Patent US10238335

Alertness prediction system and method
[+] Show Details

US10238335

1. A wearable device for monitoring and predicting alertness of an individual, the wearable device comprising: (8) (4)

one or more sensors configured to obtain information signals about the individual, the sensors comprising at least one of: a motion sensor configured to produce movement data and/or body position data of the individual,

a temperature sensor configured to produce distal skin temperature data of the individual, and

a heart rate monitor configured to produce heart rate data of the individual;

a memory configured to store: a default circadian rhythm configured to be refined with data derived from the information signals about the individual to generate an estimated circadian rhythm for the individual, and

a bio-mathematical model configured to generate a fatigue score for the individual;

a processor coupled to the one or more sensors and to the memory, configured to: receive the information signals about the individual including at least one of movement data, position data, distal skin temperature data, or heart rate data,

estimate a circadian rhythm of the individual by incorporating the information signals about the individual to refine the default circadian rhythm,

extract features from the information signals about the individual and the estimated circadian rhythm,

apply at least one pattern recognition algorithm or machine learning algorithm to the extracted features,

extract at least one coefficient from the extracted features using the at least one pattern recognition algorithm or machine learning algorithm,

apply a bio-mathematical model to the at least one extracted coefficient, and

generate the fatigue score for the individual from the at least one extracted coefficient using the bio-mathematical model; and

a support configured to support the one or more sensors, the memory, and the processor on the individual.

2. The wearable device of claim 1, wherein the bio-mathematical model is a two-process algorithm configured to predict alertness levels of the individual using assessments of sleep-wake homeostasis and the estimated circadian rhythm; and wherein the processor is further configured to: (1) (3)

make actigraphy determinations using the movement data and/or body position data,

assess, using the two-process algorithm, the sleep-wake homeostasis of the individual, including periods of sleep and wakefulness, based on the actigraphy determinations, and

combine the sleep-wake homeostasis assessments with the estimated circadian rhythm according to the two-process algorithm into the generation of the fatigue score for the individual.

3. The wearable device of claim 2, wherein the processor is further configured to: (0) (2)

4. The wearable device of claim 1, wherein the at least one coefficient includes at least one of a circadian rhythm coefficient (Φ), a wake/sleep coefficient (τ), a circadian rhythm weighting coefficient, or a wake/sleep weighting coefficient. (0)

5. The wearable device of claim 1, wherein the processor is further configured to: (0) (3)

6. The wearable device of claim 1, wherein the processor is further configured to: (1) (4)

obtain information signals about the movement and/or position of the individual;

determine actigraphy data for the individual using the information signals about the movement of the individual;

apply the bio-mathematical model to the actigraphy data, wherein the bio-mathematical model is chosen from one of an awake bio-mathematical sub-model and an asleep bio-mathematical sub-model; and

incorporate the actigraphy data into the generation of the fatigue score for the individual.

7. The wearable device of claim 6, wherein the processor is further configured to refine the actigraphy data using at least one of the information signals about the distal skin temperature of the individual or the heart rate of the individual. (0)

8. The wearable device of claim 1, wherein the extracted features include one or more of: markers indicating a shape of a circadian rhythm of the individual, information regarding sleep habits of the individual on both working days and free days, and general medical information about the individual. (0)

9. The wearable device of claim 1, wherein the processor is further configured to incorporate individual parameters about the individual into the extraction of the features when the individual parameters are entered into the wearable device. (0)

10. The wearable device of claim 1, wherein the processor is further configured to: (0) (4)

identify raw circadian data within the information signals about the individual;

identify non-circadian data caused by non-circadian events within the raw circadian data;

remove the non-circadian data from the raw circadian data to obtain refined circadian data; and

incorporate the refined circadian data into the estimation of the circadian rhythm of the individual.

11. The wearable device of claim 1, wherein the sensors of the wearable device are configured to collect environmental information signals about the individual's environment, including at least one of ambient light level data or ambient temperature data, and wherein processor is further configured to incorporate the environmental information signals into the estimation of the circadian rhythm. (2) (0)

12. The wearable device of claim 11, wherein: (0) (3)

13. The wearable device of claim 11, wherein the temperature sensor is further configured to produce ambient temperature data from the ambient temperature surrounding the individual, and wherein the processor is further configured to refine the circadian rhythm estimation with the bio-mathematical model by incorporating the ambient temperature data. (0)

14. A method for producing a fatigue score for an individual, the method comprising: (7) (7)

obtaining, with sensors of a wearable device, information signals about the individual including at least one of movement, position, distal skin temperature, or heart rate;

estimating, with a processor of the wearable device, a circadian rhythm of the individual from the information signals about the individual;

extracting, with the processor, features from the information signals about the individual and the estimated circadian rhythm;

applying, with the processor, at least one pattern recognition algorithm or machine learning algorithm to the extracted features;

extracting, with the processor and using the at least one pattern recognition algorithm or machine learning algorithm, at least one coefficient from the extracted features;

applying, with the processor, a bio-mathematical model to the extracted coefficients; and

generating, with the processor and using the bio-mathematical model, a fatigue score for the individual from the extracted coefficients.

15. The method of claim 14, wherein the method further comprises: (0) (3)

16. The method of claim 14, wherein the sensors of the wearable device are configured to: (0) (2)

17. The method of claim 14, wherein the method further comprises: (0) (4)

identifying, with the processor, raw circadian data within the information signals about the individual;

identifying, with the processor, non-circadian data caused by non-circadian events, within the raw circadian data;

removing, with the processor, the non-circadian data from the raw circadian data to obtain refined circadian data; and

incorporating, with the processor, the refined circadian data into the estimation of the circadian rhythm of the individual.

18. The method of claim 14, wherein the method further comprises entering, into the wearable device, individual parameters about the individual; and (0) (1)

19. The method of claim 14, wherein the at least one coefficient includes at least one of a circadian rhythm coefficient, a wake/sleep coefficient, a circadian rhythm weighting coefficient, or a wake/sleep weighting coefficient. (0)

20. The method of claim 14, wherein the method further comprises: (2) (4)

obtaining, with a motion sensor of the wearable device, information signals about the movement and/or position of the individual;

determining, with the processor, actigraphy data for the individual using the information signals about the movement of the individual;

applying, with the processor, the bio-mathematical model to the actigraphy data, wherein the bio-mathematical model is chosen from one of an awake bio-mathematical sub-model and an asleep bio-mathematical sub-model; and

incorporating, with the processor, the actigraphy data into the generation of the fatigue score for the individual.

21. The method of claim 20, wherein the method further comprises: (0) (1)

22. The method of claim 20, wherein the method further comprises: (2) (1)

determining, with the processor and using the actigraphy data, whether the individual is awake or asleep.

23. The method of claim 22, wherein the method further comprises: (0) (1)

24. The method of claim 22, wherein the method further comprises: (0) (1)

25. The method of claim 14, wherein the extracted features include one or more of: markers indicating a shape of a circadian rhythm of the individual, information regarding sleep habits of the individual on both working days and free days, and general medical information about the individual. (0)

View Abstract and Specification Size

PDF with Images and Document Face >

Full Text Publication >

Patent US10238335

Alertness prediction system and method

Patent Matrix® Search