Edge-intelligent iot-based wearable device for detection of cravings in individuals

A wearable device for detecting craving of a subject, comprising:

• a housing including:
• at least one sensor selected from the group consisting of an accelerometer, a thermopile sensor, and a pulse oximeter sensor
• at least one subject button selected from the group consisting of a subject stress input button and a subject craving input button;
• and
• a microprocessor electrically connected to the at least one sensor and the at least one subject button;
• wireless communications between the microprocessor and at least one of a subject cell phone, a clinician, and a cloud,
• wherein:
• during a supervised period, the patient presses the input button while experiencing stress or cravings, and the microprocessor processes signals from the at least one sensor while the input button is pressed to determine training set classification rules;
• and
• during an unsupervised period:
• the microprocessor processes signals from the at least one sensor to determine characteristics of the unsupervised sensor data;
• and
• the microprocessor compares the characteristics of the unsupervised sensor data to the training set classification rules;
• and
• at least one element selected from the group consisting of the microprocessor, the clinician, and the cloud, provides alerts to the subject cell phone based on the comparison,
• and wherein:
• processing the signals from the at least one sensor and the at least one input buttons to determine training set classification rules during the supervised period comprises:
• producing training motion data in at least one dimension by sensing the training motion of the subject during a period of no drug use;
• generating training statistics of the training motion data during the period of no drug use;
• and
• determining classification rules for the training statistics during the period of no drug use;
• and
• processing signals from the at least one sensor to determine characteristics of the unsupervised sensor data comprises producing monitoring motion data in the at least one dimension by sensing monitoring motion of the subject during the period of non-supervision,
• and wherein:
• sensing the training motion comprises sensing training accelerometer data in at least one dimension;
• and
• sensing monitoring motion comprises sensing monitoring accelerometer data in the at least one dimension,
• and wherein:
• generating training statistics comprises computing training mean and training variance of the training accelerometer data and determining the classification rules of the training statistics comprises determining classification rules of the mean and the variance of the training mean and the training variance;
• and
• generating monitoring statistics comprises computing monitoring mean and monitoring variance of the monitoring accelerometer data and comparing the monitoring statistics comprises comparing the monitoring mean and the monitoring variance to the classification rules,
• and wherein the at least one dimension comprises at least two dimensions, and further including:
• determining training phase correlation between the at least two dimensions;
• determining training classification rules further based on the training phase correlation;
• determining monitoring phase correlation between the at least two dimensions;
• including comparing the monitoring phase correlation to the classification rules to determine if the drug cravings are present.

A wearable device for detecting stress or craving of a subject, comprising:

• a housing including:
• at least one sensor selected from the group consisting of an accelerometer, a thermopile sensor, and a pulse oximeter sensor
• at least one subject button selected from the group consisting of a subject stress input button and a subject craving input button;
• and
• a microprocessor electrically connected to the at least one sensor and the at least one subject button;
• the wearable device configured for wireless communications between the microprocessor and a subject cell phone, wherein:
• during a supervised period of no drug use:
• a subject wearing the wearable device and the pressing the input button while experiencing stress or cravings;
• the wearable device producing accelerometer data by sensing motion of the subject;
• transforming the accelerometer data into training amplitude data;
• creating training histogram data from the training amplitude data;
• fitting a training curve to the training histogram data;
• determining shape and scale classification rules based on the training curve;
• and
• during an unsupervised period:
• the microprocessor processes signals from the accelerometer to determine monitoring motion data;
• transforming the monitoring motion data into monitoring amplitude data;
• creating monitoring histogram data from the monitoring amplitude data;
• fitting a monitoring curve to the monitoring amplitude data;
• determining monitoring shape and scale based on the monitoring curve;
• comparing monitoring shape and scale to the shape and scale classification rules to determine if stress or cravings are present;
• and
• providing alerts to the subject cell phone when stress or cravings are present based on the comparison.