The Role of Heart-Rate Variability Parameters in Activity Recognition and Energy-Expenditure Estimation Using Wearable Sensors
Human-activity recognition (HAR) and energy-expenditure (EE) estimation are major functions in the mobile healthcare system. Both functions have been investigated for a long time; however, several challenges remain unsolved, such as the confusion between activities and the recognition of energy-consuming activities involving little or no movement. To solve these problems, we propose a novel approach using an accelerometer and electrocardiogram (ECG). First, we collected a database of six
... ies (sitting, standing, walking, ascending, resting and running) of 13 voluntary participants. We compared the HAR performances of three models with respect to the input data type (with none, all, or some of the heart-rate variability (HRV) parameters). The best recognition performance was 96.35%, which was obtained with some selected HRV parameters. EE was also estimated for different choices of the input data type (with or without HRV parameters) and the model type (single and activity-specific). The best estimation performance was found in the case of the activity-specific model with HRV parameters. Our findings indicate that the use of human physiological data, obtained by wearable sensors, has a significant impact on both HAR and EE estimation, which are crucial functions in the mobile healthcare system. These two issues are closely related to each other in that the EE estimation is accurate assuming that the activities of the monitored person are properly recognized  . The most frequently used sensor in the mobile healthcare system is the tri-axial accelerometer. Single or multiple accelerometers are broadly used for the HAR problem and EE estimation. However, as mentioned in a recent review by Lara and Labrador  , physiological signals such as heart rate, respiration rate, and electrocardiogram (ECG) have attracted little interest. The specific reason we pay attention to physiological signals is that the information provided by the accelerometer is insufficient for recognition of some confusing activities in terms of acceleration. Furthermore, an accelerometer has a critical drawback in cases of little or no movement but with obvious energy consumption, for example, sedentary work. One previous study has proven that heart-rate variabilities reflect qualitative differences in static and dynamic activities  . Biomedical sensors for physiological signals have continuously developed. On the other hand, as reviewed by Liu and Liu, recent biomedical sensors have become wireless, portable, and wearable on the platform of mobile phone. Yet, the method of analyzing the collected physiological signals is still a challenge  . Based on these considerations, we expect that such physiological signals may provide us with additional information for better recognition of human activities and prediction of EE, even for such cases. To the best of our knowledge, no study has been performed yet to solve these drawbacks for both issues (HAR and EE estimation) with an approach that exploits human physiological signals. In this study, we aim to recognize human ambulatory activities and estimate EE using our database composed of accelerometer and physiological signals, collected from 13 voluntarily participating subjects in a laboratory environment. To investigate the role of physiological signals in both issues, we compare the recognition and estimation performances with and without ECG data. The organization of the paper is as follows. We first give a brief review of some of the existing approaches for HAR and EE estimation using wearable sensors. Then, we introduce our database and the wearable sensors used in this study. Next, we describe our approaches and experimental results of HAR and EE estimation. Finally, we conclude this paper with a discussion.