Several types of sensors have been available in off-the-shelf mobile devices, including motion, magnetic, vision, acoustic, and location sensors. This paper focuses on the fusion of the data acquired from motion and magnetic sensors, i.e., accelerometer, gyroscope and magnetometer sensors, for the recognition of Activities of Daily Living (ADL) using pattern recognition techniques. The system developed in this study includes data acquisition, data processing, data fusion, and artificial intelligence methods. Artificial Neural Networks (ANN) are included in artificial intelligence methods, which are used in this study for the recognition of ADL. The purpose of this study is the creation of a new method using ANN for the identification of ADL, comparing three types of ANN, in order to achieve results with a reliable accuracy. The best accuracy was obtained with Deep Learning, which, after the application of the L2 regularization and normalization techniques on the sensors' data, reports an accuracy of 89.51%.[13] that uses only the accelerometer sensor, this study proposes the creation of two different methods for the recognition of ADL using different number of sensors in order to adapt the method according to the number of sensors available. Firstly, it proposes the fusion of the data acquired from the accelerometer, and the magnetometer sensors. Secondly, it proposes the fusion of the data acquired from the accelerometer, gyroscope, and magnetometer sensors. The ADL proposed for the recognition are running, walking, going upstairs, going downstairs, and standing, consisting this research on the analysis of the performance of three types of ANN, such as Multilayer Perception (MLP) with Backpropagation, Feedforward neural network with Backpropagation, and Deep Learning. A dataset used for this research is composed by the sensors' data acquired in several experiments with people aged between 16 and 60 years old, distinct lifestyles, and a mobile device in the front pocket of their pants, performing the proposed ADL. This research was conducted with the use of three Java libraries, such as Neuroph [14], Encog [15], and DeepLearning4j [16], and different datasets of features, in order to identify the best dataset of features and type of ANN for the recognition of ADL, verifying that the best accuracy for the recognition of ADL with the two different methods proposed was achieved with Deep Learning methods.This paragraph concludes the section 1, and this paper is organized as follows: Section 2 summarizes the literature review for the use of data fusion techniques with accelerometer, gyroscope, and magnetometer sensors; Section 3 presents the methods used on each stage of the architecture proposed. The results obtained are presented in the Section 4, presenting the discussion about these results in the Section 5. The conclusions of this study are presented in the Section 6.
Related WorkData fusion techniques may be used with the data acquired from motion and magnetic sensors available in the off-the-shelf mobile devic...