Method and device that generates a respiration signal

A wearable device for determining a respiratory signal using motion data, comprising:

• at least one of an accelerometer or gyroscope in communication with a user;
• and
• a non-transitory computer-readable medium comprising a program configured to:
• determine movement data about the user of the wearable device over a current time window based on a plurality of movements measured by the at least one of the accelerometer or the gyroscope related to respiration of the user of the wearable device;
• determine a filter weight and associated bias for each of the plurality of movements over the current time window, wherein a respective filter weight is determined for each axis of the movement data according to the movement data, and the filter weight for each axis is determined based on a covariance of the movement data over the current time window between same or different axes of the movement data;
• dynamically filter the movement data based on a signal-to-noise ratio by applying the respective filter weight for each axis of the movement data with the associated bias, wherein the program is further configured to:
• apply the respective filter weight for each axis of the movement data;
• and
• reweigh the movement data using the associated bias determined for the current time window;
• determine a respiratory signal regarding the respiration of the user based on the dynamically filtered movement data by removing macro-movements from the dynamically filtered movement data without regard to a position of the at least one of the accelerometer or the gyroscope or a position of the user;
• and
• display an output based upon the respiratory signal, wherein the respiratory signal comprises at least one of a respiration rate or a respiration type of the user, wherein the movements measured by the at least one of the accelerometer or gyroscope comprise micro-movements and the macro-movements, and the program is configured to filter out the macro-movements when the movement data is dynamically filtered, and
• wherein the program is further configured to control the at least one of the accelerometer or gyroscope by:
• detecting a number of macro-movements over a time period;
• removing the macro-movements from the filtered movement data so that the macro-movements are not included in the respiratory signal;
• when the number of macro-movements is above a threshold amount, turning off or not activating the at least one of the accelerometer or gyroscope for measuring the micro-movements;
• and
• activating the at least one of the accelerometer or gyroscope to measure the micro-movements only when the number of macro-movements is below the threshold amount, and removing the macro-movements from the filtered movement data so that the macro-movements are not included in the respiratory signal.

A method determining a motion-related respiratory signal using a wearable device comprising at least one of an accelerometer or gyroscope, the method comprising:

• determining, by the wearable device, movement data about a user of the wearable device over a current time window based on a plurality of movements measured by the at least one of the accelerometer or gyroscope of the wearable device related to respiration of the user of the wearable device, wherein the movement data further comprises covariance of the movement data derived from the plurality of movements measured over the current time window between same or different axes of movement as measured by the wearable device;
• determining a signal-to-noise ratio for the covariance of the movement data over the current time window;
• denoising the covariance of the movement data based on the signal-to-noise ratio according to shared patterns of signals or noises across each axis of the movement data;
• determining a filter weight and associated bias for each of the plurality of movements over the current time window, wherein a respective filter weight is determined for each axis of the movement data and the associated bias according to the movement data;
• dynamically filtering the movement data by applying a respective filter weight for each axis of the movement data with the associated bias, further comprising:
• applying the respective filter weight for each axis of the movement data;
• and
• reweighing the movement data using the associated bias determined for the current time window;
• and
• determining a respiratory signal for output regarding respiration of the user based on the dynamically filtered movement data without regard to a position of the wearable device relative to the user or a position of the user, wherein the respiratory signal comprises at least one of a respiration rate or a respiration type of the user, wherein the movements measured by the at least one of the accelerometer or gyroscope comprise micro-movements and macro-movements, and the macro-movements are filtered out when the movement data is dynamically filtered, and
• wherein the method further comprises:
• detecting a number of macro-movements over a time period;
• when the number of macro-movements is above a threshold amount, turning off or not activating the at least one of the accelerometer or gyroscope for measuring the micro-movements;
• and
• activating the at least one of the accelerometer or gyroscope to measure the micro-movements only when the number of macro-movements is below the threshold amount.