.sub.i), E.sub.est(t.sub.i), k.sub.eff,est(t.sub.i), P0.sub.est(t.sub.i)} of estimated model parameter values for the respective last N+1 scanning times t.sub.i−N through t.sub.i. For this, the signal processing unit 5 uses the lung mechanical model 20, for example, the predefined model equations

P.sub.aw(t)=R*Vol′(t)+E*Vol(t)+P.sub.mus(t)+P0 and  (2)

and

P.sub.mus(t)=k.sub.eff*Sig(t).  (3)

[0276] The signal processing unit 5 derives an estimated value P.sub.mus,est(t.sub.i) for the breathing activity of the patient P in step S3 and uses for this at least one estimated model parameter value, for example, according to the model equation

P.sub.mus,est(t.sub.i)=k.sub.eff,est(t.sub.i)*Sig(t.sub.i).  (11)

[0277] In step S4, the signal processing unit 5 calculates a reliability assessment ZM(t.sub.i) for the derivation of the breathing activity value P.sub.mus,est(t.sub.i). For example, the signal processing unit 5 calculates a value for the estimation uncertainty. The calculated reliability assessment ZM(t.sub.i) or the estimation uncertainty value may also depend on values, which have been calculated for earlier scanning times t.sub.i−1, t.sub.i−2, . . . .

[0278] The signal processing unit 5 automatically makes a decision E1? whether the predefined triggering criterion E1 is met or not. The triggering criterion E1 is met when the reliability for the derivation of the breathing activity value P.sub.mus,est(t.sub.i) is low, especially when the last calculated reliability assessment ZM(t.sub.i) is below a predefined reliability limit or is significantly smaller. Furthermore, when the triggering criterion E1 is met, the signal processing unit 5 makes the decision whether an easier maneuver (“leg” branch) or a more serious maneuver (“gray” branch) will be carried out.

[0279] If the triggering criterion E1 is currently not met (“no” branch), then the reliability assessment ZM(t.sub.i) is sufficiently large. The regular operation is maintained. FIG. 6 shows the steps, which are carried out in the regular operation. The signal processing unit 5 carries out in step S5 the upper-level control as a function of the derived breathing activity value P.sub.mus,est(t.sub.i). It calculates a set point P.sub.art(t.sub.i) for the pressure, which the ventilator 1 shall generate during the mechanical ventilation of the patient P, e.g., according to the rule

P.sub.art(t.sub.i)=x*P.sub.mus,est(t.sub.i),  (1)

[0280] In step S6, the signal processing unit 5 carries out the lower-level control and calculates as a function of the pressure set point P.sub.art(t.sub.i) the necessary actuating action or each necessary actuating action SE(t.sub.i), which is carried out with the goal that the ventilator 1 actually achieves this pressure P.sub.art(t.sub.i).

[0281] The steps described up to now are carried out again for the next scanning time t.sub.i+1=t.sub.i+Δ.

[0282] FIG. 7 shows the steps that are carried out in case of an easier maneuver (“leg” branch of decision E1?).

[0283] In step S7, the signal processing unit 5 specifies a set point EW_leg(t.sub.i) for the first ventilator parameter BG, which set point is different from the standard set point EW_Std. This different set point EW_leg(t.sub.i) may depend on the calculated reliability assessment ZM(t.sub.i).

[0284] In step S8, the signal processing unit 5 carries out the easier maneuver. In this case, the first ventilator parameter BG is set at the different set point EW_leg(t.sub.i), and the ventilator 1 is operated correspondingly.

[0285] In an easier maneuver, the breathing activity value P.sub.mus,est(t.sub.i), which is likewise derived in step S3, at this scanning time t.sub.i is used for controlling the ventilator 1. The ventilator 1 is, however, by contrast to the regular operation, operated corresponding to the different set point EW_leg(t.sub.i). For example, the degree of assistance is reduced to x1<x, or the pressure P.sub.art or the volume flow Vol′ is limited.

[0286] The signal processing unit 5 carries out the upper-level control as a function of the derived breathing activity value P.sub.mus,est(t.sub.i) and optionally additionally as a function of the different set point EW_leg(t.sub.i) in step S9. The signal processing unit 5, in turn, calculates a pressure set point P.sub.art.sup.m(t.sub.i). The superscript .sup.m indicates that this occurs during a maneuver.

[0287] In step S6, the signal processing unit 5 calculates the necessary actuating actions SE.sup.m(t.sub.i) during the easier maneuver, especially as a function of the pressure set point P.sub.art.sup.m(t.sub.i). The continuation for the next scanning time t.sub.i+1 is described farther below.

[0288] FIG. 8 shows the steps that are carried out in a more serious maneuver (“grav” branch of decision E1? in FIG. 5). By contrast to an easier maneuver, the breathing activity value P.sub.mus,est(t.sub.i), which is derived and subject to great uncertainty, is not used in the more serious maneuver.

[0289] In step S10 the signal processing unit 5 calculates a different set point EW_grav(t.sub.i) for the more serious maneuver. This set point EW_grav(t.sub.i) deviates, e.g., more greatly from the standard set point EW_Std than the set point EW_leg(t.sub.i) calculated for an easier maneuver in step S7 or leads to a markedly different operation of the ventilator 1 in a different way.

[0290] In step S11 the signal processing unit 5 triggers the step that the ventilator 1 carries out the more serious maneuver, wherein the first ventilator parameter BG is set at the set point EW_grav(t.sub.i).

[0291] In step S12 the signal processing unit 5 carries out the upper-level control as a function of the airway pressure P.sub.aw.sup.m(t.sub.i) measured during the maneuver and/or of the volume flow Vol′.sup.m(t.sub.i), i.e., without using the breathing activity value P.sub.mus,est(t.sub.i) which was derived in step S3, or controls the ventilator 1 or triggers an occlusion. This control may also depend on the different set point EW_grav(t.sub.i). Step S12, in turn, yields a pressure set point P.sub.art.sup.m(t.sub.i).

[0292] In step S6 the signal processing unit 5 calculates the necessary actuating actions SE.sup.m(t.sub.i), cf. FIG. 6.

[0293] Both in an easier maneuver and in a more serious maneuver, the signal processing unit 5 generates at least one set of signal values based on measured values, which have been measured during the maneuver, and subsequently derives model parameter values and a breathing activity value.

[0294] FIG. 9 shows steps, which are carried out both during the easier maneuver and during the more serious maneuver. In step S13 the signal processing unit 5 generates a set of signal values {P.sub.aw.sup.m(t.sub.i), Vol′.sup.m(t.sub.i), Vol.sup.m(t.sub.i), Sig.sup.m(t.sub.i)}. In step S14, the signal processing unit 5 calculates an estimated set {R.sub.est.sup.m(t.sub.i), E.sub.est.sup.m(t.sub.i), P0.sub.est.sup.m(t.sub.i)} of model parameter values and uses for this the set of signal values from step S13 and optionally older sets of signal values.

[0295] If no occlusion is carried out during the maneuver (“no” branch of the decision Okk?), then the following steps are carried out: Using the lung mechanical model 20 and at least one model parameter value, the signal processing unit 5 derives a breathing activity value P.sub.mus,est.sup.m(t.sub.i) (step S3 from FIG. 10). The signal processing unit 5, in turn, calculates a value ZM.sup.m(t.sub.i) for the reliability of the derivation of this breathing activity value P.sub.mus,est.sup.m(t.sub.i) (step S4 from FIG. 10).

[0296] If an occlusion is carried out during the maneuver (“yes” branch of the decision Okk?), then the mechanical ventilation of the patient P is set for a short period of time and the intrinsic breathing activity of the patient P is stopped, and the breathing activity indicator P.sub.mus can be measured directly. In step S16, the signal processing unit 5 receives measured values from the sensor 3 and generates signal value {P.sub.aw.sup.m(t.sub.i), Vol.sup.m(t.sub.i)}. When the occlusion does not take place at the end of a breath and the volume Vol cannot be ignored, the signal processing unit 5 uses an estimated value E.sub.est(t.sub.i), which was derived before the occlusion, for the factor E as well as an estimated value P0.sub.est(t.sub.i) for the summand P0. The signal processing unit 5 derives from these signal values {P.sub.aw.sup.m(t.sub.i), Vol.sup.m(t.sub.i)} and optionally from the model parameter values E.sub.est(L) and P0.sub.est(t.sub.i) a breathing activity value P.sub.mus.sup.m(t.sub.i), without using a respiratory signal Sig. In step S17, the signal processing unit 5 calculates a reliability assessment ZM.sup.m(t.sub.i) for the derivation of the estimated breathing activity value P.sub.mus,est.sup.m(t.sub.i) and uses for this the measured breathing activity value P.sub.mus.sup.m(t.sub.i).

[0297] FIG. 11 shows three decisions E2?, E3? and E4?, which are carried out one after the other. In the decision E2?, it is decided whether the treatment of the patient P shall be continued or ended. In the decision E3?, the signal processing unit 5 decides whether the current maneuver shall be ended and returned to the regular operation. One reason for ending the maneuver is that the reliability assessment ZM.sup.m(t.sub.i) calculated during the maneuver is sufficiently high. Another reason is that a predefined period of time has elapsed, for example, for an occlusion. If the maneuver shall be ended (“yes” branch of E3?), then the signal processing unit 5 in step S18 sets the ventilator parameter BG back again at the standard set point EW_Std. Otherwise (“no” branch of E3?), the signal processing unit 5 decides in decision E4? whether an easier maneuver (“leg” branch of E4?) or a more serious maneuver (“gray” branch of E4?) shall be continued. The signal processing unit 5 preferably uses the derived breathing activity value P.sub.mus,est.sup.m(t.sub.i) or the measured breathing activity value P.sub.mus,est(t.sub.i), which it has derived during the maneuver, for the next ventilation step during the regular operation (step S15 in FIG. 6).

[0298] While specific embodiments of the invention have been shown and described in detail to illustrate the application of the principles of the invention, it will be understood that the invention may be embodied otherwise without departing from such principles.

TABLE-US-00001 LIST OF REFERENCE NUMBERS 1 Ventilator; it mechanically ventilates the patient P; it comprises the signal processing unit 5 and the pneumatic sensor 3 2.1.1, Measuring electrodes on the skin of the patient P; they provide, together with a 2.1.2, ground electrode, the measured values, from which the respiratory signal Sig is 2.2.1, 2.2.2 generated 3 Pneumatic sensor in front of the mouth of the patient P; it measures the airway pressure P.sub.aw and the volume flow Vol’ 4 Optional optical sensor with an image recording device and with an image processing unit; it measures the geometry of the body of the patient P, from which the current lung filling level Vol is derived 5 Signal processing unit; it carries out the steps of the process according to the present invention; it has reading access to the memory 9 6 Optional probe in the esophagus Sp; it measures the pneumatic pressure P.sub.es in the esophagus Sp 9 Memory, in which the lung mechanical model 20 with the model equations used is stored and to which the signal processing unit 5 has reading access 20 Predefined lung mechanical model; it comprises at least one model equation; it is stored in a computer-accessible manner in the memory 9 α(t.sub.i) Weighting factor for the set of signal values that was determined at the scanning time t.sub.i BG First ventilator parameter; it is set at the standard set point EW_Std during the regular operation and at a different set point EW_leg(t.sub.i) or EW_grav(t.sub.i) during a maneuver Δ Interval between two scanning times t.sub.i and t.sub.i + .sub.1 E Model parameter in the form of a lung mechanical factor: Elasticity of the lungs of the patient P E.sub.est(t.sub.i) Estimated value of the model parameter E at the scanning time t.sub.i; it is derived at the standard set point EW_Std E.sub.est.sup.m(t.sub.i) Estimated value of the model parameter E at the scanning time t.sub.i; it is derived during a maneuver E1 Predefined triggering criterion, which triggers a maneuver after it is detected E1? Decision: Is triggering criterion E1 met? E2? Decision: Continue treatment of the patient P? E3 Decision: Is ending criterion met for ending the current maneuver? E4 Decision: Carry out easier or more serious maneuver? EW_grav(t.sub.i) Different set point for the first ventilator parameter BG in case of a more serious maneuver; it is set at the scanning time t.sub.i EW_leg(t.sub.i) Different set point for the first ventilator parameter BG in case of an easier maneuver; it is set at the scanning time t.sub.i EW_Std Standard set point for the first ventilator parameter BG; it is used during the regular operation of the ventilator 1 k.sub.eff Model parameter in the form of a factor for the neuromuscular efficiency, i.e., how well the respiratory muscles of the patient P convert electrical pulses into breathing activity, which leads to the pneumatic pressure P.sub.mus k.sub.eff.est(t.sub.i) Estimated value of the model parameter k.sub.eff at the scanning time t.sub.i; it is derived at the standard set point EW_Std k.sub.eff.est.sup.m(t.sub.i) Estimated value of the model parameter k.sub.eff at the scanning time t.sub.i; it is derived during a maneuver Okk? Decision: Carry out occlusion? P Patient with the esophagus Sp and the diaphragm Zw; he/she generates the pressure P.sub.mus based on his/her intrinsic breathing activity; he/she is ventilated mechanically by the ventilator 1 at least from time to time P.sub.art Assisting pressure, pressure generated by the mechanical ventilation P.sub.art(t.sub.i) Current value for P.sub.art during the regular operation; it is calculated as a function of P.sub.mus.est(t.sub.i) P.sub.art.sup.m(t.sub.i) Current value for P.sub.art during a maneuver; it is calculated as a function of P.sub.mus.est(t.sub.i) or P.sub.mus.sup.m(t.sub.i) P.sub.aw Airway pressure; it is generated by a superimposition of the intrinsic breathing activity of the patient P and the mechanical ventilation P.sub.art by the ventilator 1; it is measured by the sensor 3 P.sub.aw(t.sub.i) Signal value of the airway pressure P.sub.aw; it is generated during the regular operation at the scanning time t.sub.i P.sub.aw.sup.m(t.sub.i) Signal value of the airway pressure P.sub.aw; it is generated during a maneuver at the scanning time t.sub.i P.sub.es Pressure in the esophagus Sp of the patient P; it is measured with a probe 6 in the esophagus Sp P.sub.mus Pneumatic value for the intrinsic breathing activity of the patient P P.sub.mus(t.sub.i) Actual breathing activity value at the scanning time t.sub.i P.sub.mus.sup.m(t.sub.i) Breathing activity value derived by measurements during a maneuver at the scanning time t.sub.i P.sub.mus.est(t.sub.i) Derived estimated value for the pneumatic indicator P.sub.mus; it is derived at the standard set point EW_Std; it acts as the breathing activity value P.sub.mus.est.sup.m(t.sub.i) Derived estimated value for the pneumatic indicator P.sub.mus; it is derived during a maneuver; it acts as the breathing activity value P0 Model parameter in the form of a lung mechanical summand: Residual pressure after an incomplete exhalation of the patient P P0.sub.est(t.sub.i) Estimated value of the model parameter P0 at the scanning time t.sub.i; it is derived at the standard set point EW_Std P0.sub.est.sup.m(t.sub.i) Estimated value of the model parameter P0 at the scanning time t.sub.i; it is derived during a maneuver R Model parameter in the form of a lung mechanical factor: Breathing resistance, which the airway of the patient P sets against the volume flow Vol’ R.sub.est(t.sub.i) Estimated value of the model parameter R at the scanning time t.sub.i; it is derived during the regular operation R.sub.est.sup.m(t.sub.i) Estimated value of the model parameter R at the scanning time t.sub.i; it is derived during a maneuver S1 Step: Receive and process measured values, generate set of signal values {P.sub.aw(t.sub.i), Vol’(t.sub.i), Vol(t.sub.i), Sig(t.sub.i)} S2 Step: Calculate set of estimated model parameter values {R.sub.est(t.sub.i), E.sub.est(t.sub.i), k.sub.eff.est(t.sub.i), P0.sub.est(t.sub.i)} S3 Step: Derive estimated breathing activity value P.sub.mus.est(t.sub.i) S4 Step: Calculate reliability assessment ZM(t.sub.i) for the derivation of P.sub.mus.est(t.sub.i) S5 Step: Carry out upper-level control during the regular operation, calculate the pressure set point P.sub.art(t.sub.i) as a function of P.sub.mus.est(t.sub.i) S6 Step: Carry out lower-level control, calculate actuating actions SE(t.sub.i) and SE.sup.m(t.sub.i) as a function of the pressure set point P.sub.art(t.sub.i) and P.sub.art.sup.m(t.sub.i), respectively S7 Step: Specify different set point EW_leg(t.sub.i) for the first ventilator parameter BG during the easier maneuver S8 Step: Carry out easier maneuver, set the first ventilator parameter BG at the different set point EW_leg(t.sub.i) S9 Step: Carry out upper-level control during the easier maneuver, calculate the pressure set point P.sub.art.sup.m(t.sub.i) as a function of P.sub.mus.est(t.sub.i) and of the set point EW_leg(t.sub.i) S10 Step: Specify different set point EW_grav(t.sub.i) for the first ventilator parameter BG during the more serious maneuver S11 Step: Carry out more serious maneuver, set the first ventilator parameter BG at the different set point EW_grav(t.sub.i) S12 Step: Carry out upper-level control during the more serious maneuver, calculate the pressure set point P.sub.art.sup.m(t.sub.i) as a function of the measured signal values {P.sub.aw.sup.m(t.sub.i), Vol’.sup.m(t.sub.i)} S13 Step: Generate set of signal values {P.sub.aw.sup.m(t.sub.i), Vol’.sup.m(t.sub.i), Vol.sup.m(t.sub.i), Sig.sup.m(t.sub.i)} during the maneuver S14 Step: Derive model parameter values {R.sub.est.sup.m(t.sub.i), E.sub.est.sup.m(t.sub.i), k.sub.eff.est.sup.m(t.sub.i), P0.sub.est.sup.m(t.sub.i)} during the maneuver, use set of signal values {P.sub.aw.sup.m(t.sub.i), Vol’.sup.m(t.sub.i), Vol.sup.m(t.sub.i), Sig.sup.m(t.sub.i)} and N + 1 earlier sets of signal values {P.sub.aw(t.sub.i-N), Vol’(t.sub.i-N), Vol(t.sub.i-N), Sig(t.sub.i-N)} for this S15 Step: Carry out upper-level control during the regular operation, calculate the pressure set point P.sub.art(t.sub.i) as a function of P.sub.mus.sup.m(t.sub.i) or P.sub.mus.est(t.sub.i) S16 Step: Derive (direct measurement) the breathing activity value P.sub.mus.sup.m(t.sub.i) from the signal values {P.sub.aw.sup.m(t.sub.i), Vol’.sup.m(t.sub.i) during an occlusion S17 Step: Step: [sic-Tr.Ed.] Calculate reliability assessment ZM.sup.m(t.sub.i) for the derivation of P.sub.mus.est(t.sub.i) during the maneuver, use P.sub.mus.sup.m(t.sub.i) for this S18 Step: End maneuver, set first ventilator parameter BG at standard set point EW_Std SE(t.sub.i) Actuating actions during the regular operation for EW_Std; they are calculated in the lower-level control as a function of P.sub.art(t.sub.i) SE.sup.m(t.sub.i) Actuating actions during a maneuver; they are calculated in the lower-level control as a function of P.sub.art.sup.m(t.sub.i) Sig Electrical respiratory signal (EMG signal) for the breathing activity of the patient P; it is generated from measured values, which were measured by the measuring electrodes 2.1.1 through 2.2.2 Sig(t.sub.i) Signal value of the signal Sig at the scanning time t.sub.i; it is generated during the regular operation Sig.sup.m(t.sub.i) Signal value of the signal Sig at the scanning time t.sub.i; it is generated during a maneuver Sp Esophagus of the patient P t.sub.i Scanning time T_E Time, at which the patient P begins exhalation (expiration) T_I Time, at which the patient P begins inhalation (inspiration) T_O Time period, in which an occlusion is carried out Vol Volume (current fill level) of the lungs of the patient P; it is the integral of the volume flow Vol’ over time; it is measured in one embodiment by the optical sensor 4 Vol’ Flow of air into the lungs and out of the lungs of the patient P per time unit; it is the derivation of the volume Vol after time; it is measured, e.g., by the sensor 3 X Degree of assistance; it is a proportionality factor for the mechanical ventilation in case of a proportional control during the regular operation, i.e., the ventilator 1 is operated according to P.sub.art(t.sub.i) = x*P.sub.mus.est(t.sub.i) x1 Lower degree of assistance, which is used during an easier maneuver Zw Diaphragm of the patient P