A Multiple Source Framework for the Identification of Activities of Daily Living Based on Mobile Device Data

The monitoring of the lifestyles may be performed based on a system for the recognition of Activities of Daily Living (ADL) and their environments, combining the results obtained with the user’s agenda. The system may be developed with the use of the off-the-shelf mobile devices commonly used, because they have several types of sensors available, including motion, magnetic, acoustic, and location sensors. Data acquisition, data processing, data fusion, and artificial intelligence methods are applied in different stages of the system developed, which recognizes the ADL with pattern recognition methods. The motion and magnetic sensors allow the recognition of activities with movement, but the acoustic sensors allow the recognition of the environments. The fusion of the motion, magnetic and acoustic sensors allows the differentiation of other ADL. On the other hand, the location sensors allows the recognition of ADL with large movement, and the combination of these sensors with the other sensors increases the number of ADL recognized by the system. This study consists on the comparison of different types of ANN for choosing the best methods for the recognition of several ADL, which they are implemented in a system for the recognition of ADL that combines the sensors’ data with the users’ agenda for the monitoring of the lifestyles. Conclusions point to the use of Deep Neural Networks (DNN) with normalized data for the identification of ADL with 85.89% of accuracy, the use of Feedforward neural networks with nonnormalized data for the identification of the environments with 86.50% of accuracy, and the use of DNN with normalized data for the identification of standing activities with 100% of accuracy, proving the reliability of the framework presented in this study.