VO2 estimation using 6-axis motion sensor with sports activity classification.

In this paper, we focus on oxygen consumption (VO2) estimation using 6-axis motion sensor (3-axis accelerometer and 3-axis gyroscope) for people playing sports with diverse intensities. The VO2 estimated with a small motion sensor can be used to calculate the energy expenditure, however, its accuracy depends on the intensities of various types of activities. In order to achieve high accuracy over a wide range of intensities, we employ an estimation framework that first classifies activities with a simple machine-learning based classification algorithm. We prepare different coefficients of linear regression model for different types of activities, which are determined with training data obtained by experiments. The best-suited model is used for each type of activity when VO2 is estimated. The accuracy of the employed framework depends on the trade-off between the degradation due to classification errors and improvement brought by applying separate, optimum model to VO2 estimation. Taking this trade-off into account, we evaluate the accuracy of the employed estimation framework by using a set of experimental data consisting of VO2 and motion data of people with a wide range of intensities of exercises, which were measured by a VO2 meter and motion sensor, respectively. Our numerical results show that the employed framework can improve the estimation accuracy in comparison to a reference method that uses a common regression model for all types of activities.

Applying neural network to VO2 estimation using 6-axis motion sensing data.

This paper focuses on oxygen consumption (VO2) estimation using 6-axis motion data (3-axis acceleration and 3-axis angular velocity) that are obtained from small motion sensors attached to people playing sports with different intensities. In order to achieve high estimation accuracy over a wide range of intensities of exercises, we apply neural network that is trained by experimental data consisting of the measured VO2 and motion sensing data of people with a wide range of intensities of exercises. We first investigate the gain brought by applying neural network by comparing its accuracy with an approach based on the linear regression model. Then, we analyze how much improvement the information on angular velocity can bring as compared with the estimation with the acceleration data alone. Our numerical results show that the employed framework exploiting neural network can improve the estimation accuracy in comparison to the linear regression model and the exploitation of information on the angular velocity plays an important role to improve the accuracy over higher intensities of exercises.