Inductive sensing system and method

Abstract

An inductive sensing system (8) is adapted to apply an electromagnetic excitation signal into a body, the system comprising a resonator circuit (10) including a loop antenna (12). The system senses a signal returned from the body using the same antenna based on a change in an electrical characteristic of the resonator circuit. The system is configured to separate signals received from different physiological sources within the body. This is performed based on detecting in the resonator circuit both real and imaginary parts of an additional inductive component added to the antenna by the received electromagnetic signal. Separating signals from different physiological sources is based on the relative magnitudes of the detected real and imaginary inductive components added to the resonator circuit by the returned signal.

Description

Technical Field

[0001] The present invention relates to an inductive sensing system and method, in particular for sensing electromagnetic signals emitted from a body in response to propagation into the body of electromagnetic excitation signals. Background Art

[0002] There is often a need to measure the mechanical movement and dynamic changes of internal body structures, such as the heart, lungs, or arteries. For example, it is useful to measure the cyclically changing internal volume or size of a chamber of the heart or lungs or the mechanical activity of an artery (e.g., arterial volume that changes with the cardiac cycle).

[0003] Sensors that measure mechanical activity are sometimes called kymometric sensors (kymometric biometric sensors in the clinical field). Some examples of kymometric biometric sensors are accelerometer-based biosensors, transthoracic impedance biosensors, radar-based biosensors, capacitive sensors, and photoplethysmography (PPG) sensors.

[0004] Magnetic inductive sensors also have the potential to be used as biometric sensors for sensing mechanical activity. The operating principle of inductive sensing is based on Faraday's law. An oscillating primary magnetic field is generated by a generating loop antenna, and this induces eddy currents in the tissue radiated by the signal via Faraday's law. The eddy currents generate a secondary magnetic field. The total magnetic field is then a superposition of the primary and secondary magnetic fields. The induced changes in the electrical properties of the generating antenna (antenna current) can be measured and these are used to infer the properties of the secondary magnetic field and thus the stimulated tissue.

[0005] Inductive sensing offers the potential for simple contactless measurement of the mechanical activity of the heart and lungs, or of blood vessels such as the radial artery in the human arm.

[0006] A very significant disadvantage of omnidirectional biosensors (including inductive and other types of sensors) is that it is currently very difficult to distinguish between sensed signals originating from different physiological sources or activities.

[0007] For example, it is very difficult to distinguish within a single composite returned signal signal elements related to the mechanical activity of the heart (eg, heart pulse) from elements related to the mechanical activity of the lungs (eg, breathing).

[0008] In known systems, this is often performed by assuming that the heart rate is greater than the breathing rate. In this way, different signal components can be distinguished and separated based on frequency.

[0009] However, the clinically possible ranges for heart rate and pulse rate overlap. For example, the frequencies of particularly high respiratory rates overlap with the frequencies of particularly low pulse rates, and vice versa. Therefore, a sensed (high) respiratory rate may be incorrectly interpreted by prior art systems as a low heart rate, and vice versa, which leads to erroneous clinical diagnoses and interventions.

[0010] For example, in neonatal monitoring, apnea is often not detected by impedance-based measurements because the systolic nature of the condition is incorrectly interpreted as the patient's breathing rate.

[0011] Another related problem with known systems is that motion artifacts are also difficult to distinguish from real biophysical signals, such as pulse signals or respiration signals. Distinguishing motion artifacts from real signals is again often performed by assuming that the frequency of the artifact is different from the frequency of the biometric signal and / or the waveform of the artifact is different from the waveform of the biometric signal (e.g., in terms of shape characteristics). However, these approaches are often unsatisfactory because both the frequency and waveform of the signal and the artifact are often similar to the frequency and waveform of the biometric signal.

[0012] For example, if the patient's stepping frequency during walking is close to the pulse rate (which is usually the case), the measurement of the pulse frequency becomes unreliable. Moreover, even in the case of different frequencies, the patient's stepping frequency may still be interpreted as the pulse frequency, for example due to the very similar waveforms of the two. This leads to incorrect measurements of the heart rate.

[0013] Therefore, there is a need for an improved method of inductive sensing that can more reliably distinguish different physiological signals both from each other and from signal artifacts. Summary of the invention

[0014] According to an example of an aspect of the invention, there is provided an inductive sensing system for sensing an electromagnetic signal returned from a body in response to applying an electromagnetic excitation signal to the body, the system comprising:

[0015] a resonator circuit comprising a loop antenna;

[0016] a signal generating module, adapted to excite the loop antenna to generate the electromagnetic excitation signal,

[0017] a signal sensing module adapted to sense the returned signal from the body using the loop antenna based on detecting a change in an electrical characteristic of the resonator circuit;

[0018] Wherein, the system is suitable for:

[0019] detecting, from the sensed returned signal, a measure indicative of both the real and imaginary parts of an additional inductive component induced in the resonator circuit by the received electromagnetic signal, and

[0020] Using the detected measures indicative of real and imaginary inductive components, one or more individual signal components corresponding to different respective physiological sources are extracted from the returned electromagnetic signal, the extraction being based on the relative magnitudes of the detected real and imaginary inductive components added by the returned signal to the resonator circuit.

[0021] The extracted signal components are separate individual signal components, each signal component corresponding to a different respective known physiological source in the body. In other words, each signal component is an isolated signal component separated from any other signal component corresponding to a different physiological source.

[0022] Embodiments of the invention are based on the insight that signals originating from different physiological or anatomical sources in the body exhibit different characteristic relative amplitudes of real and imaginary parts of the additional inductive component induced in the sensing antenna when the signal from the source reaches the antenna.

[0023] The proposed method is based on measuring, from changes in the electrical characteristics of the antenna, measures both real and imaginary parts relating to or indicative of the additional inductive component induced at the antenna loop (or resonator circuit) by the overall composite signal received back from the body.

[0024] In some embodiments, an independent component analysis method can be applied to a signal representing a detected measure indicative of real and imaginary components in order to determine and extract separate individual signal components corresponding to different physiological sources. This can be done in an "online" manner, such that the detected signal representing the measure indicative of real and imaginary added inductive components is processed in real time using an independent component analysis process, and the different component physiological signals are separated and extracted based thereon.

[0025] According to another set of embodiments, data sets of known characteristic relative amplitudes of the real and imaginary parts of corresponding inductance signals (or measures related thereto) for different specific physiological signal sources may be consulted, and these may be used to extract signal elements corresponding to different physiological sources from the composite signal.

[0026] For example, the detected measure indicative of the real and imaginary extra inductive components and the stored measure or information relating to the different characteristic relative amplitudes of the real and imaginary inductive components can be considered or understood as vectors in the complex plane for said extra inductive components. The extraction of different signal components for different physiological sources can be achieved by effectively projecting the stored characteristic vectors onto the measured vectors of the real and imaginary parts. Only the part of the complex signal having the same directivity in the complex plane as one of the stored characteristic vectors is output from the projection. This part then corresponds to the signal component originating from the physiological source corresponding to the vector.

[0027] Typically, when the loop antenna comes close to the body, the inductance L acquires an additional reflected inductance component Lr due to eddy currents induced in the body by the stimulation as a result of the application of the excitation signal.

[0028] These eddy currents then effectively contribute to the inductance of the loop antenna due to the secondary time-varying magnetic flux. These eddy currents combine with the antenna's primary flux, resulting in a larger induced back EMF in the antenna, and therefore a larger measurable effective inductance.

[0029] The added component of inductance due to eddy currents may be synonymously referred to in this disclosure as “reflected inductance” L r The total inductance of the coil antenna is L t can be expressed as:

[0030] L t =L0+L r (1)

[0031] where L0 is the self-inductance of the antenna in free space, and L r It is the reflected inductance caused by the proximity of the stimulating body or medium.

[0032] Reflected Inductance and Reflected Impedance Z r Closely related. The relationship is L r =Z r / iω, where ω is the radial frequency of the electromagnetic excitation signal (the time-varying field applied to the body).

[0033] Usually, the reflected inductance L r is a plural number and can be expressed as:

[0034] L r =L′ r +iL″ r (2)

[0035] Among them, L′ r Related to the reactive impedance of the antenna, and L″ r Related to the resistive impedance of the antenna.

[0036] Inductance L r The addition of the reflected component results in a detuning of the characteristics of the antenna (or resonator circuit). In particular, both the natural radial frequency of the coil antenna circuit and the damping factor of the coil antenna circuit change.

[0037] In particular, the additional inductance component L r The real part of the additional inductive component is revealed in the frequency of the resonator circuit or antenna. The imaginary part of the additional inductive component is revealed in the amplitude of the resonator circuit. Therefore, in one set of examples, the real and imaginary parts of the additional inductive component can be detected based on changes in the frequency and amplitude of the resonator circuit (current), respectively. However, other methods are possible, as will be discussed below.

[0038] The resonator circuits mentioned above are used to couple with returning electromagnetic signals for the purpose of measuring or detecting those signals.

[0039] The signal sensing in the method is thus performed simultaneously with the signal generation.

[0040] The resonator circuit may include a capacitor component electrically coupled to the loop antenna.

[0041] The resonator circuit has an electrical resonant frequency. This may be defined at least in part by a capacitor component provided and comprised by the circuit.

[0042] As mentioned above, according to one set of embodiments, the extraction of individual signal components may be based on applying an independent component analysis procedure to detection measures indicative of real and imaginary inductive components.

[0043] According to another set of embodiments, the extracting is based on consulting a data set of information indicating known relative amplitudes of real and imaginary added inductive components for different physiological signal sources.

[0044] In particular, in this set of embodiments, the system may be configured as:

[0045] accessing a data set storing, for signals from a plurality of different known physiological sources within the body, information indicating characteristic relative magnitudes of real and imaginary inductive components added to a resonator circuit by signals received from the physiological sources; and

[0046] Using said detected measures indicative of real and imaginary inductive components and based on consulting said data set, said one or more individual signal components corresponding to different respective known physiological sources are extracted from said returned electromagnetic signal.

[0047] The data set stores information indicating the relative amplitudes of real and imaginary parts of inductive components added by signals originating from different physiological sources.By relative amplitudes may be meant relative to each other, for example the amplitudes are normalized such that their sum, or the sum of their squares, is unity.

[0048] The stored information indicating the relative real and imaginary magnitudes of the characteristic may take many different specific forms or representations. In some examples, the relative magnitudes may be stored in numerical form. In some examples, the relative magnitudes may be stored in vector form, for example where each relative magnitude forms a vector element or component.

[0049] In some examples, different measures may be stored that indicate or relate to, for example, the relative amplitudes of the real and imaginary inductive components or are proportional thereto. For example, in some examples, the physiological signal may be divided by each of the real and imaginary components and these quotients or ratios may be stored. In some examples, a complex parameter or phase of an additional (complex) inductive component induced by a signal received from a given physiological source may be stored. The parameter essentially indicates the direction within the complex plane of the relative amplitudes of the real and imaginary components of the additional inductive component from which it directly follows.

[0050] According to one set of examples, the stored information for each physiological signal source may take the form of a vector or point in the complex plane of an additional inductive component added to the sensing resonator circuit by the signal received from the physiological source.

[0051] The vector or point may be referred to in the present disclosure as a characteristic vector or point. The vector or point represents the additional inductive component added to the resonator circuit by the signal received from the corresponding physiological source.

[0052] In some examples, the characteristic vector may be normalized (ie, the vector may be a unit vector in the complex plane).

[0053] Instead of a vector, in other examples the stored information may simply take the form of a representation of the direction of such a vector within the complex plane.

[0054] In other examples, the stored information may take the form of a characteristic parameter or phase of an additional inductive component induced by a signal received from a physiological source (as described above).

[0055] According to one set of examples, the relative magnitudes of the characteristics represented by the stored information are weighted to provide a multiplicative mapping between detected measures indicative of an additional sensed inductive component and real and imaginary parts of an extracted signal component for a given physiological source.

[0056] This means that the stored metric indicative of the real part and the stored metric indicative of the imaginary part may be weighted such that each multiplication by the corresponding real or imaginary part of the measured additional inductive component directly results in a signal output representative of the signal component originating from the corresponding physiological source.

[0057] According to one set of examples, the stored information for each physiological signal source may include a characteristic vector having the following items:

[0058] a first vector component indicative of a ratio between a physiological signal received from the source and a measure indicative of a real part of a received additional inductive component of the signal, and

[0059] A second vector component indicative of a ratio between a physiological signal received from the source and a measure indicative of an imaginary portion of a received additional inductive component of the signal.

[0060] Here, the relative amplitudes of the characteristics stored in the data set are determined before performing the actual induced signal measurement and extraction process. The physiological signal received from the source may mean, for example, a real physiological signal in units of the measured physiological phenomenon. For example, the lung volume signal may be in liters.

[0061] The physiological signal means the physiological parameter or quantity being measured, i.e. it is represented by the additional inductive component added to the resonator circuit. Thus, the first vector component may indicate the ratio between the physiological parameter or quantity relating to the physiological source and a measure indicating the real part of the additional inductive component added to the resonator circuit by the signal received from the physiological source, and

[0062] The second vector component may be indicative of a ratio between a physiological parameter or quantity relating to the physiological source and a measure indicative of an imaginary part of an additional inductive component added to the resonator circuit by a signal received from the physiological source.

[0063] The first and second vector components correspond to corresponding measured real and imaginary components of the actual measured signal divided by the resulting sensed additional inductive component in the resonator circuit, respectively.

[0064] Thus, each vector includes elements based on the magnitude of the relevant physiological parameter measurement divided by the real part of the corresponding additional received inductance component, and elements based on the magnitude of the relevant physiological parameter measurement divided by the imaginary part of the corresponding additional inductance component.

[0065] In case the stored measures or information relating to the relative real and imaginary inductance components are stored or represented in the form of (characteristic) vectors, the system may be adapted to:

[0066] Extracting one or more signal components from different known physiological sources based on taking the inner product of the stored vector corresponding to each physiological source and a further vector representing the measure indicative of the real and imaginary parts of the sensed returned electromagnetic signal(s). The vector elements of the further vector may be the measured additional inductance components L induced in the resonator circuit r The real and imaginary parts, that is, for example (Re{L r},Im{L r}). The vector elements may each be a signal representing the real and imaginary components as a function of time.

[0067] Optionally, a low pass filter may be applied to components of the further vector before performing the inner product in order to remove a baseline and / or a band pass filter may be applied to the further vector before performing the inner product. The band pass filter may remove a baseline in addition to, for example, high frequency noise components.

[0068] The above-mentioned signal sensing module may be adapted to detect the measure indicative of the real part of the additional inductive component based on detecting a change in the reactance of the resonator circuit. The real part of the additional inductive component Lr results in a modified antenna reactance ω·Re{L r}. This will be explained further below.

[0069] Additionally or alternatively, the signal sensing module may be adapted to detect the measure indicative of an imaginary part of the additional inductive component based on detecting a change in the resistance of the antenna coil. r The antenna resistance that results in the modification of the amplitude -ω·Im{L r}. This will be explained further below.

[0070] According to one or more embodiments, the signal sensing module may be adapted to detect the measure indicative of the real part of the additional inductive component based on determining a change in the frequency of the resonator circuit current. As described above, the real part of the additional inductive component is revealed in a modified reactance of the resonator circuit current. The modified reactance results in a modified resonator frequency. This will be further explained below.

[0071] Monitoring changes in the frequency of the resonator circuit provides a very simple, direct way of detecting changes in the real part of the additional inductive component induced at the resonator circuit.The metric detected here is therefore the frequency.

[0072] Additionally or alternatively, the signal sensing module may be adapted to detect the measure indicative of the imaginary part of the additional inductive component based on determining a change in the amplitude of the current (electrical current) in the antenna coil or resonator circuit.

[0073] As described above, the imaginary part of the additional inductive component is revealed in a modified resistance of the resonator circuit current. This modified resistance results in a modified damping of the resonator circuit. This in turn results in a modified amplitude of the resonator circuit current. This will be explained further below.

[0074] Monitoring changes in the amplitude of the resonator circuit (current) provides a very simple and direct way of detecting changes in the imaginary part of the additional inductive component induced at the resonator circuit.The metric detected here is therefore the current amplitude.

[0075] As described above, in certain examples, the stored information for each physiological signal source may take the form of a characteristic vector or point in the complex plane for the additional inductance component. In certain examples, in such a case,

[0076] The signal generation module may be adapted to excite the resonator to generate an excitation signal having a radial frequency ω, and

[0077] The controller may be adapted to perform, in at least one control mode, an adjustment of the radial frequency from a first frequency to a second frequency, thereby increasing a degree (e.g., an average degree) of orthogonality between vectors formed in the complex plane by the real and imaginary components of the additional induced inductive components for the different physiological sources.

[0078] The frequency adjustment is thus a degree of increasing orthogonality between respective vectors formed in the complex plane by means of additional inductive inductive components added to the resonator circuit by signals received from different respective physiological sources.

[0079] This set of embodiments is based on the insight that the relative directions in the complex plane of the reflected complex inductance components induced by different physiological sources vary depending on the frequency of the excitation signal. Thus, by adjusting the frequency, the relative angle in the complex plane (i.e., the parameter angle) between these corresponding inductance components (such as represented as vectors) for different physiological sources varies. In this way, the frequency can be changed from a first frequency associated with a first orthogonality (e.g., an average orthogonality) between the vectors in the complex plane for different physiological sources to a second frequency associated with a second greater orthogonality between the vectors in the complex plane for different physiological sources.

[0080] The more orthogonal the vectors for different physiological sources (representing different corresponding inductive components) are with respect to each other, the better the signal extraction or separation process works. This renders the signal extraction process less prone to motion artifacts and other noise sources. Therefore, increasing the orthogonality of the vectors is beneficial.

[0081] In other words, the frequency of the applied excitation signal can be adjusted to better orthogonalize the induced reflected inductance L at the antenna for different physiological sources within the body.r The characteristic vector of the components. These characteristics L r A greater orthogonality of the signals results in a more reliable and robust extraction and separation of signal components of different physiological origin. It is therefore advantageous to adjust the applied signal frequency to maximize the orthogonality between the characteristic vectors.

[0082] The system (eg, controller) may monitor the degree of orthogonality of the vectors representing the different inductive inductance components while executing the regulation process. A feedback loop may be implemented such that frequency regulation is guided by the monitored vector orthogonality.

[0083] According to one or more embodiments, the system may also be adapted to perform a learning process including determining or updating stored information for each of the physiological sources. The learning process may be performed, for example, prior to use of the sensing system for a particular patient, as an initial process, such as a calibration or setup process. In some examples, there may not be stored information for different physiological sources until the learning process is performed. In some examples, the learning process may include a process of learning or determining or updating a fingerprint vector.

[0084] Additionally or alternatively, according to one or more embodiments, the system may be adapted to update the stored information for each physiological source to reflect changes induced in the real and imaginary components of the additional induced inductive components for different physiological sources by the frequency modulation.

[0085] There are different ways of implementing the learning procedure. According to one or more embodiments, the system may be adapted to record information about signal characteristics of each extracted signal component, and wherein the system is adapted to perform a learning procedure comprising adjusting the stored information for each of the physiological sources based on the recorded information.

[0086] For example, the system may be adapted to monitor the same signal characteristics in future extracted signal components and adjust the characteristic measures to reduce variations or increase consistency in those characteristics of future output signal components.

[0087] The recorded information may, for example, relate to shape characteristics of the extracted signal components, or other signal characteristics such as amplitude, frequency content, baseline, or any other signal characteristic.

[0088] The system may include a controller configured to perform the above process, or the signal processing module may perform, for example, the process.

[0089] According to one or more embodiments, the learning process may include performing an independent component analysis (ICA) process on the sensed return signal sensed at the antenna. It may, for example, include performing an ICA process on the signals representing the real and imaginary components of the inductance added to the resonator circuit by the returned signal. ICA is a known method in signal processing, but is never applied in the field of inductive sensing. The method is based on the assumption that the returned signal sensed at the resonator circuit antenna (in the form of an additional inductive component) is a composite signal formed by multiple signal components, each signal component corresponding to a different physiological source. ICA can be applied to determine or extract potential physiological signal components in the sensed inductive signal.

[0090] In a preferred example, the real and imaginary components of the additional inductance sensed at the antenna are each treated as individual signals, and the ICA procedure is applied to these real and imaginary inductance signals to determine the component physiological signals forming them.

[0091] According to one or more embodiments, the system may be adapted to implement a multi-frequency drive scheme, including driving (exciting) the antenna at multiple frequencies in a continuous manner. It may cycle between multiple different frequencies. According to one or more embodiments, the system may be adapted to drive the antenna with multiple drive frequencies in a time multiplexed manner and sense a returned signal at the antenna at each of the drive frequencies. Time multiplexing means that the antenna switches continuously between the different frequencies.

[0092] The multi-frequency drive scheme thus allows obtaining one or more measurement signals for each separate frequency in each duty cycle. This thereby increases the number of signals obtained.

[0093] The antenna alternates or cycles through different drive frequencies one frequency at a time in a continuous manner and inductive sensor measurements taken simultaneously with signal generation at each of the drive frequencies.

[0094] According to one or more sets of embodiments, the system may include a signal processing module configured to process the extracted one or more signal components and derive one or more physiological parameter measurements based on the signal component(s).

[0095] The signal sensing module may be pre-programmed, for example, with one or more algorithms to facilitate the determination (ie, quantification) of the physiological parameter measurements.

[0096] In some examples, the signal processing module may be included by the signal sensing module, or the functionality may be facilitated by the signal sensing module.

[0097] In other examples, the parameter measurement derivation process may be provided by a separately provided controller, or by any other suitable component of the system.

[0098] An example according to another aspect of the present invention provides an inductive sensing method for sensing an electromagnetic signal returned from a body in response to applying an electromagnetic excitation signal to the body, the method comprising:

[0099] applying an electromagnetic excitation signal to the body using a resonator circuit including a loop antenna and an electrical coupling capacitor;

[0100] sensing the returned signal from the body using the loop antenna based on detecting a change in an electrical characteristic of the resonator circuit;

[0101] detecting, from the sensed returned signal, a measure indicative of both the real and imaginary parts of an additional inductive component induced in the resonator circuit by the received electromagnetic signal,

[0102] Based on the use of detection metrics indicative of real and imaginary inductive components, one or more individual signal components from different known physiological sources are extracted from the returned electromagnetic signal, the extraction being based on the relative amplitudes of the detected real and imaginary inductive components added by the returned signal to the resonator circuit.

[0103] Implementation options and details for each of the above steps may be understood and explained based on the explanation and description provided above for the apparatus aspect (ie, system aspect) of the present invention.

[0104] Any of the examples, options or embodiment features or details described above with respect to the apparatus aspect of the invention (ie, with respect to the system) may be applied or combined or incorporated into the present method aspect of the invention.

[0105] According to one or more embodiments, the extraction may be based on performance of an independent component analysis procedure, as discussed above with respect to the system.

[0106] According to another set of embodiments, the method may be based on the use of a data set. In particular, the method may include:

[0107] accessing a data set storing, for signals from a plurality of different known physiological sources within the body, information indicating characteristic relative magnitudes of real and imaginary inductive components added to a resonator circuit by signals received from the physiological sources; and

[0108] Based on the use of the detection measure indicative of real and imaginary inductive components and based on consulting the data set, one or more individual signal components from different known physiological sources are extracted from the returned electromagnetic signal.

[0109] According to one or more embodiments, the stored information for each physiological signal source may take the form of a vector or point in the complex plane of an additional inductive component added to the sensing resonator circuit by the signal received from the physiological source. The vector or point may be referred to as a characteristic vector or point in the present disclosure.

[0110] Thus, the stored information for each physiological signal source may take the form of a vector or point in the complex plane representing the additional inductive component added to the resonator circuit by the signal received from the respective physiological source.

[0111] According to one or more embodiments, detecting the measure indicative of the real part of the additional inductive component may be performed based on detecting a change in a frequency of a current in the resonator circuit.

[0112] Additionally or alternatively, according to one or more embodiments, detecting the measure indicative of an imaginary part of the additional inductive component is performed based on detecting a change in the amplitude of the current in the resonator circuit.

[0113] According to one or more embodiments, the resonator circuit may be controlled to generate an excitation signal having a radial frequency ω, and

[0114] The method may include performing, in at least one control mode, adjustment of the radial frequency from a first frequency to a second frequency, thereby increasing a degree (e.g., an average degree) of orthogonality between vectors in a complex plane formed by the real and imaginary components of the additional induced inductive components for different physiological sources.

[0115] In other words, the frequency adjustment is to increase the degree of orthogonality between respective vectors formed in the complex plane by means of additional inductive inductive components added to the resonator circuit by signals received from different respective physiological sources.

[0116] These and other aspects of the invention will be apparent from and elucidated with reference to the embodiment(s) described hereinafter. BRIEF DESCRIPTION OF THE DRAWINGS

[0117] For a better understanding of the invention and to more clearly show how it may be implemented, reference will now be made, by way of example only, to the accompanying drawings, in which:

[0118] Figure 1 schematically depicts applying an excitation signal to a body and receiving a return secondary signal from the body;

[0119] Figure 2 The right and left traveling potential field components at the dielectric layer interface are shown;

[0120] Figure 3 illustrates an example setup of a sensing system and application of an excitation signal to a multilayer medium;

[0121] Figure 4 shows a diagram illustrating the variation of the complex phase of the induced reflection inductance at the antenna as a function of the position of the signal source;

[0122] Figure 5 An example inductive sensing system according to one or more embodiments of the present invention is shown in block diagram form;

[0123] Figure 6 schematically depicts a workflow for extraction of signal components originating from different physiological signal sources;

[0124] Figure 7 shows a diagram illustrating extraction of physiological source signals based on measured inductive sensing signals using an independent component analysis method;

[0125] Figure 8 illustrates the directions of example characteristic vectors for reflected inductance signals from two different example physiological sources;

[0126] Fig. 9 shows the change in direction of the characteristic vector for the reflected inductance signal as a function of the changing frequency of the applied excitation signal;

[0127] Fig.10 illustrates orthogonal eyes of characteristic vectors of reflected inductance signals of different physiological sources based on adjusting the frequency of an applied excitation signal; and

[0128] Fig.11 A block diagram of an example inductive sensing method is shown in accordance with one or more embodiments. DETAILED DESCRIPTION

[0129] The present invention will be described with reference to the accompanying drawings.

[0130] It should be understood that the detailed description and specific examples, when indicating exemplary embodiments of the apparatus, system and method, are intended for illustration purposes only and are not intended to limit the scope of the invention. These and other features, aspects and advantages of the apparatus, system and method of the present invention will become better understood from the following description, claims and accompanying drawings. It should be understood that the drawings are only schematic and are not drawn to scale. It should also be understood that the same reference numerals are used throughout the drawings to indicate the same or similar parts.

[0131] The present invention provides an inductive sensing system suitable for applying an electromagnetic excitation signal into a body, the system comprising a resonator circuit including a loop antenna. The system senses a signal returned from the body using the same antenna based on a change in an electrical characteristic of the resonator circuit. The system is configured to separate signals received from different physiological sources within the body. This is performed based on detecting in the resonator circuit both real and imaginary parts of an additional inductive component added to the antenna by the received electromagnetic signal. Separating signals from different physiological sources is particularly based on the relative amplitudes of the detected real and imaginary inductive components added to the resonator circuit by the returned signal.

[0132] In one set of embodiments, a reference data set is consulted which stores, for different signals from different known physiological sources within the body, indicators of known characteristic relative magnitudes of real and imaginary inductive components added to the sensing antenna circuit by the signal originating from that source. By consulting this data set, and using the detection measures indicative of the real and imaginary inductive components, one or more signal components from different known physiological sources are extracted from the received electromagnetic signal.

[0133] In other embodiments, different methods may be used to extract different physiological signals from the received composite signal based on the detected real and imaginary components of the added inductive component. For example, in one set of embodiments, an independent component analysis process may be applied to a signal representing inductive measurements of the real and imaginary parts indicative of the additional inductive component added to the resonator circuit. This enables the extraction of different signal components that contribute to the real and imaginary inductive signals originating from different physiological sources. This will be explained in more detail below.

[0134] A number of example embodiments will be discussed below that are based on consulting a data set storing information indicating known characteristic relative magnitudes of real and imaginary added inductive components for different physiological sources. However, this represents only one example set of possible embodiments, and the use of a data set is not essential.

[0135] The present invention provides a novel solution to the problem of separating from the returned signal different signal components originating from different physiological sources (eg heart pulse, lung inflation changes, arterial behavior).

[0136] The present invention is based on the insight that electromagnetic signals returning from the body from different physiological or nomological origins exhibit different "fingerprints" in the form of the relative strengths of the real and imaginary parts of the corresponding inductive components (reflected inductance) induced at the antenna by the returned signal. This insight provides a simple and efficient method for separating signal components from different physiological sources.

[0137] A straightforward way to apply or represent this mathematically is the use of vectors in the complex plane. The complete (complex) returned signal from the body can be represented as a 2D vector, with the real and imaginary parts of the additional induced inductance at the antenna (or measures related to or indicative of it, such as antenna current frequency and amplitude) constituting the vector components. Different known signal "fingerprints" for different physiological sources can each be represented as a 2D vector, with the vector components corresponding to the relative amplitudes of the real and imaginary parts of the inductance induced by the signal from that physiological source (or measures related to or indicative of it). A detailed explanation of the exemplary embodiments will be further described below.

[0138] More broadly, signal components originating from different physiological or anatomical origins within the body may be understood as contributing added inductive components with different complex phases (different complex parameters).

[0139] The general theory underlying embodiments of the present invention will now be described.

[0140] Embodiments of the present invention provide inductive sensing systems and methods, and are therefore based on the principles of magnetic induction. First, the basic principles of magnetic induction will be briefly outlined.

[0141] Embodiments of the present invention operate on the principle of inductive coupling, whereby a coil or wire has a potential difference induced across it due to exposure to a time-varying magnetic field. Embodiments of the present invention use this principle to measure the strength of electromagnetic signals generated in an area of ​​the body by sensing changes in the inductance of a coil or loop antenna placed close to the body, wherein these changes are detected based on changing the resonant characteristics of the antenna or resonator circuit.

[0142] Any electrical conductor exhibits the property of self-inductance. Self-inductance is the property of an electrical conductor whereby a change in the current driven through the conductor results in the induction of an electromotive force in the conductor. According to Lorentz's law, the induced electromotive force is in a direction such as to oppose the change in the current that induced it. It is therefore often referred to as the "back EMF". Self-inductance can be understood as occurring due to the magnetic flux induced as a result of the change in current (Ampere's law). This flux then interacts with the conductor itself to induce a back EMF (Faraday's law of induction and Lorentz's law).

[0143] The relationship between the circuit's self-inductance L, voltage v(t), and current I(t) can be expressed as:

[0144]

[0145] By expressing v(t) as dΦB / dt (where ΦB is the magnetic flux) using Faraday's law of induction and integrating with respect to time (assuming L is a time constant), the free space self-inductance L for a coil of N turns can be expressed as:

[0146] L=NΦ B / I (4)

[0147] Embodiments of the present invention utilize resonators, including antennas (which in preferred embodiments may include only a single turn loop) to stimulate or excite the body with electromagnetic signals (waves) and sense signals emitted back from the body in response to those excitation signals.

[0148] The coil can be driven with alternating current to generate excitation signals for application to the body. These can be propagating electromagnetic signals, in some cases propagating into the medium, or the signals can include non-propagating electromagnetic fields applied to the medium, i.e., by bringing the loop antenna source close to the target medium. Alternating current produces a field of alternating field strength.

[0149] When the coil comes close to the body, the inductor L collects the additional reflected inductance component L due to the eddy currents induced in the stimulated body. r As a result of application of an excitation signal.

[0150] This is Figure 1 , which shows by way of example that the loop antenna 12 is driven with an alternating current close to the chest 16 of the subject so as to apply an electromagnetic signal 22 to the chest.

[0151] Thus, eddy currents 18 are induced within the thorax.

[0152] These eddy currents in turn effectively contribute to the inductance of the loop antenna 12. This is because they themselves result in the generation of a time-varying magnetic flux 24 of an equivalent frequency to the frequency generated by the primary antenna 12. These eddy currents combine with the primary flux of the antenna, which results in a modified induced back EMF in the antenna, and therefore a greater measurable effective inductance.

[0153] The added component of inductance due to eddy current is called "reflected inductance" L r The total inductance L of the coil antenna 12 is t can be expressed as:

[0154] L t = L0 + L r (5)

[0155] Wherein, L0 is the self-inductance of the coil antenna 12 in free space, and L r It is the reflected inductance caused by the presence of a close body.

[0156] The reflected inductance can be defined as:

[0157]

[0158] Among them, A ris the reflected part of the electromagnetic vector potential (i.e., the part generated by the eddy currents 18 in the exciting medium), and I is the coil current. r Closely related. The relationship is L r =Z r / iω, where ω is the radial frequency of the electromagnetic excitation signal 22 (the time-varying field applied to the body).

[0159] The above formula can be understood by applying the following relationship:

[0160]

[0161] Among them, B r is the “reflected” magnetic field and Stokes’ theorem is then applied to reformulate equation (6) as:

[0162]

[0163] Where N = number of turns. It can be seen that this corresponds to a simplified form of expression for the inductance outlined above in equation (4).

[0164] The amplitude of the reflected inductive component gives an indication of the strength of the "reflected" electromagnetic signal transmitted back from the body.

[0165] Usually, the reflected inductance L r is a plural number and can be expressed as:

[0166] L r =L′ r +iL″ r (9)

[0167] Among them, L′ r is related to the reactive impedance of the coil antenna, and L″ r Related to the resistance impedance of the coil.

[0168] Plural L r Of course, it can also be expressed in extreme form:

[0169]

[0170] in, Is plural L r The parameter or phase (usually given by Given).

[0171] Inductance L r The addition of the reflected component of leads to a detuning of the electrical characteristics of the resonator circuit. In particular, both the natural radial frequency of the resonator circuit and the damping factor of the resonator circuit change. By measuring this detuning of the electrical characteristics, the reflected inductance L can be detected. r The real and imaginary parts of .

[0172] In particular, the detuning of the characteristics of the coil as a result of the addition of the reflected inductance can be expressed as follows:

[0173]

[0174] in, is the undamped natural radial frequency of the coil circuit in free space, ω 0，t The natural undamped radial frequency of the coil circuit in the presence of a medium or body (the subscript t stands for "total"). is the damping factor in free space, ζ t is the (total) damping factor in the case of a medium, L′ r is the real part of the reflected inductance defined in equation (9), and L″ r is the imaginary part of the reflected inductance defined in equation (9).

[0175] It can be seen that the detuned natural radial frequency depends only on the real part of the reflected inductance L′ r The detuning damping factor depends only on the imaginary part of the reflected inductance L″ r .

[0176] For simplicity, it may be preferable to work with geometrically normalized quantities. and characteristic reflected inductance It can be defined as follows:

[0177]

[0178] Where l = the circumference of a single turn of the coil, N = the number of coil windings, L0 is the free space self-inductance (real), L r is the reflected inductance (complex), and where L r is defined as in equation (6) above. The benefit of using geometric normalization lies in the fact that: Independent of system size and number of turns in the antenna coil.

[0179] Using these characteristic quantities, the detuning of the characteristics of the coil as a result of the addition of the reflected inductance can be expressed as follows:

[0180]

[0181] in, is the undamped natural radial frequency of the coil circuit in free space, ω 0，t is the natural undamped radial frequency of the coil circuit in the presence of a medium or body (the subscript t stands for "total"). is the damping factor in free space, ζ tis the (total) damping factor in the case of a medium, is the real part of the characteristic reflected inductance defined in equation (9), and is the imaginary part of the characteristic reflected inductance defined in equation (9).

[0182] The basic principles of magnetic induction on which the sensing system of embodiments of the present invention relies have been outlined above.

[0183] As mentioned above, the present invention is based on the following insight: the additional reflected inductance component L r The relative amplitudes of the real and imaginary parts of L vary depending on where in the body the reflected signal has originated. Another way to express this is that L r The complex phase or parameter of varies depending on the location within the body from which the corresponding reflected signal component originates. This principle allows signal components from different anatomical locations to be identified.

[0184] The theory behind this physical insight will now be explained in more detail. This will be done by showing how a reflected inductance signal L originating from different locations within the body can be constructed. r The spatial mapping of phase or parameter changes is completed.

[0185] In particular, it can be shown how to arrive at the characteristic reflected inductance The spatial mapping of the origin of the inductance M(r′,z′) (in nH / m) is given by:

[0186]

[0187] The integration is defined to occur over the depth coordinate (within the stimulation body) z' and the transverse radial coordinate (within the stimulation body) r'. The spatial mapping function M(r', z') represents the spatial mapping of the sources of sensed reflected inductance across the body.

[0188] First, the basic inductive sensing equation can be stated. This can be derived from Maxwell's equations and using cylindrical coordinates Circular symmetry is assumed about the z-axis. Furthermore, the induced eddy currents in the body (which will be treated as the source of the modeled "reflected" or secondary fields) will be understood to be confined to the radial planes that circulate in the azimuthal direction (i.e., circumferentially around the z-axis). The equation thus describes the current density generated by the purely azimuthal current source Induced pure azimuth vector potential

[0189]

[0190] in, is the squared spatially varying total wave number, ε(r, z) is the permeability of the medium, σ(r, z) is the conductivity of the medium, μ0 is the vacuum permeability, and ω is the radial frequency of the eddy currents. Due to the spatial correlation of ε(r, z) and σ(r, z), the inductive sensing equation (17) allows for inhomogeneous media and is therefore not limited to layered media. For the purpose of sensing biological tissue, the median permeability μ can be approximated as μ≈μ0.

[0191] In general, the solution to the Maxwell equations for a problem involving a dielectric will be the same as the solution to the equivalent free space problem, where the source term itself includes all the induced currents and induced charges of the first problem. This allows the construction of a solution using the free space Green's function and the total azimuthal current density of the inductive sensing equation (17):

[0192]

[0193] in, is the total azimuthal current density, which includes both source and induced currents, and G0(r,z;r′,z′) is the free-space Green’s function implicitly defined by:

[0194]

[0195] in, The solution to the free-space Green's function can be obtained by using the Fourier transform in z of the contour integral in the complex plane, followed by the Hankel transform in r, and then the inverse Fourier transform in z, yielding:

[0196]

[0197] in, is a first-order Bessel function of the first kind, is the complex longitudinal wavenumber in free space (selected in the fourth quadrant), and ξ is the transverse wavenumber (the wavenumber in the radial r' direction).

[0198] The total current density in equation (18) includes the source current in the antenna loop and the eddy current in the probing body in a linear sum, i.e.,

[0199]

[0200] This allows a straightforward separation of the reflection vector potential (the quantity sought) from the source vector potential in equation (18) to achieve:

[0201]

[0202] As discussed above, consider the "characteristic" reflected inductance Rather than the reflected inductance itself, it is convenient to refer to the reflected inductance normalized by the antenna loop length and the number of windings (see equation (14) above). Considering an antenna with a single toroidal winding (i.e., N=1 in equation (14)), the characteristic reflected inductance can be obtained from the reflection vector potential at the toroidal line And it is derived using the combination of equations (14) and (6) as:

[0203]

[0204] Therein, I is the loop current, a is the loop radius, and h is the distance between the loop and the probing medium (assuming that the surface of the probing medium is at z=0 and the medium extends in the region where z>0).

[0205] Combining equations (22) and (23) yields an equation describing the (seeking) spatial mapping of the characteristic reflected inductance signal as a function of the position (r,z) of the reflected signal source within the detected body:

[0206]

[0207] where the free space Green's function G0 is given by equation (20).

[0208] The right side of equation (24) corresponds to the right side of the spatial mapping equation (16b) introduced above, namely: Therefore, the spatial mapping function M(r',z') can be expressed as follows:

[0209]

[0210] The spatial map M is a characteristic reflection inductance of the stimulated body. The spatial distribution of the origin.

[0211] Equation (24) (and Equation (18)) describes the expected magnitude of the additional component of inductance (the reflected inductance component) induced in the resonator circuit by the returning (reflected) electromagnetic signal originating from the location (r', z') in the probe body (wherein, assuming that the azimuth direction circular symmetry).

[0212] Therefore, from equation (24), it can be seen that the reflected inductance can be understood as a quantified characteristic of the return (reflected) electromagnetic signal from the body. can be expressed as normalized to the loop current Convolution of the “loop sensitivity kernel” G0(a, -h; r′, z′) on the eddy current.

[0213] The loop sensitivity kernel G0 is a function of the location (r', z') within the stimulus body and effectively quantifies the effect of the current at the location (r', z') normalized to the loop current on the characteristic reflected inductance sensed at the resonator circuit. This is expressed mathematically in equation (24) above. Therefore, a simple multiplication of the loop sensitivity kernel with the normalized eddy current provides a spatial map M(r', z') of the origin of the reflected inductance within the stimulus body, as shown in equation (24b). The spatial map M(r', z') can be directly interpreted as the characteristic reflected inductance within the body The spatial distribution of the origin.

[0214] In order to quantitatively model or map the reflected inductance as a function of position (as described in Eq. (24)), it is necessary to calculate the induced eddy current Eddy currents include conduction current and polarization current, that is,

[0215]

[0216] in, is the electric susceptibility, ε is the permittivity of the medium, σ(r, z) is the conductivity of the medium, and ω is the eddy current radial frequency.

[0217] The spatial map of the characteristic reflected inductance source M(r',z') can now be found by combining equations (25) and (24b), yielding:

[0218]

[0219] The mapping function M(r,z') consists of three factors. Factor is the vector potential normalized to the loop current. This last factor is multiplied by the intermediate factor […] to produce the eddy current density normalized to the loop current. These two factors are multiplied by the loop sensitivity kernel G0(a, -h; r′, z′) to produce the spatial map M(r′, z′). Characteristic reflected inductance can then be obtained by simply integrating with respect to r' and z' (as shown in equation (16b) above).

[0220] It is straightforward to see from equation (25b) above that the spatial map M(r',z') has real and imaginary components that depend differently on the location (r',z') of the reflected inductance source within the probe body. The relative magnitudes of the real and imaginary components of the component M(r',z') depend on the location of the source of the eddy current within the body that causes the component. From the above equation (16b), it is directly dependent on M(r',z'), so it can be directly seen that for a given eddy current source The relative magnitudes of the real and imaginary components depend on the location of the source within the body.

[0221] This relationship allows the returned signal components originating from different physiological sources within the body to be based on their The different characteristics of the real and imaginary components are distinguished relative to the amplitude.

[0222] As indicated by equation (25b), the final quantized solution of M(r′, z′) requires (Equation (17)) itself. The solution to can be determined in at least two different ways.

[0223] The first approach is more general and follows the general perturbation method. By combining equations (25), (21) and (18), the solution to equation (17) can be determined as a perturbation series:

[0224]

[0225] in, is a local perturbation of the series. The solution is general, since it applies to any circularly symmetric geometry. However, obtaining the final result is usually relatively computationally intensive.

[0226] We will now describe The second way to find the solution of . This method is computationally more efficient, but can only be applied to layered media (i.e., media comprising multiple stacked layers of media each comprising a different density, where the density is assumed to be uniform within each individual layer). The first step is to decompose the solution into right-hand (+) and left-hand (-) parts:

[0227]

[0228] In each individual uniform layer, the right and left components can be further decomposed into (r, z) separate modes of uniform inductive sensing Eq. (17). These modes are parameterized by the real wave number ξ in the radial (or transverse) r direction, yielding:

[0229]

[0230] in, is a first-order Bessel function of the first kind, and is the longitudinal wave number (the wave number in the z direction) selected in the fourth quadrant.

[0231] Amplitude Profile can be uniquely determined by the Hankel transform:

[0232]

[0233] Amplitude Profile It can be expressed in terms of the (r,z)-separation model as follows:

[0234]

[0235] in, is the amplitude profile of the Hankel transform of the left and right row domains at the layer interface, such as Figure 2 shown. is the number of longitudinal waves in layer number x, and Z1, Z2, ... Z n Corresponds to the depth in the z direction of the layer number 1, 2, ...n.

[0236] Figure 2 The diagram shows the corresponding left and right row domains on multiple stacked layers and layer interfaces. The x-axis corresponds to the distance in the z-direction, where Z1, Z2, .., Z n Corresponds to the depth in the z direction of layer 1, 2, ...n.

[0237] The incoming amplitude profile can be determined from equations (18) and (20) using the source current density:

[0238]

[0239] where I is the antenna current, a is the antenna loop radius, and h is the distance between the antenna loop and the probing medium (assuming that the surface of the probing medium is at z=0 and the medium extends in the region where z>0). This yields:

[0240]

[0241] Use with continuous and continuous The transfer matrix theory of (well known in the art, e.g. for modeling light waves in multilayer media) is known from equation (34) Expressing the above equation (32) and Figure 2 All other coefficients in are possible.

[0242] For the sake of brevity, the full set of expressions is omitted for this purpose, as these expressions are quite long. However, it can be seen that there is a computation A second approach, which works for layered media and is computationally more efficient than the first perturbation method of equation (26), is based on using equation (34) and applying transfer matrix theory to calculate the total Hankel transform vector potential of equation (32): Vector potential in real space It is also calculated via the inverse Hankel transform defined by equation (31) and equation (27), thereby enabling (via equation (25b)) the characteristic reflected inductance The spatial mapping M(r′, z′) is more clearly expressed as follows:

[0243]

[0244] in, is the total Hankel transformed vector potential calculated according to equation (32) normalized to the loop current.

[0245] As discussed above, it can be seen that the relative amplitudes of the real and imaginary components of M(r',z') (and hence its complex quantity or phase) depend on the L r The location of the signal source (r',z'). Therefore, the same applies to the reflected inductance itself which directly depends on M(r',z'). Therefore, for a given physiological signal source The relative amplitudes of the real and imaginary components depend on the location of the source within the detecting body.

[0246] Use the and therefore for The derived solution (Equation (25)) derives the induced reflection inductance component L at the antenna 12 for the signal r A spatial mapping of the expected amplitude and phase as a function of the origin location of the detected signal within the body is possible.

[0247] For those with Figure 3 The illustrated arrangement of the example system derives an example spatial mapping. An example antenna 12 with a single toroidal winding of radius d is modeled with a separation distance h from the body being probed. The probe body is modeled as three layers 42, 44, 46 comprising different dielectric conductivities σ and permittivities ε. Layer 42 is modeled as a fat layer with a depth of 10 mm, layer 44 as a spongy bone layer with a depth of 10 mm, and another layer representing an inflated lung. The antenna height h is modeled at 5 mm and the antenna diameter is modeled at 50 mm. The frequency f of the generated electromagnetic excitation signal is modeled as f=200 MHz. The corresponding normalized radial frequency where c is the speed of light and ω = 2πf is the radial frequency.

[0248] Figure 4 Shown for Figure 3 The characteristic reflected inductance of the setting The left figure shows the phase and amplitude profiles of (as expressed in equation (24)) as a function of depth (z direction) and radial position (r direction) within the medium. The right figure shows the phase (parameter) of The x-axis corresponds to the depth in the z-direction within the medium (unit: mm). The y-axis corresponds to the radial position (along the r-direction) (unit: mm).

[0249] It can be seen that the desired characteristic reflected inductance The phase of varies as a function of the depth z and radial position r in the medium. As mentioned above, the characteristic reflection inductance The references are for convenience only, where pass With the non-normalized reflected inductance L r It is related to, where l is the length of the antenna loop winding and N is the number of antenna windings. It can be seen that the phase is therefore also used as for L r varies as a function of the position within the medium.

[0250] The theory and Figure 3 The figure in the figure proves that the reflected inductance L r The complex parameter or phase of the (or characteristic reflected inductance) varies depending on the position of the source of the returned electromagnetic signal in the body, and therefore the relative amplitudes of the real and imaginary parts of the detected reflected inductance also vary depending on the position of the source of the returned electromagnetic signal in the body (see Equation 10).

[0251] It follows that the relative amplitudes of the real and imaginary parts of the detected reflected inductance can be used to distinguish and separate the returned signals from different physiological sources. Embodiments of the present invention are based on this principle.

[0252] Figure 3 An example inductive sensing system 8 is shown in schematic form in accordance with one or more embodiments. The example system will now be summarized.

[0253] The inductive sensing system 8 is used to sense electromagnetic signals returning from the body in response to applying electromagnetic excitation signals into the body.

[0254] The system includes a resonator circuit 10, which includes a loop antenna 12 and an electrically coupled capacitor 13. The capacitance of the capacitor 13 at least partially defines the natural resonant frequency of the resonator circuit (in the absence of stress or damping). When the antenna 12 is excited, it will tend to resonate naturally at the defined resonant frequency, generating an electromagnetic signal at the same frequency. Thus, selecting the capacitance of the capacitor allows at least partial tuning of the frequency of the generated electromagnetic signal.

[0255] The system 8 further comprises a signal generation module 14 adapted to excite the loop antenna to generate an electromagnetic excitation signal. The signal generation module may comprise a driver module for driving the antenna, for example, at a radial frequency ω, i.e., driving the antenna with an alternating current of frequency ω. The drive module may be or comprise, for example, an oscillator.

[0256] The signal generation module may drive the antenna and the resonator circuit with a current of a radial frequency ω, wherein an excitation signal of the radial frequency ω is required.

[0257] By exciting the resonator, a resonant current is induced to flow back and forth through the loop antenna into the capacitor. This can thereby stimulate the antenna to generate a driving alternating current through an oscillating electromagnetic signal (wave).

[0258] The same antenna that is used to sense electromagnetic signals received from the body in response is used to generate the excitation signal.

[0259] For the avoidance of doubt, an "electromagnetic excitation signal" simply means an electromagnetic signal applied to the body for the purpose of exciting or stimulating the generation of eddy currents within the body, which in turn stimulate electromagnetic signals emitted back out of the body which can be sensed by a sensing system.

[0260] By "electromagnetic signal" may generally mean electromagnetic radiation emissions or electromagnetic near-field oscillations or electromagnetic oscillations and / or electromagnetic waves.

[0261] The system 8 also includes a signal sensing module (“Signal Sensing”) 20 adapted to sense the returned signal from the body using the loop antenna 12 based on detecting changes in the electrical characteristics of the resonator circuit 10. The signal sensing module may include a signal processing or analysis module for detecting or monitoring the electrical characteristics of the current in the resonator circuit 10.

[0262] For example, the signal sensing module 20 may be adapted to monitor at least the frequency of the resonator circuit current, and the amplitude of the resonator circuit current. These properties of the current will change depending on the strength of the reflected electromagnetic signal returned from the body and detected at the antenna. Changes in these properties indicate the amplitude of the real and imaginary parts of the additional inductive component (reflected inductance Lr) induced at the resonator circuit by the received returned electromagnetic signal, respectively.

[0263] The sensing of these signal characteristics is performed simultaneously with (ie, concurrently with) the excitation of the antenna for generating the excitation signal. Thus, signal transmission and sensing are performed simultaneously.

[0264] exist Figure 5 In FIG. 1 , the signal sensing module 20 is shown connected to the resonator circuit 10 via the signal generating module 14. However, this is not necessary: ​​the signal sensing module and the signal processing module may be connected to the resonator independently.

[0265] The system 8 is adapted to perform a series of steps using the above components. The system may include a controller or microprocessor ("MPU") 56, which is adapted to perform or facilitate these steps. The exemplary microprocessor 56 is Figure 5 The example system of is shown for illustration. However, a dedicated controller or microprocessor is not necessary. In other examples, one or more of the other components of the system, such as the signal sensing module 20 and / or the signal generating module 14 can be adapted to perform the steps.

[0266] The system 8 is adapted to perform the following steps:

[0267] detecting from the sensed returned signal a measure of both the real and imaginary parts of an additional inductive component induced in the resonator circuit 10 by the received electromagnetic signal, and

[0268] Using the detected measures indicative of real and imaginary inductive components, one or more individual signal components corresponding to different respective physiological sources are extracted from the returned electromagnetic signal, the extraction being based on the relative amplitudes of the detected real and imaginary inductive components added to the resonator circuit 10 by the returned signal.

[0269] In some embodiments, the system is adapted to access a data set that stores, for signals from a plurality of different known physiological sources within the body, information indicating characteristic relative magnitudes of real and imaginary inductive components added to the resonator circuit by the signals received from the physiological sources; and

[0270] Using the detected measures indicative of real and imaginary inductive components, and based on consulting the data set, one or more individual signal components corresponding to different respective known physiological sources are extracted from the returned electromagnetic signal.

[0271] In other embodiments, different modules for separating different physiological signals may be used, for example, independent component analysis methods.

[0272] For the sake of simplicity, the additional inductance component mentioned above can be referred to as the reflected inductance L r , and corresponds to the reflected inductance discussed in detail above. The complex number L r It can be expressed as L r =L′ r +iL″ r , where L′ r is the real part, and L″ r It is the imaginary part.

[0273] The sense measure indicative of the real and imaginary parts of the additional inductive component may take different forms. However, in one advantageous set of examples, the sense measure indicative of the real part may be a sensed change in the frequency of the resonator circuit 10 current. r The sensing measure of the imaginary part of can be the sensed change in the amplitude of the resonator circuit current.

[0274] Table 1 below sets forth various electrical characteristics that change or vary depending on changes in the real and imaginary parts of the reflected inductance at the antenna. Thus, each electrical characteristic provides an indication of the reflected inductance L r A potential measure of one of the real or imaginary parts of . Different properties are measurable in different components of system 8.

[0275]

[0276] Table 1

[0277] Therefore, any measure indicative of the change in reactance and resistance, respectively, of the loop antenna can be used as a measure indicative of the real and imaginary parts of the reflected inductance.A convenient direct measure is a measure of the change in frequency and damping of the resonator circuit.

[0278] A change in the damping of the resonator circuit results in a change in the current amplitude of the resonator circuit. Therefore, L r The imaginary part of can be measured by measuring the change in the amplitude of the resonator circuit current. It is also possible to measure it by measuring the change in the amplitude of the voltage of the resonator circuit, where the relationship between the current amplitude I0 and the voltage amplitude V0 is given by I0=2πCV0, where C is the total capacitance of the resonator circuit.

[0279] These may in turn be sensed, for example, by sensing changes in frequency and amplitude of the current in the resonator circuit or in the overall circuit including the signal generating module (eg, an oscillator according to the above examples).

[0280] As described above, according to at least one set of embodiments, the system is adapted to consult a data set storing information indicating the reflected inductive component L detected at the antenna for signals originating from different known physiological sources. r The relative amplitudes of the real and imaginary parts of the characteristic.

[0281] The data set may be stored locally, for example in a memory further comprised by the system, or the data may be stored externally, for example remotely from the system 8, for example in a remote computer or server or data store. The system may be adapted to communicate with such a remote computer or data store to access information stored therein.

[0282] A data set may take the form of, for example, a database, or a table, or any other form of data structure.

[0283] According to a set of advantageous examples, the indicative information stored for each physiological source may take the form of a vector having a content indicating the reflected inductance L r The characteristics of the real and imaginary parts are the relative magnitudes of the vector elements.

[0284] These stored characteristic vectors for different physiological sources X may thus take the form:

[0285]

[0286] Among them, P X，re and P X，im Indicates the relative amplitudes of the real and imaginary components of the reflected inductance signal expected to be induced by a signal originating from a physiological source X.

[0287] These vectors may be referred to as "fingerprint vectors" because they indicate the inductive signal fingerprint of each physiological source.

[0288] Examples of different physiological sources include, by way of non-limiting example, "breathing," "pulse," "exercise," or "radial artery vessel diameter." A source may include a more specific source, such as "speaking" or "eating."

[0289] The characteristic fingerprint vectors can be determined in advance and stored statically in a database, or can be dynamically updated during the use of the system. The latter option will be described in more detail in the following article.

[0290] Initially deriving the fingerprint vector may be performed empirically, or may be performed based on, for example, modeling. The empirical derivation may be based on an iterative sampling process, tracking different possible vectors, and iteratively moving towards a vector for each physiological source that accurately results in the extraction of signal components corresponding to the physiological signal.

[0291] In the simplest case, the feature fingerprint vector The expected induced reflected inductance L at the antenna for a signal originating from a physiological source X can be taken as r_X In this case,

[0292]

[0293] In this example, the feature vector represents only the relative magnitudes of the real and imaginary components. It can be understood as effectively representing the desired complex number L in the complex plane r direction.

[0294] The system 8 is further adapted to extract one or more signal components from different known physiological sources from the sensed electromagnetic signal using the stored information (ie, the characteristic vector in this example).

[0295] The fully complex returned signal can also be expressed in vector form, based on the real and imaginary parts of the additional inductance component (reflected inductance) sensed at the antenna. This can take the form

[0296] Among them, the detected additional inductance component L r is given by:

[0297] L r (t) = L′ r (t)+iL″ r (t).

[0298] Extract signal component A from a specific physiological source X X(t) may then include the following steps: Projected onto the stored feature vector for this source Mathematically, this can be achieved by performing the sensing signal and the characteristic vector for a given physiological source A simple inner product implementation of .

[0299]

[0300] This will result in an inductance signal A X The component of (t) corresponds only to the electromagnetic signal received from a specific physiological source X, which has Relative L r Real and imaginary parts.

[0301] This projection may be done for any one or more of the physiological sources having characteristic vectors stored in the data set.

[0302] Optionally, the projection in equation (39) can be solved by first removing , for example, by adding Apply a high pass filter on:

[0303]

[0304] in, yes The slowly varying baseline component of . Note that Only The 3-dB point of this filter should preferably be low enough to ensure that relevant physiological changes are still present in the preprocessed signal middle.

[0305] Optionally, the projection in equation (39) can be further improved by applying a bandpass filter before taking the inner product:

[0306]

[0307] in, yes The 3-dB point of this filter should preferably be high enough to ensure that relevant physiological information is still present in the preprocessed signal. middle.

[0308] According to one or more sets of embodiments, the system 8 may include a module for quantifying one or more physiological parameters or phenomena corresponding to the extracted signal components. For example, the extracted L originating from the lungs rThe signal component may correspond to a physiological parameter of lung volume.The system may include an algorithm or operator or function for converting or transforming or mapping the corresponding signal component into a signal representing the corresponding physiological phenomenon (eg, lung volume).

[0309] In some cases, this may include simple multipliers applied to, for example, derived (original) signal components. In other cases, this may include more complex functions or operators that process or transform the original derived signal components into signals representing physical phenomena or parameters.

[0310] The algorithms, operators and / or functions used to perform the transformation or mapping into physiological signals may be pre-stored locally or remotely and accessed by the system 8 .

[0311] In some examples, the mapping of real physiological signals can be established as a feature (fingerprint) vector For example, instead of using a simple normalized L having vector components of the form represented in equations (36) and (37) above r vector, according to one or more examples, The components of the vector can be given by the magnitude of the true physiological parameter or phenomenon X itself divided by the real and imaginary parts of the corresponding reflected inductance.

[0312] For example, if the physiological phenomenon X is the change in lung volume, then the corresponding fingerprint vector component P X，re and P X，im will have units of liters / henry. This then means that when the fingerprint vector is compared with the sensed L r When inner products are taken between vectors (in Henrys), the resulting output signal has simply units of liters (ie, the units of the true physiological parameter).

[0313] Figure 6 A schematic representation of the processing steps performed by the system 8 in extracting different signal components is illustrated. Figure 6 Shown is a loop antenna 12 and a coupled signal processing module 14. By way of example, the generated excitation signal is shown transmitted to the chest region of the subject's body.

[0314] The returning electromagnetic signal from the body is received at the same antenna. The additional inductive component L r As a result, the received electromagnetic signal is induced at the resonator circuit.

[0315] L r The real and imaginary parts are detected and determined, and these are output as two independent signals.

[0316] As mentioned above, according to a set of examples, L rThe real part of can be measured by measuring the frequency of the resonator circuit (including the antenna 12 and the coupled capacitor), or the frequency of the signal generating module 14 (e.g., a free-running oscillator). This is because the radial frequency of the resonator depends on the real part of the reflected inductance via the following relationship:

[0317]

[0318] where L0 is the characteristic self-inductance in free space, and ω 0,0 is the radial frequency of the resonator in free space.

[0319] In other examples, L r The real part of L can be measured in different ways, wherein the frequency of the resonator circuit and / or the oscillator is kept fixed by a control circuit and wherein L r The real part of is measured via the frequency feedback signal to the control circuit. Such a circuit is often called a frequency locked loop circuit.

[0320] For example, a control circuit may be implemented to maintain the frequency of the resonator circuit at a fixed frequency. To this end, the frequency of the circuit may be measured in a continuous loop and the frequency fed back to a variable capacitance capacitor component (as a feedback signal). The variable capacitor may adjust the capacitance of the resonator circuit depending on the measured frequency in order to compensate for any sensed frequency changes and maintain the frequency at a defined value.

[0321] The frequency feedback signal on the variable capacitor can then be used to determine L r The real part of , as it contains information representing the change in the natural frequency of the resonator (and oscillator) circuit.

[0322] Frequency locked loop circuits such as these have advantages, for example, for enabling regulatory compliance, since the transmit frequency remains fixed and cannot, for example, be moved into an electromagnetic band that is prohibited by law.

[0323] As mentioned above, according to a set of examples, L r The imaginary part of can be measured by measuring the damping of the resonator. In an example, the damping of the resonator can be measured by measuring the amplitude of the signal generating module (oscillator 14 in this example). It can also be measured by measuring the voltage amplitude of the resonator circuit across the capacitor (often referred to as "slot swing" in the art), or the current amplitude of the resonant circuit.

[0324] This is because of the additional inductive resistance component R in the resonator circuit with the following magnitude r , the amplitude of the oscillator 14 depends at least in part on the imaginary part of the reflected inductance:

[0325] R r =-ωIm[Lr ] (41)

[0326] where ω is the angular frequency of the resonator circuit current. This additional resistance component causes a reduction in the amplitude of the oscillator 14, where the amplitude decreases with increasing resistance.

[0327] According to some examples, the damping of the resonator can be determined by using the same method as used to measure L r The real part of the fixed current and voltage amplitude control loop circuit discussed above operates on similar principles to the fixed frequency feedback circuit.

[0328] In particular, the resonator circuit may be coupled to a gain controller or damping controller for controlling (setting) the gain or damping of the resonator circuit. During operation, the current or voltage amplitude of the resonator circuit may be monitored and this fed back to the gain controller via a feedback signal. The gain or damping controller may be adapted to adjust the gain or damping of the resonator circuit based on the return feedback signal so as to compensate for any sensed changes in the current or voltage amplitude and maintain the current amplitude or voltage amplitude at a defined fixed level. The defined level may be adjustable (e.g., user adjustable) according to a preferred example.

[0329] The current or voltage amplitude feedback signal fed to the gain or damping controller can then be used to measure L r The imaginary part of , because it contains information representing the instantaneous current or voltage amplitude of the resonator circuit.

[0330] The gain may be controlled based on controlling the tail current of the resonator circuit (or oscillator circuit) or the collector voltage (ie, Vcc) of the gain circuit.The damping of the resonator circuit may be adjusted based on the use of a variable resistance component in the resonator circuit.

[0331] Furthermore, the use of a feedback loop has the advantage of enabling easier compliance with regulatory standards, such as requirements for a circuit to operate within certain defined voltage or current amplitude limits.

[0332] Signals from different physiological origins are extracted by the system 8 based on any of the procedures and methods outlined above.

[0333] As discussed, one approach is to use vector representations, and in which the system accesses stored characteristic (or fingerprint) vectors for different physiological sources This represents the real and imaginary L for signals originating from the corresponding sources r The characteristic relative amplitude of a component.

[0334] Fingerprint vectors for different physiological sources according to a set of examples It may be pre-stored, for example in a local memory, or remotely at a remote computer or server.

[0335] In some examples, the system may extract signal components of all physiological sources for which fingerprint vectors are stored in the data set, i.e., the system may apply all fingerprint vectors stored in the data set to the complex L sensed by the system 8. r Signal.

[0336] According to other examples, the system may include a user interface, and the system may be adapted to allow selection by the user of a set of one or more of the physiological sources for which fingerprint vectors stored in the data set exist.

[0337] According to one or more examples, the system can be adapted to receive user input indicating one or more patient demographic parameters (e.g., weight, age, height), and wherein the system is adapted to determine one or more relevant physiological sources of derived signal components based on the input patient parameters. The relevant fingerprint vectors corresponding to the determined physiological sources can then be retrieved from the data set and used by the system 8 to extract the corresponding signal components.

[0338] In some cases, more than one vector may be stored in a data set for a given physiological source or one or more of the stored vectors may include patient-related parameters that may be adjusted. A user interface may be used to receive user input indicating one or more of these patient parameters, and these are used to tune or adjust the vectors, or select the most relevant fingerprint vector to apply in a given scenario. This option is based on the recognition that certain patient demographic parameters may be expected to affect the optimal setting for the fingerprint vector (e.g., the thickness of the fat layer may affect the direction of the fingerprint vector in the complex plane).

[0339] According to one or more examples, system 8 may include a display or other sensory output device for conveying the derived signal components to a user of the system.

[0340] As described above, the system may include a module for further processing the extracted signal components to derive from them corresponding signals for the real physiological phenomena or parameters to which the extracted signal components correspond. Examples may include, for example, a respiration rate signal, a pulse rate signal, a blood vessel diameter, blood pressure, or any other physiological parameter or measurement result or signal. The extracted parameters or measurement signals may optionally be displayed using a display output.

[0341] According to one or more examples, the system 8 may also be adapted to perform a method for adjusting or improving the stored fingerprint vectors based on the characteristics or features of previously extracted signal components for a given patient. For example, different patients have different geometries associated with their bodies, and these different geometries mean that a single set of stored characteristic fingerprint vectors will not necessarily be optimally suited for each individual patient.

[0342] For example, each patient has a slightly different internal and external body shape (eg, thickness of fat layer, muscle mass, position of the heart, lung inflation, work of breathing).

[0343] Thus, the system 8 can be configured with a learning algorithm that allows for initial learning and / or updating of fingerprint vectors (e.g., automatically and dynamically) based on the particular patient at hand. The optimal fingerprint vector can be identified for a particular patient based on the process. This can involve the application of a learning or adjustment algorithm adapted to trial different fingerprint vectors, which results in an output score of the likelihood of the corresponding extracted signal component (e.g., pulse) originating from a certain target physiological origin.

[0344] The preferred method for users to dynamically learn or update fingerprint vectors would be to apply one of the various known methods in the category known as Independent Component Analysis (ICA). Methods in this category are never used in the context of inductive sensing. The inventors have found that these methods are extremely effective in dynamically identifying the best fingerprint vectors.

[0345] In essence, the ICA algorithm assumes that the detection signal at the resonator circuit (of dimension n>1, i.e., multiple signals detected simultaneously) is a (potential, yet unknown, to be determined) linear mixture or combination of dimension m≤n. To further explain this, now consider an example where there are two source signals a1(t) and a2(t) and two detection signals s1(t) and s2(t). These two are related via the following:

[0346]

[0347] or in matrix notation

[0348]

[0349] Among them, the weight vector matrix W Describes how the sources a1(t) and a2(t) are mapped or transformed into the detected signals s1(t) and s2(t). Independent component analysis involves To determine the matrix W , so that the unknown potential sources a1(t) and a2(t) can be obtained based on the measurement results of signals s1(t) and s2(t).

[0350] In particular, derive the inverse of the weight vector matrix W -1 can be most useful because this more directly describes how to detect a signal based on Building unknown sources Right now,

[0351]

[0352] Therefore, independent component analysis can be used to detect signals based on Get the matrix W -1 In order to calculate the potential source signal

[0353] This ICA method can be directly applied to the determination of fingerprint vectors for embodiments of the present invention. In particular, there is a direct match between the forms of equation (39) and equation (44). The detected signals s1(t) and s2(t) can be given, for example, by the measured frequency and amplitude of the inductive sensor (or the real and imaginary parts of the reflected inductance, respectively, or the reactance and resistance of the loop antenna, respectively - see Table 1 above for different options), and the potential sources to be reconstructed, the sources a1(t) and a2(t) are different physiological signal sources picked up by the antenna, such as breathing and pulse (or breathing and movement, or pulse and movement).

[0354] Although the above equation only shows a matrix with two rows, this is for illustration only and the matrix may have more than two rows, ie, n, for extracting more than two physiological source signals, ie, n source signals a(t).

[0355] Furthermore, when applied to embodiments of the present invention, the matrix W -1 The rows in correspond to the fingerprint vectors for the inductive sensing method discussed above with respect to equation (39). Therefore, the fingerprint vector described in equation (39) is a matrix W -1 From a set of measured signals The number of source signals to be extracted is the matrix W -1 Using the nomenclature of the “fingerprint” vector adopted above, the inverse matrix W -1 Can be described as a "fingerprint matrix".

[0356] Therefore, we can see that independent component analysis involves deriving the matrix W -1 A class of signal processing methods, and this can be advantageously applied to the identification and / or extraction of a set of sensing inductive signals from two or more physiological signal sources included in the sensing signal. Therefore, according to an embodiment of the present invention, this can be used to dynamically determine an unknown fingerprint vector.

[0357] The method may be applied as an initial procedure, such as an initial learning or calibration procedure, before the inductive sensing system is actually used to learn fingerprint vectors (or more broadly, the relative magnitudes of the real and imaginary components of reflected inductance) associated with different physiological sources for the particular patient at hand. It may additionally or alternatively be used to update the fingerprint vectors (or relative magnitudes).

[0358] The method of independent component analysis has never been applied in the field of inductive sensing. It is a non-obvious development that despite 50 years of research on inductive sensors, it was never recognized that frequency and damping can be considered as two independent signals whose relative values ​​can be used to identify a specific physiological source, i.e., its originator.

[0359] A popular approach to ICA is called Fast ICA, for example in the paper [ A., & Oja, E. (2000). Independent component analysis: algorithms and applications. Neural Networks, 13, 411-430.] It uses the potential source signal is an efficient approximation of the entropy of , and it has the optimization W -1 In order to achieve an efficient iterative method to minimize entropy (maximize non-Gaussianity).

[0360] Figure 7 Example results of applying FastICA for detecting fingerprint vectors according to one or more embodiments of the present invention are shown. The fingerprint vectors are determined based on applying FastICA to the detected signals of the amplitude and frequency of the resonator circuit current (as a function of time). The identified fingerprint vectors effectively decompose the detected composite (mixed) signal (amplitude and frequency) into potential components (or sources) corresponding to signals generated by respiration and pulse (for example).

[0361] exist Figure 7 , detected signals 102, 104 are obtained from an inductive sensor placed at the sternum of a volunteer breathing relatively rapidly (approximately 36 breaths per minute - see airway volume signal 100). Both the inductive sensor amplitude (signal 102) and frequency (signal 104) are recorded. Potential physiological signal sources of pulse and respiration (signals 108 and 110, respectively) are extracted or reconstructed based on fingerprint vectors determined using the FastCA method discussed above (applied to the detected inductive sensing signals 102, 104).

[0362] Therefore, the ICA method is very effective in isolating different physiological sources in inductive sensors where both amplitude and frequency are recorded simultaneously.

[0363] It should be noted that although the above example is described in terms of detecting fingerprint vectors for different physiological sources, it can be understood to be more broadly applicable to detecting the relative magnitudes of the real and imaginary components of reflected inductance for different physiological sources. The use of fingerprint vectors is one representation of the general concept discussed in this application.

[0364] More broadly, the learning process for fingerprint vectors can be based on the temporary use of any alternative method for separating sensory signal components from different physiological sources. According to some examples, these separated signals using alternative methods can then be analyzed to determine their corresponding real and imaginary L r Components, and the relative amplitudes of these components for a given signal source are then stored in the data set in the form of, for example, an updated fingerprint vector. Alternatively, it may directly lead to the determination of a fingerprint vector, as in the above ICA method.

[0365] Alternative methods may be any method currently known in the art for separating or distinguishing signal components from different physiological sources, such as based on different frequency compositions / spectra of the signal components, different amplitudes / amplitudes of the signal components, different quality measures of the signal components. By way of non-limiting example, separation of signals based on these methods may employ the use of any of the following: wavelets, signal frequency analysis, and / or signal quality measures. ICA is an example method for determining different signal components or fingerprint vectors, and other methods may be alternatively used in other examples.

[0366] In more detail, methods exist in the prior art for signal processing for separating sensed signal components having different physiological origins based on certain a priori known characteristics associated with the signals originating from those sources. For example, the breathing rate is (most often) slower than the pulse rate, and therefore the corresponding induced signals at the antenna will have corresponding different frequencies. Therefore, in some examples, these signal components can be separated from each other based on frequency analysis of the overall composite signal sensed at the antenna.

[0367] Another existing method is based on the analysis of the shape of the signal component waveform. For example, a typical breathing signal has a different shape than a pulse rate signal. Waveform-based methods can often make use of wavelets.

[0368] As discussed above, these methods employed in the prior art can be misleading or unreliable, particularly in the presence of certain pathologies that typically cause different signals to appear closer to each other, such as when the respiratory rate is very high, or the heart rate is very low. The use of fingerprint vectors is more reliable across all clinical scenarios.

[0369] In an example, a learning process using the known classical signal separation methods discussed above can be performed in certain scenarios that are known to be clinically simple and do not cause confusion or errors in the separation of the signals. Once the learning process has updated the fingerprint vector, the real and imaginary L r Use of the relative amplitudes of the components may then be employed with methods according to embodiments of the invention.

[0370] Simple clinical scenarios (during which the learning process can be employed) can include (by way of non-limiting example) any of the following: situations where motion artifacts are low (e.g., static patients; situations where the patient is not apneic; situations where these are not arrhythmic; situations where the respiratory rate and pulse rate are well separated; situations where the pulse waveform is not abnormal; situations where the respiratory signal has a normal waveform).

[0371] Examples of non-simple clinical situations during which the implementation of the learning process of classical signal separation methods can be avoided and instead the improved methods of embodiments of the present invention can be used include, for example, one or more of the following:

[0372] - Situations in which there are episodes of apnea. Here, intermittently there is no breathing signal at all. In particular, for example in the case of neonatal patients, the pulse rate can also drop during apnea. During the absence of a breathing signal, classical signal separation algorithms can misinterpret the pulse rate for the breathing rate.

[0373] - Situations where the breathing rate and pulse rate are similar. In this case, frequency analysis methods may fail to distinguish the two signals.

[0374] - Situations where there are motion artifacts (e.g. movement of the patient or antenna during sensing). These motion artifacts cause noise in the signal that may interfere with reliable signal separation.

[0375] In some examples, common machine learning techniques may also be applied to improve or update the fingerprint vectors. For example, a machine learning algorithm trained to separate signal components from different physiological sources based on the classical methods discussed (e.g., frequency or waveform analysis) may be employed.

[0376] It should be noted that although in the examples outlined above, L for different physiological sources is indicated r The information about the relative magnitudes of the real and imaginary components is stored in the form of vector , but this is not necessary. In other examples, the relative magnitudes themselves may simply be stored in numerical form. In other examples, other measures involving or derived from or otherwise indicating the relative magnitudes of the real and imaginary components may be used, i.e., L r A proxy for the relative magnitudes of the real and imaginary components of .

[0377] In some examples, one may simply store the complex L expected to be induced at the antenna by a given physiological source. r The complex phase or parameter of the component.

[0378] According to a set of advantageous embodiments, the system 8 is further configured to adjust the frequency of the generated excitation signal in order to increase the characteristic vectors for different physiological sources in the body. The theory underlying this option will now be outlined in detail.

[0379] It can be seen from the resonator circuit derived above (characteristic reflected inductance ) is derived from the determined expression of the additional inductive component received at the , the quantity or phase of the induced inductive component of the signal originating from different areas of the body varies depending on the frequency of the applied excitation signal. Thus, by changing the frequency of the applied excitation signal, the corresponding The phase or magnitude of the signal components will change.

[0380] In particular, the induction at the antenna The depth dependence of the phase of the signal (in the z direction) is strongly frequency dependent. Therefore, by adjusting the frequency of the applied excitation signal, the reflected inductance originating from different locations in the body is adjusted. The relative phase (ie parameter) of the signals is possible.

[0381] This means that the complex The relative directions of the signals can be varied by adjusting the frequency of the applied excitation signal. An equivalent way of expressing this is to use the complex The relative amplitudes of the real and imaginary components of the signal vary depending on the frequency of the applied excitation signal.

[0382] Additionally, adjusting the frequency of the applied excitation signal may also result in a resulting change in the relative phase or magnitude of signals originating from different tissue types in the body (even where those tissue types are at the same general location in the body).

[0383] This arises from the fact that the phase of the induced eddy currents in the body can often make relatively large jumps or discontinuities across tissue type boundaries. This can be seen, for example, from the loss tangent δ≡-Arg(-iωσ+ω 2 ε0χ e ) affects the relative phase of the generated signal. At low frequencies, all tissues will have δ = 90°. However, at a certain tissue-dependent critical frequency At δ = 0°, the loss tangent will move closer to δ = 0°. Since each tissue has a different conductivity and permittivity, the critical frequency at which the loss tangent changes to different values ​​is strongly tissue-dependent. Therefore, by changing the frequency, the relative phase of the signals from different tissues in the body can be tuned.

[0384] Figure 8 and Fig. 9 Figure 2 shows the induced reflected inductance component at the antenna for a given physiological source. The frequency dependence of the phase.

[0385] Figure 8 Schematic illustration of two characteristic vectors for two different physiological sources in the body By way of example, two vectors are shown corresponding to the signals for respiration and for the heart pulse. The x-axis corresponds to The real part of the y-axis corresponds to The imaginary part of .

[0386] The starting point of each vector corresponds to the characteristic reflected inductance at the beginning of the corresponding physiological cycle: the deflated lung state and the systolic heart state for the respiration and pulse vectors, respectively. The ending point corresponds to the end of the corresponding physiological process. Therefore, the difference between the two in each vector provides a representation of the corresponding physiological process.

[0387] In each case, the ending point minus the starting point for each vector may be understood to correspond to a fingerprint vector.

[0388] Fig. 9 The different characteristics of different lung inflation levels of the example patient are shown in Figure 2. r vector Each row represents the vector of lung inflation for a different frequency of the applied electromagnetic excitation signal. The x-axis corresponds to the reflected inductance L r The real part of the y-axis corresponds to the reflected inductance L r The arrows show the change in reflected inductance when the lung is inflated: when the lung is deflated, the starting point is L r , and when the lungs are inflated, the end point is L r . It can be seen that the starting point varies for different frequencies and the ending point varies for different frequencies. Therefore, L at the beginning and end of the breathing process r The relative amplitudes of the real and imaginary components change depending on the frequency.

[0389] It can also be seen from the figure that L in the complex plane r The direction of the vector varies depending on the selected frequency of the applied excitation signal.

[0390] By adjusting the frequency, the corresponding L can be changed rThe direction of the fingerprint vector.

[0391] The fingerprint vector may be understood to be given either by (the relative magnitude of the real and imaginary components of) one or the other of the start or end points or by the change in the real and imaginary components over the course of lung inflation.

[0392] Thus, according to an advantageous set of embodiments, the frequency of the applied excitation signal can be adjusted so as to better orthogonalize the induced reflected inductance L at the antenna for different physiological sources within the body. r Characteristic vector of the component These characteristics L r Greater orthogonality of the signals results in a more reliable and robust extraction and separation of signal components for different physiological sources. It is therefore advantageous to adjust the applied signal frequencies to maximize the orthogonality between the characteristic vectors.

[0393] The process Fig.10 The schematic diagram is shown in Fig.10 The relative directions of two example characteristic fingerprint vectors at a first frequency ("Frequency 1") and a second frequency ("Frequency 2") are illustrated. As shown, at the first frequency, the vectors are at a first relative phase separated from each other. By adjusting the frequency to the second frequency, the vectors can be caused to be orthogonal to each other in the complex plane.

[0394] To achieve this functionality, the system 8 may be configured with a frequency adjustment algorithm that is configured to detect, for example, in real time, the complex L sensed at the resonator circuit 10 corresponding to different physiological sources in the body. r The degree of orthogonality in the complex plane between the component vectors.

[0395] The feedback loop can be implemented as a signal generating module (e.g., an oscillator) whereby the frequency of the generated electromagnetic excitation signal is adjusted to improve orthogonality with frequency adjustment guided or informed by the frequency feedback signal. The algorithm thereby adjusts the free space frequency of the oscillator to better orthogonalize the characteristic vectors of the reflected inductance signals for different physiological sources.

[0396] Frequency adjustment may be based on the temporary use of an alternative method for separating sensed signal components from different physiological sources. These separated signals using the alternative method may then be analyzed to determine their corresponding real and imaginary L r components, and thus determine the corresponding fingerprint vector for each signal component.

[0397] These alternative methods can be used in conjunction with frequency adjustment to detect how the fingerprint vector direction changes as the frequency is adjusted. The degree of orthogonality between a new set of fingerprint vectors for all received signal components can then be determined at each new frequency. This can be used to inform how to further adjust the frequency to maximize orthogonality. A feedback loop can be implemented.

[0398] Additionally or alternatively, in order to adjust the frequency to increase orthogonality, according to one or more embodiments, the resonator circuit antenna can be driven (excited) at multiple frequencies quasi-simultaneously, for example 2 or 3 frequencies in a quasi-simultaneous manner. This is done, for example, by multiplexing the frequencies, i.e., continuously alternating or cycling the frequencies between a set of two or more frequencies, for example, exciting the loop at a first frequency, and at a second frequency, etc., and then returning to the first frequency. This allows quasi-simultaneous measurements of the inductive sensor signal at multiple different excitation frequencies.

[0399] This frequency switching is preferably performed at a high frequency, for example, at a sufficiently high frequency that the Nyquist sampling frequency for each of the recording frequencies is greater than the information bandwidth of the underlying physiological source. The information bandwidth of a typical source signal is at least about 5 Hz for pulse and respiration. Thus, in a preferred embodiment, the minimum frequency at which all frequency selection cycles should repeat themselves is approximately 10 Hz.

[0400] Thus, switching the frequency of the drive signal to the antenna between two or more frequency settings (e.g., completing a full frequency cycle of multiple frequencies to the nearest 1 / 10 Hz = 0.1 seconds) will allow quasi-simultaneous measurements of the inductive sensor signal at multiple frequencies. For signals representing tissue motion, the information bandwidth may be greater, and in this case, a higher switching frequency may be preferred. The method of rapidly switching between multiple different frequency settings may be referred to as time multiplexing of the drive or excitation frequency of the resonator circuit or antenna.

[0401] Recording inductive sensing signals at multiple frequencies in a quasi-simultaneous manner may be particularly beneficial in the context of processes for dynamically determining or learning or updating fingerprint vectors (or more broadly the relative real and imaginary components of reflected inductance for different physiological sources). This is because these methods may only perform poorly when the number of potential physiological sources exceeds the number of detected signals. This may often be the case for inductive sensing, as sources such as "motion", "pulse", "respiratory volume", and "respiratory effort" may all be present at the same time. In this example, four different sources would be present, which is more than the 2 signals (e.g., amplitude and frequency of the resonator circuit current) that are typically recorded for an inductive sensor operating at a single frequency.

[0402] This problem can be solved at least in part by measuring the inductive sensor received signals or readings at multiple frequencies in a quasi-simultaneous manner as described above. By way of example, the driving frequency of the antenna can be alternated or multiplexed between two frequencies (e.g., 150 MHz and 250 MHz). If this is done using two sensing signals (e.g., of frequency and amplitude) measured for each frequency of each duty cycle, this results in a total of four independent detection measurement signals for each multiplexed duty cycle: "resonator amplitude at 150 MHz", "resonator frequency at 150 MHz", "resonator amplitude at 250 MHz", "resonator frequency at 250 MHz". Each of these is a (statistically) independent signal, and therefore the collection of each additional signal adds real additional information. Its independence can be shown and understood from the physics of induction: different frequencies will induce eddy currents in bodies of different geometries and / or positions, resulting in statistically independent returned signals. The fact that these signals are independent allows for an improved determination of four independent fingerprint vectors.

[0403] By way of example, the inductive sensor may be operated in a multiplexed manner using two multiplexed frequencies: 150 MHz and 250 MHz. As discussed above, this allows, for example, four detected signals to be measured. The FastICA method (see above description) may then be applied to the set of four signals to obtain corresponding four fingerprint vectors associated with different physiological signal sources.

[0404] Preferably, the number of detection signals should match or exceed the number of physiological signal sources received at the antenna at the same time. For example, this ensures that the FastICA method works most accurately. Therefore, where frequency reuse is used, the number of reused frequencies in each duty cycle is preferably equal to the number of physiological signal sources divided by two.

[0405] Although ICA is discussed above, alternative methods for separating signal components can be any method currently known in the art for separating or distinguishing signal components from different physiological sources, such as based on different frequency compositions / spectra of signal components, different amplitudes / amplitudes of signal components, and different quality measures of signal components. By way of non-limiting example, separation of signals based on these methods can employ the use of any of the following: wavelets, signal frequency analysis, and / or signal quality measures. These alternative methods and their use for detecting fingerprint vectors of signals from different physiological sources have been discussed in detail above.

[0406] The stored fingerprint vectors in the dataset can be updated simultaneously with the frequency being adjusted (and thus consistent with the real physical L r vector changes simultaneously), so that the extraction of different signal components remains accurate.

[0407] To facilitate the above frequency adjustment feature, according to one set of examples, a system 8 may be provided having a signal generating module 14 including a frequency adjustable oscillator circuit. This may be achieved in many different ways.

[0408] According to one approach, banks of capacitors may be provided, each with a high-Q switch that facilitates connection or disconnection of the capacitor to the resonator. By controlling the switches, the total capacitance of the resonator may be tuned, and the frequency of the generated excitation signal may also be tuned.

[0409] According to a particular set of examples, the system can be configured in at least one control mode to switch the frequency of the signal generating module (e.g., oscillator) between different frequencies at high speed. The switching can be done at a frequency between, for example, about 5-150 Hz. In this way, all different physiological sources in the body are detected at multiple different frequencies within a short time space (e.g., quasi-simultaneously). In this way, L r Signal components can be sensed at many different frequencies providing a greater amount of information. By combining fingerprint vectors from many frequencies, signal separation can be rendered more reliable.

[0410] Embodiments have been discussed above which involve accessing a data set which stores, for signals from a plurality of different known physiological sources within the body, information indicating characteristic relative magnitudes of real and imaginary inductive components added to the resonator circuit by the signals received from the physiological sources. However, the use of such a data set is not essential.

[0411] By way of example, according to another set of embodiments, the extraction of individual signal components can be based on applying an independent component analysis procedure to a detection measure indicative of real and imaginary inductive components. In particular, the independent component analysis can be applied to the signal sensed at the antenna in an "online" or real-time manner, such that different physiological signal components are extracted from the sensed measurement signal in real-time using ICA. In this method, no stored reference data sets of relative real and imaginary signal components for different physiological sources are required.

[0412] The independent component analysis (ICA) process applied to the received measurement signal has been described above with respect to the determination of fingerprint vectors for different physiological signal sources. However, the same ICA process can be applied to the incoming signal in an online or real-time manner to determine and extract different component physiological signals without the need for a fingerprint vector or at least without the need to store the fingerprint vector. The ICA method can be simply applied continuously to the sensed signal (e.g., s1, s2) of the measured additional real and imaginary inductance components added to the resonator circuit as a function of time. The weight vector matrix W can be derived in real time (see the above description), and this is used to extract physiological signal components a1, a2, ..., etc. The fingerprint vector can be derived and applied in real time (but not stored), or the weight vector matrix can be directly applied without the fingerprint vector being explicitly extracted. Since the ICA process has been described in detail above, it will not be described again here. The same process as discussed above can be applied in real time to the measurement signal and its received and real-time derived component physiological signals.

[0413] According to another aspect of the present invention, an inductive sensing method is provided. Fig.11 An example method is shown.

[0414] The method 70 is based on sensing electromagnetic signals returning from the body in response to propagating electromagnetic excitation signals into the body.

[0415] The method comprises propagating 72 an electromagnetic excitation signal into a body using a resonator circuit comprising a loop antenna and an electrical coupling capacitor.

[0416] The method further comprises sensing 74 the returned signal from the body using a loop antenna based on detecting a change in an electrical characteristic of the resonator circuit.

[0417] The method further comprises the following steps for extracting signal components from different physiological sources from the sensed returned signal:

[0418] The sensed returned signal detection 76 indicates a measure of both the real and imaginary parts of the additional inductive component induced in the resonator circuit by the received electromagnetic signal,

[0419] One or more individual signal components from different known physiological sources are extracted 80 from the returned electromagnetic signal based on the use of detected metrics indicative of real and imaginary inductive components, the extraction being based on the relative amplitudes of said detected real and imaginary inductive components added to the resonator circuit by the returned signal.

[0420] According to one set of embodiments, the method comprises:

[0421] accessing 78 a data set storing, for signals from a plurality of different known physiological sources within the body, information indicating characteristic relative magnitudes of real and imaginary inductive components added to a resonator circuit by signals received from the physiological sources; and

[0422] Based on the use of the detection measure indicative of real and imaginary inductive components and based on consulting the data set, one or more individual signal components from different known physiological sources are extracted 80 from the returned electromagnetic signal.

[0423] Embodiments of the present invention relate to an inductive sensing system and an inductive sensing method.

[0424] Inductive sensors have the potential to become a highly valuable technology in the field of patient monitoring, particularly because the biometric signal is relatively motion tolerant (which means that it is relatively unaffected by movement of the patient or the measuring sensor). Possible applications advantageously include wearable patient monitors for motion tolerant respiration measurements, contactless patient monitoring, and spot check measurements.

[0425] In a preferred embodiment, an antenna with a single loop (single winding) is provided, but this is not essential. A single loop winding provides benefits in terms of signal strength due to reduced parasitic capacitive coupling between the windings. The same single loop antenna is used for both generating the excitation signal and sensing the returned signal. The returned signal is sensed simultaneously with the signal generation by detecting changes in the electrical characteristics of the resonator circuit when the signal is generated (and the returned signal is received at the antenna). The combined sensing and generating antenna allows for the provision of a high quality sensed signal as well as providing the benefits of low cost, low complexity, and low power.

[0426] Embodiments of the invention are based on sensing the real and imaginary parts of the additional inductive component (reflected inductance) induced by the received (returned) electromagnetic signal from the body at the antenna 12. These parts can be detected by sensing changes in the frequency of the resonator circuit current and changes in the amplitude of the signal generator (e.g. oscillator) or resonator circuit current.

[0427] A wide variety of potential applications exist for embodiments of the present invention. Applications include, by way of non-limiting example: patient monitoring; telemetry; spot check monitoring; implementation in wearable devices (e.g., chest patches or wrist-worn devices); neonatal monitoring; sleep monitoring; obstetric monitoring; use of mattress-based sensors.

[0428] As described above, certain embodiments may utilize a controller. The controller can be implemented in a variety of ways using software and / or hardware to perform the various functions required. A processor is an example of a controller that employs one or more microprocessors that can be programmed using software (e.g., microcode) to perform the desired functions. However, a controller may be implemented with or without a processor, and may also be implemented as a combination of dedicated hardware that performs some functions and a processor (e.g., one or more programmed microprocessors and associated circuits) that performs other functions.

[0429] Examples of controller components that may be employed in various embodiments of the present disclosure include, but are not limited to, conventional microprocessors, application specific integrated circuits (ASICs), and field programmable gate arrays (FPGAs).

[0430] In various embodiments, a processor or controller may be associated with one or more storage media, such as volatile and non-volatile computer memory, such as RAM, PROM, EPROM, and EEPROM. The storage media may be encoded with one or more programs that perform the desired functions when run on one or more processors and / or controllers. The various storage media may be fixed within a processor or controller, or may be transferable so that one or more programs stored thereon may be loaded into a processor or controller.

[0431] Those skilled in the art can understand and implement variations of the disclosed embodiments when practicing the claimed invention by studying the drawings, the disclosure and the claims. In the claims, the word "comprising" does not exclude other elements or steps, and the words "one" or "an" do not exclude multiple. A single processor or other unit can implement the functions of several items recorded in the claims. Although certain measures are recorded in mutually different dependent claims, this does not indicate that a combination of these measures cannot be used to advantage. The computer program may be stored / distributed on a suitable medium, such as an optical storage medium or solid-state medium supplied together with other hardware or as part of other hardware, but may also be distributed in other forms, such as via the Internet or other wired or wireless telecommunications systems. Any reference numerals in the claims should not be interpreted as limiting the scope.

Claims

1. An inductive sensing system (8) for sensing an electromagnetic signal returning from a body in response to applying an electromagnetic excitation signal to the body, the system comprising: A resonator circuit (10) comprising a loop antenna (12); a signal generating module (14) adapted to excite the loop antenna to generate the electromagnetic excitation signal, a signal sensing module (20) adapted to sense electromagnetic signals returning from the body using the loop antenna based on detecting changes in electrical characteristics of the resonator circuit; as well as A controller, wherein the controller is adapted to: detecting, from the sensed returned electromagnetic signal, a measure indicative of both the real and imaginary parts of an additional inductive component induced in the resonator circuit by the received electromagnetic signal, and Using detected measures indicative of real and imaginary inductive components, one or more individual signal components corresponding to different respective physiological sources are extracted from the returned electromagnetic signal, the extraction being based on relative amplitudes of the detected real and imaginary inductive components added to the resonator circuit by the returned electromagnetic signal.

2. The inductive sensing system according to claim 1, wherein: The extraction of the individual signal components is based on applying an independent component analysis procedure to the detected measures indicative of real and imaginary inductive components.

3. The inductive sensing system according to claim 1, wherein: The system is suitable for: accessing a data set storing, for signals from a plurality of different known physiological sources within the body, information indicating characteristic relative magnitudes of real and imaginary inductive components added to a resonator circuit by signals received from the physiological sources; and Using the detected measure indicative of real and imaginary inductive components and based on consulting the data set, the one or more individual signal components corresponding to different respective known physiological sources are extracted from the returned electromagnetic signal.

4. The inductive sensing system (8) according to claim 3, wherein: The stored information for each physiological source takes the form of a vector or point in the complex plane that represents the additional inductive component added to the resonator circuit by the signal received from the respective physiological source.

5. The inductive sensing system according to claim 4, wherein: The relative magnitudes of the characteristics represented by the stored information are weighted to provide a multiplicative mapping between detected measures indicative of real and imaginary parts of additional sensed inductive components and extracted signal components for a given physiological source.

6. An inductive sensing system (8) according to any one of claims 4 to 5, wherein: The stored information for each physiological source includes a vector having the following: a first vector component indicative of a ratio between a physiological parameter relating to the physiological source and a measure of the real part indicative of an additional inductive component added to the resonator circuit by a signal received from the physiological source, and A second vector component indicative of a ratio between a physiological parameter related to the physiological source and a measure of the imaginary part indicative of an additional inductive component added to the resonator circuit by a signal received from the physiological source.

7. An inductive sensing system (8) according to any one of claims 4 to 5, wherein: Extracting a given signal component comprises taking the inner product of the corresponding stored vector and a further vector representing said measure indicative of the real and imaginary parts of the sensed returned electromagnetic signal, and Optionally wherein a low pass filter is applied to components of said further vector before performing said inner product so as to remove a baseline and / or a band pass filter is applied to said further vector before performing said inner product.

8. An inductive sensing system (8) according to any one of claims 1 to 5, wherein: the signal sensing module being adapted to detect the measure indicative of the real part of the additional inductive component based on detecting a change in the reactance of the resonator circuit or based on detecting a change in the frequency of the current in the resonator circuit; and / or The signal sensing module is adapted to detect the measure indicative of the imaginary part of the additional inductive component based on detecting a change in resistance of the loop antenna or based on detecting a change in amplitude of a current in the resonator circuit.

9. An inductive sensing system (8) according to any one of claims 4 to 5, wherein: The signal generation module is adapted to excite the resonator to generate an excitation signal having a radial frequency ω, and wherein the controller is adapted to perform, in at least one control mode, an adjustment of the radial frequency from a first frequency to a second frequency, thereby increasing the degree of orthogonality between respective vectors formed in the complex plane by means of an additional inductive inductive component added to the resonator circuit by the signals received from the different respective physiological sources.

10. The inductive sensing system (8) according to claim 9, wherein: The system is further adapted to update the stored information for each physiological source to reflect changes induced in the real and imaginary components of the additional induced inductive components for different physiological sources by adjustment of the radial frequency.

11. The inductive sensing system (8) according to any one of claims 3 to 5, wherein: The system is adapted to perform a learning procedure comprising determining or updating the stored information for each of the physiological sources.

12. The inductive sensing system of claim 11, wherein: The learning process includes performing an independent component analysis (ICA) process on the sensed returned electromagnetic signals.

13. The inductive sensing system of claim 12, wherein: The learning procedure includes performing an ICA procedure on signals representing the real and imaginary parts of the additional inductive component added to the resonator circuit by the returned electromagnetic signal.

14. The inductive sensing system according to any one of claims 1 to 5, wherein: The system is adapted to drive the loop antenna with a plurality of drive frequencies in a time multiplexed manner and to sense the returned electromagnetic signal at the loop antenna at each of the drive frequencies.

15. The inductive sensing system (8) according to any one of claims 1 to 5, wherein: The system includes a signal processing module configured to process the extracted one or more signal components and to derive one or more physiological parameter measurements based on the one or more signal components.

16. An inductive sensing method (70) based on sensing an electromagnetic signal returning from a body in response to applying an electromagnetic excitation signal to the body, the method comprising: applying (72) an electromagnetic excitation signal to the body using a resonator circuit, the resonator circuit comprising a loop antenna; sensing (74) electromagnetic signals returning from the body using the loop antenna based on detecting changes in electrical characteristics of the resonator circuit; detecting (76) from the sensed returned electromagnetic signal a measure indicative of both the real and imaginary parts of an additional inductive component induced in the resonator circuit by the received electromagnetic signal; and Based on the use of detected measures indicative of real and imaginary inductive components, one or more individual signal components from different known physiological sources are extracted (80) from the returned electromagnetic signal, the extraction being based on the relative amplitudes of the detected real and imaginary inductive components added to the resonator circuit by the returned electromagnetic signal, wherein a controller is adapted to perform the detecting (76) and the extracting (80).

Images