Fatigue estimation device, fatigue estimation method, and storage medium

1 . A fatigue estimation device comprising:

at least one memory configured to store instructions; and

at least one processor configured to execute the instructions to:

receive, from a sensor device, biological data including an R-R Interval of a test subject which have been measured by the sensor device;

acquiring attribute information including a muscle mass of the test subject;

perform a normalization process that adjusts differences based on respective attributes of test subjects by dividing the R-R Interval by the muscle mass, which is used as the attribute of the test subject, based on the muscle mass included in the attribute information of the test subject with respect to the R-R Interval included in the biological data of the test subject;

extract a heart rate variability which is an index of an autonomic neural tension as features, based on the biological data after the normalization process; and

estimate a fatigue level of the test subject based on the features, wherein

(a) the at least one memory further stores a trained fatigue estimation model generated using a dataset normalized according to physiological attributes and environmental parameters;

(b) the processor is configured to dynamically adjust the parameters in real time based on both the environmental information and a stored physiological attribute of the test subject;

(c) the normalization process comprises an adaptive normalization that compensates for inter-subject physiological variability and environmental interference in the biological data;

(d) the extracted features comprise time-domain features and frequency-domain features derived from the normalized biological data; and

(e) the processor is configured to estimate the fatigue level by inputting the extracted features into the trained fatigue estimation model, wherein the model outputs a fatigue score indicative of at least one of a physical fatigue state or a mental fatigue state of the test subject.

2 . The fatigue estimation device according to claim 1 , wherein

the attribute corresponds to a muscle mass of the test subject, and

the processor performs the normalization process based on the muscle mass of the test subject.

3 . The fatigue estimation device according to claim 1 , wherein

the attribute is at least any one of an age, a gender, an occupation, or a type of sport which the test subject performs or position in the sport; and

the processor performs the normalization process based on at least any one of the age, the gender, the occupation, or the type of sport which the test subject performs or position in the sport.

4 . The fatigue estimation device according to claim 1 , wherein the processor calculates the fatigue level based on a fatigue estimation model which has been trained to output the fatigue level estimated regarding the test subject of the biological data in a case of inputting the biological data or the features of the biological data.

5 . The fatigue estimation device according to claim 1 , wherein the processor adjusts parameters used for the normalization process based on external inputs.

6 . The fatigue estimation device according to claim 1 , wherein the processor is further configured to

determine a state concerning a fatigue of the test subject; and

output control means configured to output regarding a result of a determination.

7 . A fatigue estimation method performed by a computer, the fatigue estimation method comprising:

receiving, from a sensor device, biological data including an R-R Interval of a test subject which have been measured by the sensor device;

acquiring attribute information including a muscle mass of the test subject;

performing a normalization process that adjusts differences based on respective attributes of test subjects by dividing the R-R Interval by the muscle mass, which is used as the attribute of the test subject, based on the muscle mass included in the attribute information of the test subject with respect the R-R Interval included in the biological data of the test subject;

extracting a heart rate variability which is an index of an autonomic neural tension as features, based on the biological data after the normalization process; and

estimating a fatigue level of the test subject based on the features, wherein

(a) the at least one memory further stores a trained fatigue estimation model generated using a dataset normalized according to physiological attributes and environmental parameters;

(b) the processor is configured to dynamically adjust the parameters in real time based on both the environmental information and a stored physiological attribute of the test subject;

(c) the normalization process comprises an adaptive normalization that compensates for inter-subject physiological variability and environmental interference in the biological data;

(d) the extracted features comprise time-domain features and frequency-domain features derived from the normalized biological data; and

(e) the processor is configured to estimate the fatigue level by inputting the extracted features into the trained fatigue estimation model, wherein the model outputs a fatigue score indicative of at least one of a physical fatigue state or a mental fatigue state of the test subject.

8 . A non-transitory computer readable recording medium storing a program, the program causing a computer to perform a process comprising:

receiving, from a sensor device, biological data including an R-R Interval of a test subject which have been measured by the sensor device;

acquiring attribute information including a muscle mass of the test subject;

performing a normalization process that adjusts differences based on respective attributes of test subjects by dividing the R-R Interval by the muscle mass, which is used as the attribute of the test subject, based on the muscle mass included in the attribute information of the test subject with respect to the R-R Interval included in the biological data of the test subject;

extracting a heart rate variability which is an index of an automatic neural tension as features, based on the biological data after the normalization process; and

estimating a fatigue level of the test subject based on the features, wherein

(a) the at least one memory further stores a trained fatigue estimation model generated using a dataset normalized according to physiological attributes and environmental parameters;

(b) the processor is configured to dynamically adjust the parameters in real time based on both the environmental information and a stored physiological attribute of the test subject;

(c) the normalization process comprises an adaptive normalization that compensates for inter-subject physiological variability and environmental interference in the biological data;

(d) the extracted features comprise time-domain features and frequency-domain features derived from the normalized biological data; and

(e) the processor is configured to estimate the fatigue level by inputting the extracted features into the trained fatigue estimation model, wherein the model outputs a fatigue score indicative of at least one of a physical fatigue state or a mental fatigue state of the test subject.

9 . The fatigue estimation device according to claim 1 , wherein the processor adjusts parameters used for the normalization process based on environmental information representing an environment at time of a measurement of the biological data.