Una contribución a la evaluación de la adherencia a hábitos de vida saludables basado en aplicaciones móviles

The adherence to a healthy lifestyle plays a key role for increasing life expectancy and living better. The main habits of healthy lifestyle are: physical activity, diet and sleep quality. Nowadays, many people use a smartphone and carry it all day. The objective of this thesis is to demonstrate the feasibility of the evaluation of the adherence to a healthy lifestyle by means of smartphone applications and sensors, whether internal or externally connected. On the one hand, the accelerometer sensor is used to evaluate the physical activity and the associated energy expenditure. In previous research, we can found classifiers of physical activity from data of this sensor but the measurements were performed in a laboratory environment or with smartphone fixed to a specific position. From the collected data during a week of 26 subjects, a 75.6% of F1-score of the classification of activities has been achieved and a 3.18% of error in the energy expenditure estimation. On the other hand, the heart rate variability (HRV) can serve as indicator of behaviours related to health and physical condition. A system has been designed to evaluate the HRV using the rear camera of the smartphone as a sensor. For this purpose, the photoplethysmography technique has been used. In previous research, this technique has been used in smartphones in order to obtain the heart rate but it has not been assessed the beat-to-beat HRV. The proposed system uses the GPU for image processing in real time. The obtained results have been compared with the electrocardiogram and with a reference photoplethysmography device. For that, the standard deviation of error made for the beat detection and the level of agreement of HRV indices have been assessed. This assessment has been performed with 23 subjects and the results obtained for two different smartphone models have been compared. The standard deviation of error of heart rate detection between smartphone and electrocardiogram obtained was 5.4 ms, while between electrocardiogram and reference photoplethysmography device was 4.9 ms. On the other hand, an application for the ensemble analysis of physical activity and heart rate has been developed. Using this application, the data of 11 people was collected, they have divided in two groups of 5 and 6 people during 3 and 6 weeks respectively. From the analysis of the collected data, it has been found that the level of physical activity decreases over the time and there is some association between the constancy of the practice of physical activity and changes in mood. However, these association should be taken with caution due to the reduced number of subjects which were involved in this study. Therefore, the developed system is a starting point in order to evaluate the adherence to a healthy lifestyle in a unified way with an single application. Finally, one of the consequences of leading an unhealthy lifestyle is the decreasing of quality of sleep that can cause daytime sleepiness. This can be a serious health risk, for example if it occurs while driving. To prevent this, an early drowsiness detection system based on the analysis of respiratory signal and respiratory rate variability has been proposed and validated. The designed algorithm has been assessed with 15 subjects and a specificity of 96.6% and a sensitivity of 90.3% has been obtained.