Systems for processing respiratory rate

Abstract

In one aspect, a computer-implemented method includes: receiving a signal corresponding to the impedance of a patient's chest; filtering the signal using one or more filters, the one or more filters reducing noise and centering the signal around a zero baseline; adjusting the amplitude of the filtered signal based on a threshold; separating the amplitude-adjusted signal into component signals, wherein each of the component signals represents a frequency-limited band; detecting fractional phase transitions of the component signals in the component signals; selecting a main component signal from the component signals based on the amplitude of the component signals at a time corresponding to the detected fractional phase transition; determining a frequency of the main component signal at a time corresponding to the detected fractional phase transition; and determining the patient's breathing rate based on the determined frequency.

Description

[0001] CROSS-REFERENCE TO RELATED APPLICATIONS

[0002] This application claims the benefit of U.S. Provisional Application No. 62 / 883,339, filed on August 6, 2019, and U.S. Utility Application No. 16 / 922,616, filed on July 7, 2020. The contents of the foregoing applications are incorporated herein by reference in their entirety. Background Art

[0003] The present disclosure relates to respiratory rate processing.

[0004] Respiratory rate is the rate at which breathing occurs. Typically, respiratory rate is measured when the patient is at rest, and is measured in breaths per minute. Respiratory rate may increase with fever, illness, or other medical conditions, and is widely used to monitor the physiology of sick patients. Respiratory rate is measured to identify changes in physiology and other vital signs. Summary of the invention

[0005] The present disclosure describes systems and techniques related to respiratory rate processing. The described systems and techniques can be used to estimate the average respiratory rate based on periodic changes detected in a raw continuous impedance signal received from a sensor. Respiratory rate processing includes pre-conditioning the impedance signal, followed by rate estimation using wavelet groups and quarter phase (QP) domain periodic tracking. Signal pre-conditioning includes pre-filtering to reduce the effects of out-of-band artifacts and generally improve the signal-to-noise ratio (SNR), followed by adaptive limiting of the filtered signal through dynamic tracking of local amplitude statistics to suppress signal spikes and transient increases in signal amplitude. Localized loss of signal amplitude may occur due to complete signal loss, such as due to lead detachment, or due to a lack of impedance change, such as due to breath holding or apnea. Under these conditions, the respiratory rate may not be well-defined. The tracked amplitude statistics are themselves estimates of the local amplitude of the filtered signal and are therefore used as a basis for conforming the respiratory rate signal amplitude to a sufficient level.

[0006] The present disclosure incorporates by reference the entire contents of the following U.S. patents: U.S. Patent No. 7,702,502, entitled “Apparatus for Signal Decomposition, Analysis and Reconstruction,” filed on February 23, 2006, U.S. Patent No. 7,706,992, entitled “System and Method for Signal Decomposition, Analysis and Reconstruction,” filed on February 23, 2006, U.S. Patent No. 8,086,304, entitled “Physiological Signal Processing to Determine a Cardiac Condition,” filed on November 26, 2008, and U.S. Patent No. 9,530,425, entitled “Method and Apparatus for Signal Decomposition, Analysis, Reconstruction and Tracking,” filed on March 17, 2014, which include disclosures of signal decomposition and analysis techniques that can be applied to respiratory rate processing.

[0007] In general, in one innovative aspect, a system includes: a sensor configured to detect an impedance of a patient's chest; a display device; one or more data processing devices coupled to the sensor and the display device; and a non-transitory computer-readable storage medium encoded with a computer program. The program includes instructions that, when executed by the one or more data processing devices, cause the one or more data processing devices to perform operations. The operations include receiving a signal corresponding to the impedance of the patient's chest; filtering the signal using one or more filters, the one or more filters reducing noise and centering the signal around a zero baseline; adjusting the amplitude of the filtered signal based on a threshold; separating the amplitude-adjusted signal into component signals, wherein each of the component signals represents a frequency-limited frequency band; detecting a fractional phase transition of the component signals in the component signals; selecting a main component signal from the component signals based on the amplitude of the component signal at a time corresponding to the detected fractional phase transition; determining the frequency of the main component signal at a time corresponding to the detected fractional phase transition; and determining the patient's respiration rate based on the determined frequency. Separating the amplitude-adjusted signal into the component signals may include filtering the signal using a bandpass filter bank. Detecting fractional phase transitions of the component signal may include detecting a positive zero crossing, a negative zero crossing, a positive peak, and a negative peak of the component signal. Adjusting the amplitude of the filtered signal based on a threshold may include determining the threshold based on an average amplitude of the filtered signal, and limiting the filtered signal based on the threshold. Selecting a primary component signal may include generating a data object at a detected fractional phase transition of the component signal, the data object containing values ​​representing the time, amplitude, and phase corresponding to the component signal at the detected fractional phase transition; and selecting the component signal with the highest amplitude as the primary component signal based on the amplitude value contained in the data object generated for the component signal corresponding to the time of the detected fractional phase transition. Determining the frequency of the primary component signal may be based on determining the time and phase difference between fractional phase transitions. Determining the patient's breathing rate may include determining the average breathing rate over a period of time.

[0008] In another aspect, an apparatus includes a non-transitory computer-readable storage medium encoded with a computer program. The program includes instructions that, when executed by one or more data processing devices, cause the one or more data processing devices to perform operations. The operations include: receiving a signal corresponding to an impedance of a patient's chest; filtering the signal using one or more filters, the one or more filters reducing noise and centering the signal around a zero baseline; adjusting the amplitude of the filtered signal based on a threshold; separating the amplitude-adjusted signal into component signals, wherein each of the component signals represents a frequency-limited band; detecting fractional phase transitions of the component signals in the component signals; selecting a main component signal from the component signals based on the amplitude of the component signals at a time corresponding to the detected fractional phase transition; determining the frequency of the main component signal at a time corresponding to the detected fractional phase transition; and determining the patient's respiratory rate based on the determined frequency. Separating the amplitude-adjusted signal into component signals may include: filtering the signal using a bandpass filter bank. Detecting the fractional phase transitions of the component signals may include: detecting positive zero crossings, negative zero crossings, positive peaks, and negative peaks of the component signals. Adjusting the amplitude of the filtered signal based on a threshold may include: determining the threshold based on an average amplitude of the filtered signal, and limiting the filtered signal based on the threshold. Selecting a primary component signal may include: generating a data object at a detected fractional phase transition of the component signal, the data object containing values ​​representing the time, amplitude, and phase corresponding to the component signal at the detected fractional phase transition; and selecting the component signal with the highest amplitude as the primary component signal based on the amplitude value contained in the data object generated for the component signal corresponding to the time of the detected fractional phase transition. Determining the frequency of the primary component signal may be based on determining the time and phase difference between fractional phase transitions. Determining the patient's breathing rate may include: determining an average breathing rate over a period of time.

[0009] In another aspect, a method includes receiving a signal corresponding to an impedance of a patient's chest; filtering the signal using one or more filters, the one or more filters reducing noise and centering the signal around a zero baseline; adjusting the amplitude of the filtered signal based on a threshold; separating the amplitude-adjusted signal into component signals, wherein each of the component signals represents a frequency-restricted band; detecting fractional phase transitions of the component signals in the component signals; selecting a primary component signal from the component signals based on the amplitude of the component signals at a time corresponding to the detected fractional phase transition; determining a frequency of the primary component signal at a time corresponding to the detected fractional phase transition; and determining a respiration rate of the patient based on the determined frequency. Separating the amplitude-adjusted signal into component signals may include filtering the signal using a bandpass filter bank. Detecting the fractional phase transitions of the component signals may include detecting positive zero crossings, negative zero crossings, positive peaks, and negative peaks of the component signals. Adjusting the amplitude of the filtered signal based on a threshold may include determining a threshold based on an average amplitude of the filtered signal, and limiting the filtered signal based on the threshold. Selecting the primary component signal may include: generating a data object at a detected fractional phase transition of the component signal, the data object containing values ​​representing the time, amplitude, and phase corresponding to the component signal at the detected fractional phase transition; and selecting the component signal with the highest amplitude as the primary component signal based on the amplitude value contained in the data object generated for the component signal corresponding to the time of the detected fractional phase transition. Determining the frequency of the primary component signal may be based on determining the time and phase difference between the fractional phase transitions. Determining the respiration rate of the patient may include determining an average respiration rate over a period of time.

[0010] Various other features and benefits of such systems and methods will be apparent from the preceding detailed description and claims. BRIEF DESCRIPTION OF THE DRAWINGS

[0011] Figure 1 is a block diagram illustrating an example of a respiratory rate processing system.

[0012] FIG. 2A to FIG. 2C is a graph showing the impulse response of the frequency bands of the analysis wavelet group of the respiratory rate processing system.

[0013] Figure 2D is a graph showing the frequency response of the real part of the analysis wavelet group.

[0014] Figure 3 is a waveform showing a cycle from one component signal decomposed into four quarter phases.

[0015] Figure 4 is a flow chart illustrating an example of operations performed by a respiratory rate processing system.

[0016] Figure 5 An example of a respiration rate processing system including a patient-worn sensor in communication with a display device is shown.

[0017] Figure 6 is a block diagram illustrating an example of a computing system that may be used to implement a respiratory rate processing system.

[0018] Figure 7 is a block diagram illustrating an example of a computing system depicting a host system that may be used to implement a respiratory rate processing system.

[0019] Like reference numbers in the various drawings indicate like elements. DETAILED DESCRIPTION

[0020] Figure 1 is a block diagram illustrating an example of a respiratory rate processing system 100. System 100 includes electronic components, such as hardware and / or software, that facilitate respiratory rate processing. In some implementations, a portion of system 100 is included in a patient monitoring device that can continuously collect physiological vital sign data, such as an electrocardiogram (ECG) waveform, heart rate, respiratory rate, blood oxygen saturation, blood pressure, temperature, and pulse rate. Examples of patient monitoring devices include chest sensors, such as those described in U.S. Publication No. 2015 / 0302539, entitled “Patient Care and Health Information Management Systems and Methods” filed on October 22, 2015, and U.S. Publication No. 2016 / 0228060, entitled “Patient Worn Sensor Assembly” filed on February 9, 2016; and other sensor assemblies connected to peripheral devices, such as the sensor assemblies described in U.S. Publication No. 2018 / 0213859, entitled “Sensor Connection Assembly for Integration with Garment” filed on January 25, 2018, the disclosures of which are incorporated herein by reference in their entirety. In some implementations, a portion of system 100 is included in a monitoring station that uses signals received from one or more peripheral devices, such as a chest sensor, a fingertip probe, an earlobe probe, or a non-invasive blood pressure (NIBP) cuff, to analyze and display physiological vital sign data. Examples of monitoring stations include bedside monitors and central server stations also described in US Publication No. 2015 / 0302539.

[0021] System 100 includes a sensor 102 that injects a high frequency, low amplitude current across the chest cavity, measures the voltage resulting from the current injection, and derives impedance (magnitude of resistance and phase shift) from the injected current and the measured voltage. When the lungs are full, current does not flow easily through the chest, so resistance increases as lung volume increases. Sensor 102 can be affixed to the patient's chest, for example.

[0022] The raw impedance signal from sensor 102 may contain high frequency noise and low frequency artifacts that, while relatively large, reside primarily outside of the expected frequency band associated with respiration rate (e.g., nominally from 5 breaths per minute to 60 breaths per minute (BrPM), or from 0.0833 Hz to 1 Hz, based on physiological limitations and conventions). Therefore, the artifacts may be suppressed by applying a low pass filter 104 and a high pass filter 106 to the impedance signal.

[0023] The raw impedance signal may contain high frequency artifacts, including spurious noise components typically above 1.5 Hz, and 50 Hz and 60 Hz power line EMI products. For an impedance signal sampling rate F of, for example, 33.33 Hz, s_imp , these EMI products appear at 16.66 ... Hz and 6.66 ... Hz due to frequency folding effects associated with aliasing. These artifacts can be mitigated by applying a cascade of low-pass filters 104, where each stage is designed to match the signal passband and specifically suppress the EMI products. The low-pass filter 104 can be implemented using a computationally efficient scaling integration kernel. The scaling integration kernel described in U.S. Patent No. 7,706,992 can effectively implement the low-pass operation by performing a moving average low-pass through the transfer function:

[0024]

[0025] With scale parameter w p and the order parameter N i The filter kernel size and order are set to meet specific design criteria.

[0026] The signal bandwidth can be matched by placing a low pass with a -6dB corner frequency set to the highest frequency of interest, e.g. 1.0 Hz in this case. The scale parameter can be calculated from the formula w = round(F s_imp *k 6dBNul / F 6dB ) to set, where k 6dBNul is determined numerically to establish the ratio of the corner frequency to the first zero frequency of the kernel and depends on the order N of the kernel i For example, it is possible to determine N i= 2 to provide acceptable transition sharpness and out-of-band suppression, especially in combination with an EMI suppression stage, so that k 6dBNul =0.44295.

[0027] EMI artifacts can be mitigated by applying two filter cores in cascade, where each filter core nulls its first response F null Placed at one of the two EMI frequencies, this is done by using the formula w = round (F s_imp / F null ) to set the scaling parameters. The first-order kernel can be used for EMI suppression because the response zeros provide complete attenuation at those zero frequencies.

[0028] Based on the above derivation, and for the impedance signal sampling period T s = 30 ms, the parameters of each low-pass filter stage are as follows:

[0029] class <![CDATA[w p ]]> <![CDATA[N i ]]> Signal 15 2 EMI 1 2 1 EMI 2 5 1

[0030] Table 1

[0031] A high pass filter 106 is applied to the low pass filtered impedance signal. The high pass filter 106 removes the large offset near DC introduced by the original impedance value, centers the signal around a zero baseline, and also suppresses as much as possible low frequency artifacts that are actually below the lower nominal breathing rate band edge (5BrPM or 0.0833 Hz corresponding to a 12 second breathing interval). Given the size of the time scale, a minimum phase response can be selected for the high pass filter 106 to minimize the time lag of the system due to the inherent group delay of the filter. Using an IIR filter topology in recursive form can produce a memory efficient filter architecture. A second order response can be selected because it provides good stability and adequate rejection in the stop band. The high pass filter 106 is specified according to the following analog domain transfer function:

[0032]

[0033] Among them, ω c controls the corner frequency (in radians / second) and Q controls the resonance of the filter at the corner. A value of Q = 0.720 can be chosen to maximize the flatness of the response in the passband and the sharpness of the response at the corner, while significantly minimizing ringing in the impulse response. For this value of Q, the formula ω c =2πk Q f c Set the -6dB point of the filter at frequency f c where factor k Q =1.325.

[0034] The model can be implemented numerically using a second-order difference equation of the following form:

[0035] a 0 Y HP (n) = b 0 x(n)+b 1 x(n-1)+b 2 x(n-2)-a 1 Y HP (n-1)-a 2 Y HP (n-2)

[0036] Among them, x(n) and y HP (n) are the signal input and output of high-pass filter 106 at sampling time n. The filter response is given by coefficient a i and b j To produce a digital filter that approximates the analog model, a bilinear transform can be employed as it provides a very close match of the analog filter response, particularly in its passband and transition band. An example of the resulting digital filter coefficients is as follows:

[0037] <![CDATA[a 0 ]]> 1.0 <![CDATA[a 1 ]]> -1.97108032037207 <![CDATA[a 2 ]]> 0.971507303518756 <![CDATA[b 0 ]]> 0.985646905972705 <![CDATA[b 1 ]]> -1.9712938119454 <![CDATA[b 2 ]]> 0.985646905972706

[0038] Table 2

[0039] The output of the high pass filter 106 is a filtered impedance signal s(n) used to estimate the respiration rate. The filtered impedance signal s(n) itself can be displayed on a display 118 to provide a clean, zero-centered waveform showing the impedance fluctuations (waves) in response to respiration.

[0040] The filtered impedance signal s(n), although improved by the filtering process, may still contain occasional artifacts due to step changes observed in the raw impedance readings. These step changes may be quite small in percentage of the average raw impedance value, but can still be an order of magnitude larger than the relatively small impedance changes due to breathing, and therefore can appear as relatively large spikes on the observed signal after the above filtering. In addition, the amplitude of the impedance changes is occasionally observed to become two or three times (or more) larger than the average over a few respiratory cycles, resulting in amplitude transients. This may occur in association with deep breathing, for example, during pauses in talking, or after brief pauses in breathing. In either case, the large transient condition can, in some cases, over-excite the analysis wavelet group 110 during the respiratory rate estimation process, which can cause the respiratory rate processing to be inaccurate or completely lost for a considerable period of time (e.g., tens of seconds or longer, depending on the size of the transient).

[0041] These types of transients can actually be mitigated by some form of amplitude control. In some implementations, this is accomplished by controlled clipping of the signal by the amplitude processing unit 108, which 1) estimates and tracks the local average signal amplitude to form an adaptive threshold, and then 2) clips the signal by limiting the absolute excursion of the signal in either direction from zero to that threshold. Because the spike transients are relatively short in time, such clipping serves to reduce their energy, thereby mitigating their excitation in the analysis wavelet set 110. Clipping also limits the excessive excitation power of waves with increased amplitude, while still maintaining their periodic information, thereby allowing respiratory rate processing to continue without interruption during these amplitude transients.

[0042] The signal amplitude estimate forms a clipping threshold that generally follows the steady-state amplitude of the signal, but is relatively slow to respond to sudden, large upward changes in amplitude associated with transient conditions. A clipper based on this threshold clips most transients while recovering relatively quickly back to a steady-state level.

[0043] In some implementations, signal amplitude tracking can be achieved using a nonlinear averaging (NLA) process, which is generally robust and operates efficiently in real time. First, a raw amplitude signal is formed by taking the absolute value of the filtered impedance signal s(n) sample by sample, to which a nonlinear recursive filter is applied to form a local time-averaged amplitude estimate. The basic block operation of the NLA process is characterized in an input-output sense by the following nonlinear difference equation:

[0044] y(n)=(1-a n )x(n)+a n y(n-1)

[0045] in,

[0046] x(n) = instantaneous / original signal amplitude estimate for each sample index n;

[0047] y(n) = filtered amplitude estimate of the signal; and

[0048]

[0049] The NLA block updates the output y once for each sample n of the impedance signal. The output value y(n) of the current sample is the output value of the previous sample multiplied by the attenuation factor a n (where 0<a n <1) and the input value x(n) of the current sample multiplied by the complementary coefficient (1-a n ). For a given feedback coefficient a n, the NLA block takes the form of a first-order feedback low-pass filter with unity gain at DC. Thus, the output y(n) can be considered as a local average of the amplitude signal x(n), where the average is taken over the most recent past values. The feedback coefficient a n The value of determines the length of time for which the averaging is performed, and therefore determines the settling time of the output in response to a step function input. n The closer it is to 1, the longer it will take to stabilize.

[0050] The instantaneous suppression action of the NLA block is performed by making a n The value of depends on the relationship between the input and output values. When the input amplitude value x(n) is equal to or greater than the average amplitude y(n), the feedback coefficient a n Take the rise value a r ; When the input amplitude is lower than the average amplitude, the feedback coefficient a n Take the drop value a f By making the rise coefficient closer to 1 than the fall coefficient, the filter responds more slowly to input amplitudes above the local average, thereby reducing tracking sensitivity to large or rapid rises in transient input signal amplitude. Similarly, the fall coefficient is set so that the settling time is short enough to facilitate detection of signal loss. The behavior of signal amplitude tracking under steady-state conditions generally follows the average of the settling times corresponding to the two coefficients.

[0051] Through the impulse invariant mapping from the s-plane poles to the z-plane, according to Digital feedback coefficient a n The analog time constant τ of the first-order analog low-pass filter n (expressed in seconds) related, where T s is the sampling period of the impedance signal (also in seconds). Therefore, the filter output can be expected to be nominally about 5τ after a step change in signal amplitude. n seconds to stabilize. Therefore, the feedback factor a will be increased r and the decreasing feedback coefficient a f Assigned to the corresponding rise time constant τ in physical units of seconds r and the fall time constant τ f .

[0052] Proper design of the rise time constant and the fall time constant optimizes suppression of large upward transients, smooth tracking under steady-state amplitude conditions without over-suppressing small upward amplitude changes, and fast downward tracking in response to signal loss. Good amplitude tracking performance can be obtained using a cascade of three (3) stages of NLA blocks that are chained together so that the output y of the previous block is connected to the input x of the consecutive block. The input x to the first block in the chain is then the raw absolute value of the filtered impedance signal s(n), and the output y of the last block in the chain is then used to form a tracked amplitude estimate according to aEst(n) = kAclp·y(n), where kAclp is a scaling factor that adjusts the optimal clipping level. Examples of optimal values ​​for the design parameters are as follows:

[0053] <![CDATA[τ r ]]> 10 seconds <![CDATA[τ f ]]> 0.3 seconds kAclp 9

[0054] Table 3

[0055] Note that for this example, the time constant τ r and τ f The value of is the same for all three stages of the NLA block. At initialization, the amplitude tracking value is reset to zero and must rise to a steady-state level, which may take an excessive amount of time because the rise time constant τ r To accelerate the initial stabilization to steady state, after the reset, the time constant is kept at τ f Up to an initialization period, such as 10 seconds.

[0056] The filtered impedance signal s(n) is clipped relative to the tracked amplitude estimate according to the following rule (operating for each sample n):

[0057]

[0058] This produces an amplitude normalized impedance change signal s used to estimate the respiration rate c (n).

[0059] The tracked amplitude estimate can be used to detect a loss of signal amplitude consistent with conditions where the respiration rate (and therefore its estimate) is not well defined - for example, lead detachment, breath holding, and apnea. Signal loss can generally be demonstrated as a drop in the measured value aEst(n), where the lowest value is limited by background artifacts due to electrical noise or parasitic impedance fluctuations. The threshold used to determine such a condition can be set midway between the typical levels encountered during quiescent conditions and the typical levels encountered during active breathing conditions to maximize the margin of discrimination between these conditions. Flag f sig Can be used to limit the signal level of respiration rate processing. sigIt is generally effective for amplitude levels above this threshold, but ineffective for amplitude levels below this threshold.

[0060] In the case where the measured value moves slowly through the defined threshold region and / or has superimposed small ripples, it may cause jittering of the defined mark. This can be alleviated by adding hysteresis to the threshold (e.g., making the threshold anti-correlated with the defined state), resulting in two thresholds thr corresponding to the upper threshold and the lower threshold, respectively. U and thr D In addition, due to the time lags present in the various processing stages, a time delay can be added to the validation and invalidation of the flag so that the indication of signal quality is aligned in time with the output of the system. Both of these goals can be met by implementing the signal qualification process using a state machine.

[0061] Examples of state definitions and state transition rules are as follows:

[0062] Initialization: state = NOT_QUAL / / f sig = False

[0063]

[0064]

[0065] Table 4

[0066] Examples of amplitude-limited parameter values ​​are as follows:

[0067] Threshold (Ohm AC) <![CDATA[thr U ]]> 3.08 <![CDATA[thr D ]]> 2.27 Delay (seconds) <![CDATA[T NQ ]]> 2 <![CDATA[T Q ]]> 25

[0068] Table 5

[0069] The threshold and delay times can be determined empirically based on experimental data. The delay for each flag state can be selected to invalidate the flag before signal conditions deteriorate and re-validate the flag after signal conditions recover, seeking to ensure favorable signal conditions only when the associated flag is valid.

[0070] Signal limit sign f sig may be presented to the output interface 116 as a means of suppressing the display of false or erroneous respiration rate outputs during adverse signal conditions. sig Can be used to detect apnea. By detecting loss of contact (e.g., by showing raw impedance out of range condition), signal loss due to lead detachment can be distinguished from signal loss due to apnea. sig When false and the lead contacts are intact, apnea can be detected.

[0071] The respiratory rate estimation process can be based on a method for tracking and estimating rate that is based on embedded periodicity, as described in U.S. Pat. No. 9,530,425. The method provides accurate, robust rate estimates at a low computational cost. The method utilizes the techniques described in U.S. Pat. No. 7,706,992 and extends these techniques to be able to track the fundamental rate of the fundamental component in the input signal over a wide rate range (e.g., greater than one octave). The respiratory rate estimation process in the current example covers a nominal physiological rate range of 5BrPM to 60BrPM, over a range of more than 3.5 octaves. For respiratory rate estimation, processing and tracking of components can be performed in quarter phase (QP) space. In this space, updates are performed when the input signal changes at important information points (also called "important points") without losing measurement or timing accuracy. This "smart sampling" can translate into significant computational savings because the frequency of QP-based information updates typically occurs much less frequently than signal updates that occur once per sampling period.

[0072] As described in U.S. Pat. No. 9,530,425, the amplitude normalized impedance change signal s c (n) may first be processed with a multi-band analysis wavelet set to analyze the signal over a set of rate bands to cover the range of expected breathing rates. The band outputs are analysis component signals to which a QP transform (as described in U.S. Pat. No. 7,706,992) is applied by a QP object transform and processing unit 112 to detect extrema and zero crossings of the component signals. An updated QP object containing key time-frequency information associated with periodicity in the signal is formed at these points. This process produces an array of streams of QP objects, one for each component signal (or band). The QP objects may then be processed in QP space by a QP periodicity tracking unit 114 to identify and track the periodicity of the amplitude-normalized impedance change signal s c The main periodic component identified in (n) (as described in US Pat. No. 9,530,425) can be extracted from this component with high temporal accuracy to determine the local average respiration rate.

[0073] The analysis wavelet group 110 is usually designed as a bandpass filter group with adjacent spaced frequency band centers. Each bandpass filter in the group responds to the periodicity present in the group input signal by changing the amount according to its frequency response, and has the maximum output amplitude for the frequency closest to the center frequency of the bandpass. For the signal applied to the filter group input containing (quasi) periodic components, the filters with the maximum output amplitude relative to their neighbors at a certain point in time will be those frequency bands whose centers are closest to the frequencies of those signal components. Tracking of the frequency band amplitude as a function of the frequency band frequency at that point in time will reveal the "peaks" at those associated frequencies. Finding and tracking these amplitude peaks as a function of time, as well as the basic information in the frequency bands indexed by these peaks, forms the basic principle of the bandpass filter group. Another advantage of using bandpass filters is that each filter in the group has a relatively narrow bandwidth, and therefore generally has a higher SNR relative to the associated frequency components.

[0074] A design consideration for a filter bank is the so-called "time-frequency frame" of the system, which is controlled by the "Q" (relative bandwidth) of the bandpass filters. Closely related to the time-frequency frame is the frequency spacing of the bands. Excessive band overlap represents a finer sampling of the frequency domain, but increases computational overhead that may constitute redundant spectral information. However, insufficient band overlap may result in excessive amplitude loss in the crossover region between the band centers, leaving "holes" in the spectral representation of the signal, and in extreme cases, leading to blurring of the spectral peak pattern. A higher Q will produce a narrower filter bandwidth, allowing the filters to be more closely spaced in frequency for a given amount of band overlap, thereby providing increased frequency resolution. However, achieving a narrow filter bandwidth necessarily requires having a wider impulse response in time, which in turn translates into lower time resolution, as well as increased processing delays, which may result in excessive response lag in the system. Therefore, the choice of Q represents a trade-off between frequency resolution and time resolution and / or response time.

[0075] As described in U.S. Pat. No. 9,530,425, processing in QP space allows time-frequency information to be captured with much higher time and frequency accuracy than would be conveyed by simply using the center frequency of a given band as a rate estimate. This property can relax the frequency resolution requirements between band centers that are stringent for resolving frequencies. In theory, a minimum of two bands per octave are required to isolate the fundamental harmonic from the second harmonic of the input periodic component, although in practice more bands are required to achieve low crossover losses between band responses.

[0076] The band centers can be distributed so that their spacing is not too sparse over the frequency range where respiratory rate processing should exhibit maximum accuracy, such as for a normal range where long-term trends can be monitored and / or smoothness of system response is desired. For monitoring respiratory rate, a band spacing of approximately 1 BrPM to 2 BrPM can produce the desired response smoothness over the physiological rate range of up to 30 BrPM. Rates above this value are associated with hyperventilation, requiring faster response times to a greater degree than response smoothness, so Q can be set lower in these ranges, allowing resolution requirements to be relaxed over the higher rate ranges. Using purely logarithmic band spacing over the entire rate range can result in excessive compression of the band spacing at the low end of the frequency range and / or excessive sparseness of the band spacing at the high end of the frequency range, while using purely linear band spacing can result in the opposite disadvantage. The curve of the target band spacing can follow a power law according to the following formula:

[0077] f n =(g 1 +n(g 2 -g 1 ) / (N b -1)) p

[0078] in

[0079] And, n∈{0, 1, ..., N b -1}

[0080] Target rate f 0 =5BrPM, Target frequency band number N b = 24, power law p = 1.5. This distribution results in an average of about 7.2 bands per octave and a minimum band spacing of 3.4 bands per octave.

[0081] As described in U.S. Pat. No. 7,706,992, the analysis wavelet set 106 may be implemented using a parallel version of the Kovtun-Ricci wavelet transform with analysis component bands. In addition to the scaled integral kernel and the scaled difference kernel, the analysis wavelet set 110 may include an additional kernel type in the kernel cascade, referred to herein as a scaled comb kernel, having a z-domain transfer function:

[0082]

[0083] For a frequency band p with a scale parameter k p and c p And the order parameter N cp The intrinsic latency of this core is D cp =N cp ·k p ·(cp –1). For N cp = 1, the frequency response of the scaled comb kernel is a Dirichlet function that closely approximates a sine function at its first main lobe around DC, and then repeats these main lobes in frequency at intervals controlled by the scale parameter Kp (specifically, the frequency interval is F s / (2k p ), where F s is the sampling rate). p The value of can be set to the same scaling value k that appears in the expression for the scaled difference kernel. p This centers the first repeating main lobe of the scaled comb kernel on the first main lobe of the scaled difference kernel, defining a new, more selective bandpass function centered on the peak of the main lobe. The scaled comb feature responds to zeros on either side of each repeating main lobe, according to the scale parameter c p is spaced from each center so that for c p = 2, the zero is in the middle of the main lobe center, and for larger c p , the zeros are proportionally moved closer to the center of the main lobe. These zeros have the effect of narrowing the bandwidth of the main lobe, effectively increasing the Q of the filter. Q also increases with the order N cp The core is therefore characterized by two parameters by which Q can be additionally controlled. The performance can be further enhanced because even for c p = 2, the increased Q obtained due to the increased side zero values ​​results in less impulse response broadening than when similar Q is obtained by simply increasing the filter order. On the other hand, increasing the order does reduce the height of the side lobes between the main lobe images, so the order N cp becomes a means of controlling stopband suppression in the region just beyond the first passband zero.

[0084] In a broadband sense, the scaling comb kernel acts as a comb filter with lobes around DC and scaled according to k. p Repeating lobes. To obtain a pure bandpass response, it may be desirable to isolate the first repetition of the main lobe response above DC, which is at frequency F s / (2k p ) is centered. To suppress the lobes at DC, a scaled difference kernel can be used since it has a main response zero at DC. A scaled integration kernel can then be used to suppress the mainlobe image above the first, as well as provide an overall smoothing of the impulse response. In practice, using these two kernels in the role of stopband suppression often does not require a high order (e.g., an order of 4 may provide adequate out-of-band suppression).

[0085] Therefore, the parameter used to program the bandpass wavelet response is k p 、c p 、Ncp 、N dp 、w p and N ip For the pth passband, k p controls the center frequency, and Q is determined (primarily) by c p and (minor) N cp and the scaling derivative order N dp Parameter N cp and N dp Controls the stopband rejection near DC and scales the integration scale w p and order N ip Controls stopband rejection above the passband.

[0086] This series of bandpass wavelets can be computationally efficient and characterized by a linear phase response for each band. As described in U.S. Patent No. 7,706,992, this group of wavelets performs a time-frequency transform. The transform can be achieved with an additional time delay per band to maintain time alignment across the bands. Although retaining the timing information between the bands in the transform is useful for extracting characteristics within multiple components in the basic signal, in order to achieve this, it is necessary to adapt the longest delay in the group, such as the delay of the lowest rate band, while increasing the delay of the higher rate band. The lowest supported respiratory rate may be 5BrPM, with a basic period of 12 seconds per breath. Since the band wavelets typically support the order of several cycles in their impulse response, this causes a large lag for the higher frequency bands. The purpose is to estimate only the frequency of the (single) basic component of the respiratory rate, and to provide a rapid response to a rapidly increased respiratory rate with as little delay as possible, such as during a sudden attack of hyperventilation. Therefore, the analysis wavelet group 110 can be implemented without time alignment between frequency bands, and therefore without additional delay elements, so that each frequency band only exhibits the intrinsic delay of its own wavelet structure. The resulting characteristic of the group 110 is that the intrinsic delay will generally decrease with increasing frequency band center frequency and / or rate.

[0087] The Q of the filter bank can then be set to the point where the amplitude response of a given band crosses the amplitude response of the next adjacent band, based on the desired inter-band overlap point, i.e., the amplitude drop relative to the amplitude response peak. To resolve a single dominant frequency component, the overlap at the -6dB point is theoretically sufficient to ensure amplitude coverage across the entire group to prevent amplitude ambiguity due to "holes" in the coverage between bands. In some embodiments, additional margin for overlap is implemented for the analysis group to provide a certain amount of smoothing in the transition between bands. The overlap point can be placed at a nominal -1dB (minimum -1.3dB) to ensure smooth coverage while still maintaining high selectivity.

[0088] Based on the above description and derivation, and for the sampling period as previously specified for the impedance signal, the bandpass wavelets of the analysis wavelet group 110 may be implemented using the following design parameters:

[0089]

[0090]

[0091] Table 6

[0092] Value F Cp is the center frequency of the band, defined as the frequency at which the pth band exhibits its peak amplitude response. Complete coverage of the desired rate range can be achieved with 24 bands. For purely logarithmic-spaced bands, if the Q parameter of the band is kept constant (producing a so-called "constant Q" group), the band crossing points are regularized. Since the band spacing in the group is not logarithmic but superlinear, maintaining regularity in the band crossing points may require increasing certain Q parameters at well-placed points along the frequency axis (determined empirically for optimal band crossing characteristics). Therefore, there are several representative responses for various scales, one response for each specific Q in the group.

[0093] FIG. 2A to FIG. 2C are example graphs 200, 202, 204 showing impulse responses for bands 1, 9, and 17, respectively (specifically, each lowest frequency analysis wavelet band for each band range in the group with a particular Q). The time domain impulse responses generated for these wavelets are characterized by effective time lengths between 3 and 5 cycles, thus representing very good time resolution of local periodic characteristics. For each of these graphs, the maximum time amplitude point is normalized to 1, and the x-axis is time in samples. Since the wavelets are analytic, the impulse responses are complex values. The "real" and "imag" traces are the real and imaginary parts of the impulse response, respectively. The "mag" trace is the instantaneous complex amplitude at each time point. Figure 2D is an example graph 250 showing the frequency response of the real part of the analytical wavelet group after compensation as described below.

[0094] To ensure calibration of the relative responses across the filter bank, the band output signals are normalized to unity gain at the center of the band. Cp The real and imaginary magnitude responses are evaluated (separately) and the multiplicative inverse is taken to determine the calibration factor A. NR and A NI In some implementations, the real and imaginary filter outputs are multiplied by their respective calibration factors to obtain a calibrated filter bank. Thus, the calibration factor for each respective frequency band is as follows:

[0095]

[0096]

[0097] Table 7

[0098] Reference Figure 1 , the QP object transformation and processing unit 112 detects QP transitions on each of the component signals received from the analysis wavelet group 110 . Figure 3 3 is a waveform 300 showing an example of a cycle from one component signal being decomposed into four quarter phases labeled A, B, C, and D. The component signals are decomposed and, in some implementations, defined according to certain unique features, such as positive and negative peaks and positive and negative zero crossings of the component signal. These unique features are shown below in a tabular format:

[0099] QP Mark Started in Ends at A Positive zero crossing Positive peak B Positive peak Negative zero crossing C Negative zero crossing Negative peak D Negative peak Positive zero crossing

[0100] Table 8

[0101] Refer again Figure 1 , the QP object transformation and processing unit 112 forms a QP object (e.g., a data structure) that collects QP attributes (e.g., in data fields) at each QP transition, as described in U.S. Pat. No. 7,706,992. For the purpose of respiration rate estimation, the QP object contains the phase value (denoted by the phase marker L) at the time point of the QP detection. QP Indicates), time value t QP and component amplitude a QP , and the index of the frequency band for the basic component. The current QP at the current point in time is collected into an array across frequency bands to form a QP state vector that is indexed by frequency band and updated when any QP object on any frequency band is updated, thus containing the QP state update (per U.S. Patent No. 7,706,992). The concentrated time-frequency information contained in this state vector forms the basis for being able to identify and track dominant frequency signal components and to make high-precision estimates of their instantaneous frequency and amplitude.

[0102] The breathing rate estimate is based on calculating the atomic period of the rate component at each QP state update, specifically across all QP objects (frequency bands) updated at the time of the QP state update. As described in U.S. Patent No. 7,706,992, this estimate is based on the calculation time and phase difference from one QP object update to the next QP object update (i.e., QP target transitions), which in turn yields information about the period of the fundamental frequency component with high accuracy. The atomic period estimate for a given QP object update can be expressed as follows:

[0103] TR QP =dT QP / (60·dP QP )

[0104] Value TR QP Corresponds to the atomic breathing period in minutes / cycle (minutes / breath). Value dT QP Corresponds to the time interval from the previous QP object update to the current QP object update, in seconds. QP corresponds to the phase advance value associated with the number of quarter-phase markers traversed from the previous QP state update to the current QP state update, and is evaluated according to the following table:

[0105]

[0106] Table 9

[0107] According to U.S. Patent No. 7,706,992, the number of QP phase marks traversed corresponds to the increments encountered along the cyclic sequence A→B→C→D→A in the current QP object transition. A fundamental phase advance (e.g., a mark increment of 1) is by far the most common phase change. A zero phase change is not possible, but is possible during noise or transients, especially at startup. Phase mark changes of -1 or 2 are also possible in the case of sudden phase reversals, typically associated with transients in the signal, or more commonly, with updates of the QP object occurring in a frequency band far from any significant periodic components in the signal, and therefore exhibiting low relative frequency band amplitude and / or SNR. The QP phase mark can then be updated after the QP phase mark has changed and the dP QP Update the atomic breathing period estimate TR at each QP object update for the frequency band greater than zero QP .

[0108] To form stable QP statistics, each QP amplitude and atomic period is smoothed by a scaled integration kernel operating in QP space. The integration scale can be set to 4QP to naturally match one period of the basic wavelet scale. A first-order scaled integration kernel can provide sufficient smoothing at this stage. The smoothing operation results in an instantaneous component amplitude a evaluated on each band when the associated QP object is updated. inst and the instantaneous respiratory cycle TR inst These estimates are then appended to the QP state vector for each band, forming an instantaneous amplitude vector and an instantaneous breathing period vector that are also updated at each QP state update.

[0109] At each QP state update, the current state of the associated smoothed instantaneous amplitudes acquired across the frequency bands forms the QP amplitude spectrum. The local periodic components generated by fluctuations in the respiratory rate input signal will tend to excite those frequency bands having a center frequency (center period) about the period of the fundamental component. Therefore, the QP amplitude spectrum at a given QP time point will generally exhibit peaks at the corresponding frequency band indices, where the principal component exhibits the main (or highest amplitude) peak. Observation of the Fourier series characteristics of the respiratory signal waveform (particularly the amplitudes of the harmonics in the series) shows that, in practice, the first harmonic in the series tends to always have the highest amplitude and will therefore appear as the main periodic component. This characteristic tends to remain true even when the respiratory signal waveform is undergoing the above-mentioned limiting process. Since the period of the first harmonic also corresponds to the cyclic period of the signal waveform, the frequency band with the highest amplitude in the QP amplitude spectrum is associated with the fundamental rate component of the basic cyclic signal waveform.

[0110] Therefore, the QP periodicity tracking unit 114 performs the following steps each time the QP state is updated:

[0111] • The dominant frequency band is selected by finding the maximum magnitude on the QP magnitude spectrum.

[0112] If the maximum amplitude band also undergoes a QP object update, then:

[0113] ο Tag tracking updates,

[0114] o update the index of the tracked frequency band with the index of the main frequency band,

[0115] οInstantaneous amplitude a of the main frequency band inst To update the instantaneously tracked a trInst Amplitude, and

[0116] οUse instantaneous respiratory cycle TR inst Update the instantaneously tracked respiratory cycle TR trInst .

[0117] Therefore, the instantaneously tracked respiratory cycle estimate TR trInst Represents the instantaneous period of respiratory signal fluctuations in minutes / cycle, updated four times per local cycle according to quarter-phase processing.

[0118] A statistics-based smoothing filter 115 may be applied to the instantaneously tracked respiratory cycle estimate TR trInst , to produce a smooth moving estimate TR of the respiratory interval est The smoothing filter 115 is designed to remove any remaining large, peripheral transient excursions from the final output and generally stabilize the output reading. The smoothing time T sm= 10 seconds may produce an acceptable level of smoothing while keeping the time lag as low as possible. In some implementations, the processing delay until the smoothing filter 115 totals about 4 cycles (~3 cycles in the analysis wavelet group 110 plus 1 cycle in the QP MA process), or about 20 seconds for a nominal respiratory rate of 12 BrPM (5 second respiratory cycle). This results in a nominal system delay of 30 seconds that is consistent with clinical practice.

[0119] The average within the smoothing filter 115 may use at each tracking update the interval T prior to (and including) the current tracking update. sm TR appears within seconds trIns The value of is updated to obtain the final respiratory cycle estimate TR expressed in minutes / cycle est Finally, the respiration rate estimate can be evaluated as BR est =1 / TR est And expressed in BrPM.

[0120] Rather than performing a simple moving average, the smoothing filter 115 may compute a smoothed median by taking the average of the central values ​​in a percentile sense within a range around the pure median. The central range average may be set to include those values ​​within the interquartile range above and below the median, which are selected to maintain robustness to outliers while also producing a well-smoothed output over time.

[0121] The respiratory rate processing system 100 can be useful for any environment where a patient's respiratory rate is to be monitored or statistically analyzed, including critical care, surgery, rehabilitation, and emergency and hospital ward environments. The system 100 can provide continuous and instantaneous respiratory rate values, which is critical in emergency medicine and is also very useful for patients with respiratory or cardiac problems.

[0122] Figure 4 4 is a flow chart illustrating an example of operations 400 performed by a respiratory rate processing system. Operations 400 may be performed by a system of one or more computers, such as Figure 1 Operation 400 may include the details discussed above.

[0123] At 402, a signal corresponding to an impedance of a patient's chest is received. At 404, the received signal is filtered using one or more filters that reduce noise and center the signal around a zero baseline.

[0124] At 406, the amplitude of the filtered signal is adjusted based on the threshold. Adjusting the amplitude of the filtered signal based on the threshold may include determining the threshold based on an average amplitude of the filtered signal, and clipping the filtered signal based on the threshold.

[0125] At 408, the amplitude adjusted signal is separated into component signals, wherein each of the component signals represents a frequency limited band. Separating the amplitude adjusted signal into the component signals may include filtering the signal using a bandpass filter bank.

[0126] At 410, fractional phase transitions of the component signals in the component signals are detected. Detecting the fractional phase transitions of the component signals may include detecting positive going zero crossings, negative going zero crossings, positive peaks, and negative peaks of the component signals.

[0127] At 412, a primary component signal is selected from the component signals based on the amplitude of the component signal at the time corresponding to the detected fractional phase transition. Selecting the primary component signal may include: (i) generating a data object at the detected fractional phase transition of the component signal, the data object containing values ​​representing the time, amplitude, and phase corresponding to the component signal at the detected fractional phase transition; and (ii) and selecting the component signal with the highest amplitude as the primary component signal based on the amplitude value contained in the data object generated for the component signal corresponding to the time of the detected fractional phase transition.

[0128] At 414, a frequency of the primary component signal at a time corresponding to the detected fractional phase transition is determined.Determining the frequency of the primary component signal may be based on determining a time and phase difference between fractional phase transitions.

[0129] At 416, a breathing rate of the patient is determined based on the determined frequency.Determining the breathing rate of the patient may include determining an average breathing rate over a period of time.

[0130] Figure 5 A system 500 is shown including a patient-worn sensor 502 in communication with a display device 504, such as a bedside monitor, a central server station, or a display of a mobile device as described in U.S. Publication No. 2015 / 0302539. The display device 504 displays information for the patient. The display device 504 can display a user interface 506, which includes information received from the patient-worn sensor 502 and / or information associated with the patient associated with the patient-worn sensor 502. The patient-worn sensor 502 can be, for example, a chest sensor as described in U.S. Publication No. 2015 / 0302539. The patient-worn sensor 502 can include contacts for attaching to the patient's skin and recording various patient vital signs such as blood pressure, body temperature, respiratory rate, body impedance, blood oxygenation, heart rhythm (via ECG), and heart rate.

[0131] The patient-worn sensor 502 may wirelessly communicate with the display device 504 via a wireless connection 508 using a wireless communication protocol such as Bluetooth, WiFi, or a cellular protocol. The patient-worn sensor 502 may send vital sign information about the patient to the display device 504 via the wireless connection 508. In some implementations, the patient-worn sensor 502 may process the collected vital sign information before sending the information to the display device 504, such as the respiratory rate processing described above, and in some implementations, the patient-worn sensor 502 may send raw vital sign information to the display device 504 instead of or in addition to the processed information.

[0132] The display device 504 may be a touch screen device capable of receiving touch screen input, such as a tablet computer. In some implementations, the display device 504 may receive input from a keyboard, a mouse, an input button, or one or more devices capable of recognizing voice commands. In some implementations, the display device 504 may be controlled using a device that wirelessly communicates with the display device 504, such as a mobile phone. In some implementations, the display device 504 may be a "headless" device that does not include direct user input and / or output functions, but is used only as a processing device for processing raw vital sign information received from the patient-worn sensor 502, detecting alarm conditions, sending warnings to other devices that communicate with the display device 504, and sending patient information to one or more central servers. In such a case, the display device 504 will not include a display.

[0133] The user interface 506 displays information about the patient associated with the patient-worn sensors 502. For example, electrodes of the patient-worn sensors 502 may be in contact with the patient's skin and collect vital sign information, which is transmitted to the display device 504 via the wireless connection 508, processed by the display device 504, and displayed as part of the user interface 506. The user interface 506 shows various vital sign waves and numerical levels.

[0134] For example, user interface 506 shows a patient's breathing waveform 510 and a patient's numerical breathing rate (RR) value 512. Respiration waveform 510 and breathing rate value may be, for example, based on the above description of Figures 1 to 4The process described is determined. In the example shown, the patient's respiratory rate value 512 is 15 breaths per minute. The user interface 506 indicates that the patient's acceptable respiratory rate range 514 falls between 8 and 30 breaths per minute. Since the patient's current respiratory rate of 15 breaths per minute falls within the indicated acceptable range 514, there is currently no alarm state regarding the patient's respiratory rate. This is indicated by a bell-shaped icon 516 superimposed with an "X" symbol. Icon 516 indicates that the patient's current respiratory rate is within an acceptable range. In the case where the patient's respiratory rate is not within an acceptable level, icon 516 can change to indicate an alarm state. For example, "X" can disappear from icon 516 and icon 516 can light up or flash to indicate an alarm state. In addition, the display device 504 can emit an audible alarm to warn nearby caregivers about the patient's alarm state. In some implementations, other parts of the user interface 506 can flash or otherwise indicate an alarm state. For example, when the patient's respiratory rate is outside the acceptable range 514, the displayed respiratory rate value 512 can flash. In some implementations, icon 516 (or other portions of user interface 506) may flash at varying rates to indicate the severity of a particular alarm state. For example, the further the patient's breathing rate is from an acceptable range, the faster icon 516 flashes.

[0135] Figure 6 6 is a block diagram illustrating an example of a computing system 600 executing a computer program that implements the above-described operations, data structures, and / or programs contained on a computer-readable medium 602 and executed by a central processing unit (CPU) 604. The CPU 604 may be an x86 series CPU or other CPU known in the art. In some implementations, input for executing the program may be implemented through a keyboard 606. In some implementations, input for executing the program may be implemented through a peripheral device such as a mouse, a light pen, a touch screen, a touch pad, or any other input device known in the art. In some implementations, the described operations may be called by another program to perform and process oximetry data. Once processed, the data processed using the above-described operations, data structures, and / or programs may be output to a display 608, or sent to another user terminal 612 via the Internet 610 or other wireless communication channels. In some implementations, the output may be placed in a database 614 or another database located off-site.

[0136] Figure 7is a block diagram showing an example of a computing system depicting a host computer 700, which in some implementations may include a set of units or modules, such as a communication module (COMM) 702 and associated communication links 704 for communicating with sensors worn by a patient, a microprocessor unit (MPU) 706 including a computer readable medium (CRM) 708 containing a set of application instructions, an associated storage module 710 for storing patient data, a human-computer interface (HCI) 712 through which an operator (e.g., a physician, technician, or patient) can interact with the MPU 706 and may include a visual display, an auditory display, a keyboard, a mouse, a touch screen, a tactile device, or other means of interaction. The host computer 700 also includes an input / output module (I / O) 714 for typically connecting to and communicating with computer peripherals (e.g., printers, scanners, and / or data acquisition or imaging devices); and a network communication module (NET) 716 for communicating with other hosts on a computer network such as Ethernet and / or a wireless network or WiFi.

[0137] The features described in the present disclosure may be implemented in digital electronic circuit systems, or in computer hardware, firmware, software, or a combination thereof. The apparatus may be implemented in a computer program product tangibly implemented in an information carrier such as a machine-readable storage device for execution by a programmable processor, and the method steps may be performed by a programmable processor executing an instruction program to perform the functions of the described implementation by operating on input data and generating output. The described features may advantageously be implemented in one or more computer programs, one or more computer programs being executable on a programmable system including at least one programmable processor, at least one programmable processor being coupled to receive data and instructions from a data storage system, at least one input device, and at least one output device, and to send data and instructions to the data storage system, at least one input device, and at least one output device. A computer program is a set of instructions that can be used directly or indirectly in a computer to perform a specific activity or produce a specific result. A computer program may be written in any form of programming language including a compiled language or a parsed language, and it may be deployed in any form, including as a standalone program or as a module, component, subroutine, or other unit suitable for use in a computing environment.

[0138] Suitable processors for executing instruction programs include, for example, both general-purpose microprocessors and special-purpose microprocessors, and the sole processor or one of multiple processors of any kind of computer. Typically, the processor will receive instructions and data from a read-only memory or a random access memory or both. The basic elements of a computer are a processor for executing instructions and one or more memories for storing instructions and data. Typically, a computer will also include one or more mass storage devices for storing data files, or be operably coupled to communicate with one or more mass storage devices for storing data files; such devices include disks, such as internal hard disks and removable disks; magneto-optical disks; and optical disks. Storage devices suitable for tangibly implementing computer program instructions and data include all forms of non-volatile memory, including, for example, semiconductor memory devices, such as EPROM, EEPROM, and flash memory devices; disks, such as internal hard disks and removable disks; magneto-optical disks; and CD-ROM and DVD-ROM disks. The processor and memory can be supplemented by or incorporated into an ASIC (Application Specific Integrated Circuit). To provide interaction with a user, the features may be implemented on a computer having a display device, such as a CRT (cathode ray tube) or LCD (liquid crystal display) monitor, for displaying information to the user and a keyboard and pointing device, such as a mouse or trackball, through which the user can provide input to the computer.

[0139] Features can be implemented in a computer system that includes a back-end component such as a data server, or includes a middleware component such as an application server or an Internet server, or includes a front-end component such as a client computer with a graphical user interface or an Internet browser, or any combination thereof. The components of the system can be connected by digital data communications of any form or medium, such as a communication network. Examples of communication networks include LANs, WANs, and computers and networks that form the Internet, for example. The computer system may include a client and a server. The client and server are usually remote from each other and usually interact through a network such as the one described. The relationship between the client and the server is generated by means of computer programs that run on respective computers and have a client-server relationship with each other.

[0140] Although this specification contains many specific implementation details, these implementation details should not be interpreted as limitations on any invention or the scope that can be claimed, but as descriptions of features specific to a particular embodiment of a particular invention. Certain features described in this specification in the context of different embodiments may also be implemented in a combined manner in a single embodiment. Conversely, the individual features described in the context of a single embodiment may also be implemented individually or in any suitable sub-combination in multiple embodiments. In addition, although the features may be described as acting in a certain combination as described above and even initially claimed as such, in some cases, one or more features from the claimed combination may be deleted from the combination, and the claimed combination may involve a sub-combination or a variation of the sub-combination.

[0141] Similarly, although operations are depicted in a particular order in the accompanying drawings, this should not be understood as requiring that such operations be performed in the particular order shown or in a continuous order, or that all of the operations shown be performed to achieve the desired results. In some cases, multitasking and parallel processing may be advantageous. In addition, the separation of various system components in the above-described embodiments should not be understood as requiring such separation in all embodiments, and it should be understood that the described program components and systems can generally be integrated together in a single software or hardware product or packaged into multiple software or hardware products.

[0142] Thus, specific embodiments of the subject matter have been described. Other embodiments are within the scope of the appended claims. In some cases, the actions recited in the claims can be performed in a different order and still achieve the desired results. In addition, the processes depicted in the drawings do not necessarily require the particular order or sequential order shown to achieve the desired results. In certain implementations, multitasking and parallel processing may be advantageous.

Claims

1. A non-transitory computer-readable storage medium encoded with a computer program, the computer program comprising instructions that, when executed by one or more data processing devices, cause the one or more data processing devices to perform operations comprising: receiving a signal corresponding to the impedance of the patient's chest cavity; filtering the signal using one or more filters that reduce noise and center the signal around a zero baseline; adjusting the amplitude of the filtered signal based on a threshold value; separating the amplitude adjusted signal into component signals, wherein each of the component signals represents a frequency limited band, detecting a fractional phase shift of a component signal in the component signals; selecting a primary component signal from the component signals based on amplitudes of the component signals at times corresponding to detected fractional phase transitions; determining a frequency of the principal component signal at a time corresponding to the detected fractional phase transition; and A breathing rate of the patient is determined based on the determined frequency.

2. The non-transitory computer-readable storage medium of claim 1, wherein: Separating the amplitude adjusted signal into the component signals includes filtering the signals using a bandpass filter bank.

3. The non-transitory computer-readable storage medium of claim 1, wherein: Detecting the fractional phase transition of the component signal includes detecting a positive zero crossing, a negative zero crossing, a positive peak, and a negative peak of the component signal.

4. The non-transitory computer-readable storage medium of claim 1, wherein: Adjusting the amplitude of the filtered signal based on the threshold comprises: determining the threshold based on an average amplitude of the filtered signal; and The filtered signal is clipped based on the threshold.

5. The non-transitory computer-readable storage medium of claim 1, wherein: Selecting the main component signal comprises: generating a data object at the detected fractional phase transition of the component signal, the data object comprising values ​​representing time, amplitude, and phase corresponding to the component signal at the detected fractional phase transition; and Based on the amplitude values ​​contained in the data objects generated for the component signals corresponding to the times of the detected fractional phase transitions, the component signal having the highest amplitude is selected as the main component signal.

6. The non-transitory computer-readable storage medium of claim 1, wherein: Determining the frequency of the principal component signal is based on determining the time and phase differences between fractional phase transitions.

7. The non-transitory computer-readable storage medium of claim 1, wherein: Determining a breathing rate of the patient includes determining an average breathing rate over a period of time.

8. A computer-implemented method comprising: receiving a signal corresponding to the impedance of the patient's chest cavity; filtering the signal using one or more filters that reduce noise and center the signal around a zero baseline; adjusting the amplitude of the filtered signal based on a threshold value; separating the amplitude adjusted signal into component signals, wherein each of the component signals represents a frequency limited band; detecting a fractional phase shift of a component signal in the component signals; selecting a primary component signal from the component signals based on amplitudes of the component signals at times corresponding to detected fractional phase transitions; determining a frequency of the principal component signal at a time corresponding to the detected fractional phase transition; and A breathing rate of the patient is determined based on the determined frequency.

9. The computer-implemented method of claim 8, wherein: Separating the amplitude adjusted signal into the component signals includes filtering the signals using a bandpass filter bank.

10. The computer-implemented method of claim 8, wherein: Detecting fractional phase transitions in the component signals includes detecting positive going zero crossings, negative going zero crossings, positive peaks, and negative peaks of the component signals.

11. The computer-implemented method of claim 8, wherein: Adjusting the amplitude of the filtered signal based on the threshold comprises: determining the threshold based on an average amplitude of the filtered signal; and The filtered signal is clipped based on the threshold.

12. The computer-implemented method of claim 8, wherein: Selecting the main component signal comprises: generating a data object at the detected fractional phase transition of the component signal, the data object comprising values ​​representing time, amplitude, and phase corresponding to the component signal at the detected fractional phase transition; and Based on the amplitude values ​​contained in the data objects generated for the component signals corresponding to the times of the detected fractional phase transitions, the component signal having the highest amplitude is selected as the main component signal.

13. The computer-implemented method of claim 8, wherein: Determining the frequency of the principal component signal is based on determining the time and phase differences between fractional phase transitions.

14. The computer-implemented method of claim 8, wherein: Determining a breathing rate of the patient includes determining an average breathing rate over a period of time.

15. A system for processing a respiratory rate, comprising: a sensor configured to detect impedance of a thorax of a patient; Display device; one or more data processing devices coupled to the sensor and the display device; as well as A non-transitory computer-readable storage medium encoded with a computer program, the computer program comprising instructions that, when executed by the one or more data processing devices, cause the one or more data processing devices to perform operations comprising: receiving a signal corresponding to the impedance of the patient's chest cavity; filtering the signal using one or more filters that reduce noise and center the signal around a zero baseline; adjusting the amplitude of the filtered signal based on a threshold value; separating the amplitude adjusted signal into component signals, wherein each of the component signals represents a frequency limited band; detecting a fractional phase shift of a component signal in the component signals; selecting a primary component signal from the component signals based on amplitudes of the component signals at times corresponding to detected fractional phase transitions; determining a frequency of the principal component signal at a time corresponding to the detected fractional phase transition; and A breathing rate of the patient is determined based on the determined frequency.

16. The system of claim 15, wherein: Separating the amplitude adjusted signal into the component signals includes filtering the signals using a bandpass filter bank.

17. The system of claim 15, wherein: Detecting the fractional phase transition of the component signal includes detecting a positive zero crossing, a negative zero crossing, a positive peak, and a negative peak of the component signal.

18. The system of claim 15, wherein: Adjusting the amplitude of the filtered signal based on the threshold comprises: determining the threshold based on an average amplitude of the filtered signal; and The filtered signal is clipped based on the threshold.

19. The system of claim 15, wherein: Selecting the main component signal comprises: generating a data object at the detected fractional phase transition of the component signal, the data object comprising values ​​representing time, amplitude, and phase corresponding to the component signal at the detected fractional phase transition; and Based on the amplitude values ​​contained in the data objects generated for the component signals corresponding to the times of the detected fractional phase transitions, the component signal having the highest amplitude is selected as the main component signal.

20. The system of claim 15, wherein: Determining the frequency of the principal component signal is based on determining the time and phase differences between fractional phase transitions.

21. The system of claim 15, wherein: Determining a breathing rate of the patient includes determining an average breathing rate over a period of time.

Images