Spectroscopic monitoring for the measurement of multiple physiological parameters

Abstract

The present disclosure relates to devices, systems, methods and computer program products for continuously monitoring, diagnosing and providing treatment assistance to patients using sensor devices, location-sensitive and power-sensitive communication systems, analytical engines, and remote systems. The method of non-invasively measuring multiple physiological parameters in a patient includes collecting photoplethysmograph (PPG) signal data from a wearable sensor device, applying one or more filters to correct the signal data and extracting a plurality of features from the corrected data to determine values for blood glucose, blood pressure, SpO2, respiration rate, and pulse rate of the patient. An alert may be automatically sent to one or more computing devices when the value falls outside a custom computed threshold range for the patient. The method offers ease of usage, allows continuous real-time monitoring of the patient in any setting for timely intervention, and results in improved accuracy of the signal data.

What is claimed is: 
     
         1 . A computer-implemented method of non-invasively measuring multiple physiological parameters for health monitoring, comprising:
 receiving, by a local computing device or a remote computing device, sensor data from a wearable sensor device attached to a patient's body, the sensor data comprising photoplethysmograph (PPG) signal data;   applying, by the local computing device or the remote computing device, one or more filters to remove motion artifacts, noise related interferences, effects of shivering, applied pressure, horizontal or vertical movements associated with the received sensor data;   extracting, by the local computing device or the remote computing device, a plurality of features from the PPG signal data, the plurality of features comprising at least systolic duration, diastolic duration, systolic slope, diastolic slope, pulse duration, overall mean, peak amplitude, left half and right half;   predicting, by a classifier associated with the remote computing device, values for blood glucose, blood pressure, SpO2, respiration rate, pulse rate, or a combination thereof, from the extracted plurality of features based on historical PPG signal data of the patient and one or more additional features including age, gender and disease status of the patient; and   sending an alert to one or more computing devices when the values for blood glucose, blood pressure, SpO2, respiration rate, pulse rate, or a combination thereof, falls within or above a computed threshold range for the patient.   
     
     
         2 . The method of  claim 1 , wherein motion artifacts from the PPG signal is removed using low pass Butterworth filtering, wavelet transform and thresholding. 
     
     
         3 . The method of  claim 1 , wherein external interferences on the PPG signal is corrected using one or more additional sensors present in the wearable device or the local computing device. 
     
     
         4 . The method of  claim 1 , wherein the alert comprises a summary of the patient's physiological parameters. 
     
     
         5 . The method of  claim 1 , wherein the local computing device is connected to one or more additional local computing devices for performing first level of sensor signal analysis, context aware monitoring, resilient communication, and context aware prioritization. 
     
     
         6 . The method of  claim 1 , wherein the one or more additional features are extracted from a hospital information system (HIS). 
     
     
         7 . The method of  claim 1 , wherein the threshold range for the patient is computed further based on one or more additional sensors present in the wearable device or the local computing device. 
     
     
         8 . The method of  claim 1 , wherein the threshold range is computed by a machine learning module trained to detect anomalies based on multiple factors. 
     
     
         9 . The method of  claim 1 , wherein the alert is a moderate alert when the value is within the computed threshold for the patient and wherein the alert is a severe alert when the value is above the computed threshold for the patient. 
     
     
         10 . A wearable, non-invasive, health-monitoring IoT device for use in the method of  claim 1 . 
     
     
         11 . A non-invasive remote health monitoring system for measuring multiple physiological parameters, comprising:
 one or more processing units; and   one or more memory units coupled to the one or more processing units; wherein the one or more memory units comprises:
 a signal analytics module configured to:
 receive sensor data of a patient from a wearable, non-invasive IoT sensor device, the sensor data comprising photoplethysmograph (PPG) signal data; 
 apply one or more filters to remove motion artifacts, noise related interferences, effects of shivering, applied pressure, horizontal or vertical movements associated with the received sensor data; and 
 extract a plurality of features from the PPG signal data, the plurality of features comprising at least systolic duration, diastolic duration, systolic slope, diastolic slope, pulse duration, overall mean, peak amplitude, left half and right half; 
 
 a machine learning module configured to:
 predict values for blood glucose, blood pressure, SpO2, respiration rate, pulse rate, or a combination thereof, from the extracted plurality of features based on historical PPG signal data of the patient and one or more additional features including age, gender and disease status of the patient; and 
 compute a threshold range for the patient; 
 
 an alert module configured to send an alert to one or more computing devices when the values for blood glucose, blood pressure, SpO2, respiration rate, pulse rate, or a combination thereof, falls within or above the computed threshold range for the patient; and 
 a summarization module configured to display a summary of the patient health status on the one or more computing device. 
   
     
     
         12 . The system of  claim 11 , wherein the wearable sensor device comprises:
 an optical sensor unit comprising a LED source coupled to a photodetector, the optical sensor unit configured to obtain photo-plethysmograph (PPG) signal data from a patient's body using near infrared (NIR) spectroscopy, wherein the LED source comprises at least a red LED source configured to be detected by the photodetector at  660 nm and an infrared (IR) LED source configured to be detected by the photodetector at  910  nm;   an analog front end (AFE) unit configured to convert the received PPG signal data to digital signal data;   a power source; and   a microcontroller configured to wirelessly transmit the digital signal data to a computing device for determining physiological parameters of the patient.   
     
     
         13 . The system of  claim 10 , wherein the machine learning module is trained to detect anomalies based on multiple factors. 
     
     
         14 . The system of  claim 11 , wherein the alert is a moderate alert when the value is within the computed threshold for the patient and wherein the alert is a severe alert when the value is above the computed threshold for the patient 
     
     
         15 . A non-transitory machine-readable storage medium storing instructions which, when executed by one or more processors, cause the one or more processors to perform operations comprising:
 receiving, by a local computing device or a remote computing device, sensor data from a wearable sensor device attached to a patient's body, the sensor data comprising photoplethysmograph (PPG) signal data;   applying, by the local computing device or the remote computing device, one or more filters to remove motion artifacts, noise related interferences, effects of shivering, applied pressure, horizontal or vertical movements associated with the received sensor data;   extracting, by the local computing device or the remote computing device, a plurality of features from the PPG signal data, the plurality of features comprising at least systolic duration, diastolic duration, systolic slope, diastolic slope, pulse duration, overall mean, peak amplitude, left half and right half;   predicting, by a classifier associated with the remote computing device, values for blood glucose, blood pressure, SpO2, respiration rate, pulse rate, or a combination thereof, from the extracted plurality of features based on historical PPG signal data of the patient and one or more additional features including age, gender and disease status of the patient; and   sending an alert to one or more computing devices when the values for blood glucose, blood pressure, SpO2, respiration rate, pulse rate, or a combination thereof, falls within or above a computed threshold range for the patient.