Quantitative falls risk assessment through inertial sensors and pressure sensitive platform

1. A method for assessing falls risk using a system comprising at least one inertial sensor and at least one processor, the method comprising:

performing a standing balance assessment of a person by:

(a) receiving inertial sensor data using the at least one processor from the at least one inertial sensor configured to be attached to a lumbar region of the person; and

(b) receiving pressure sensor data using the at least one processor from a pressure sensor matrix that is detecting pressure exerted by the person on the pressure sensor matrix, wherein the pressure sensor matrix comprises a plurality of pressure sensors; then

generating center of pressure data for the person, using the at least one processor, based on at least the received pressure sensor data from the standing balance assessment;

deriving a feature value indicative of when the person has fallen during a predetermined interval using the at least one processor;

training a classifier model using the at least one processor to recognize the feature value, based on the received inertial sensor data and the center of pressure data, wherein the classifier model is configured to generate a prediction risk of future falls based on the center of pressure data and based on the inertial sensor data as input into the classifier model, and

using the at least one processor to generate falls risk estimate based on the prediction risk of future falls to reduce the risk of future falls of the person.

2. The method of claim 1 , wherein the classifier model is a support vector machine model.

3. The method of claim 2 , wherein training the classifier model comprises relating values related to a statistical property of the inertial sensor data to the feature value indicative of when the person has fallen.

4. The method of claim 3 , wherein the method further comprises determining using the at least one processor a root mean squared value of the inertial sensor data, wherein the inertial sensor data includes multi-axial acceleration data or multi-axial angular velocity data, and wherein the root mean squared value is used by the classifier model to determine the falls risk estimate indicative of falls risk.

5. The method of claim 3 , wherein the method further comprises determining using the at least one processor a spectral edge frequency of the inertial sensor data, wherein the spectral edge frequency is determined based on a power spectrum of the inertial sensor data, the power spectrum being based on a frequency profile of a portion of the inertial sensor data, and wherein at least half of the power spectrum's power is at or below the spectral edge frequency.

6. The method of claim 3 , wherein training the classifier model comprises determining a value of a pressure-based feature and relating the value of the pressure-based feature to the feature value indicative of when the person has fallen, wherein the pressure-based feature is related to a statistical property of the center of pressure data.

7. The method of claim 6 , wherein the statistical property of the center of pressure data is indicative of how the person's center of pressure changes during an interval in which the pressure sensor data is received.

8. The method of claim 6 , wherein training the classifier model comprises relating the feature value indicative of when the person has fallen with the person's age, gender, height, or weight.

9. The method of claim 1 , wherein the method further comprises receiving inertial sensor data and pressure sensor data from a plurality of persons and training the classifier model to recognize feature patterns associated with falling using the at least one processor, and wherein the training of the classifier model is based on the inertial sensor data and the pressure sensor data from the plurality of persons, and wherein the plurality of persons includes only men or only women.

10. The method of claim 1 , wherein the method further comprises receiving via the least one processor inertial sensor data and pressure sensor data from a plurality of persons, wherein the training of the classifier model is based on the inertial sensor data and the pressure sensor data from the plurality of persons, the plurality of persons all having eyes open or all having eyes closed during the receiving of the inertial sensor data and the pressure sensor data.

11. The method of claim 3 , wherein the method further comprises determining using the at least one processor a root mean squared value of the inertial sensor data, wherein the inertial sensor data includes multi-axial acceleration data and multi-axial angular velocity data, and wherein the root mean squared values of multi-axial acceleration data and multi-axial angular velocity data are used by the classifier model to determine the falls risk estimate indicative of falls risk.

12. A system for assessing falls risk, comprising:

an inertial sensor configured to be attached to a lumbar region of a person;

a pressure sensor matrix, the pressure sensor matrix configured to sense a pressure exerted by the person thereon, wherein the pressure sensor matrix comprises a plurality of pressure sensors;

at least one processor configured to perform a standing balance assessment of a person by being configured to: (a) receive inertial sensor data from the inertial sensor and (b) receive pressure sensor data from the pressure sensor matrix, wherein the at least one processor is further configured to:

generate center of pressure data for the person based on at least the received pressure sensor data from the standing balance assessment;

derive a feature value from the received inertial sensor data and/or the received pressure sensor data indicative of when the person has fallen during a predetermined interval,

train a classifier model using the at least one processor to recognize the feature value, based on the received inertial sensor data and the center of pressure data, wherein the classifier model is configured to generate a prediction risk of future falls based on the center of pressure data and based on the inertial sensor data input into the classifier model, and

use the at least one processor to generate falls risk estimate based on the prediction risk of future falls to reduce the risk of future falls of the person.

13. The system of claim 12 , wherein the classifier model is a support vector machine model.

14. The system of claim 13 , wherein the at least one processor is configured to relate values related to a statistical property of the inertial sensor data to the feature value indicative of when the person has fallen.

15. The system of claim 14 , wherein the at least one processor is further configured to determine a root mean squared value of the inertial sensor data, wherein the inertial sensor data includes acceleration data or angular velocity data, and wherein the root mean squared value is used by the classifier model to determine the falls risk estimate indicative of falls risk.

16. The system of claim 14 , wherein the at least one processor is further configured to determine a spectral edge frequency of the inertial sensor data, wherein the spectral edge frequency is determined based on a power spectrum of the inertial sensor data, the power spectrum being based on a power spectral density of a portion of the inertial sensor data, and wherein at least half of the signals' power is at or below the spectral edge frequency.

17. The system of claim 14 , wherein the at least processor is configured to determine a value of a pressure-based feature and to relate the value of the pressure-based feature to the feature value indicative of when the person has fallen, wherein the pressure-based feature is related to a statistical property of the center of pressure data.

18. The system of claim 17 , wherein the statistical property of the center of pressure data is indicative of how the person's center of pressure changes during an interval in which the pressure sensor data is received.

19. The system of claim 17 , wherein the at least processor is configured to train the classifier model by relating the feature value indicative of falls risk of the person with the person's age, gender, height, or weight.

20. The system of claim 12 , wherein the at least processor is further configured to receive inertial sensor data and pressure sensor data from a plurality of persons and training the classifier model to recognize feature patterns associated with falling, and wherein the training of the classifier model is based on the inertial sensor data and the pressure sensor data from the plurality of persons, and wherein the plurality of persons includes only men or only women.

21. A method for assessing falls risk using a system comprising at least one inertial sensor and at least one processor, comprising:

performing a standing balance assessment of a person by:

(a) receiving inertial sensor data by the at least one processor from the at least one inertial sensor configured to be attached to a lumbar region of the person; and

(b) receiving pressure sensor data using the at least one processor from a pressure sensor matrix that is detecting pressure exerted by the person on the pressure sensor matrix, wherein the pressure sensor matrix comprises a plurality of pressure sensors; then

generating center of pressure data for the person, using the at least one processor, based on at least the received pressure sensor data from the standing balance assessment; and

training a classifier model using the at least one processor to recognize a feature value indicative of when the person has fallen during a predetermined interval using the at least one processor, based on the received inertial sensor data and the center of pressure data, wherein the classifier model is configured to generate a prediction risk of future falls based on the center of pressure data and based on the inertial sensor data input into the classifier model, and

using the at least one processor to generate falls risk estimate based on the prediction risk of future falls to reduce the risk of future falls of the person.