Abstract
Human activity recognition (HAR) techniques can significantly contribute to the enhancement of health and life care systems for elderly people. These techniques, which generally operate on data collected from wearable sensors or those embedded in most smart phones, have therefore attracted increasing interest recently. In this paper, a random forest-based classifier for human activity recognition is proposed. The classifier is trained using a set of time-domain features extracted from raw sensor data after being segmented into windows of 5 seconds duration. A detailed study of model parameter selection is presented using the statistical t-test. Several simulation experiments are conducted on the WHARF accelerometer benchmark dataset, to compare the performance of the proposed classifier to support vector machines (SVM) and Artificial Neural Network (ANN). The proposed model shows high recognition rates for different activities in the WHARF dataset compared to other classifiers using the same set of features. Furthermore, it achieves an overall average precision of 86.1% outperforming the recognition rate of 79.1% reported in the literature using Convolution Neural Networks (CNN) for the WHARF dataset. From a practical point of view, the proposed model is simple and efficient. Therefore, it is expected to be suitable for implementation in hand-held devices such as smart phones with their limited memory and computational resources.
Highlights
In daily life, a person performs diverse set of activities such as standing up, sitting down, walking, climbing stairs, etc
The contribution of this work can be highlighted as follows: (1) introducing Random Forest (RF)-based effective and efficient Human Activity Recognition (HAR) system with average precision of 86.1% and average accuracy of 84.8% which improves the state-of-the-art rate of 79.1% for Wearable Human Activity Recognition Folder (WHARF) dataset, (2) testing the proposed system on the challenging WHARF datase which is considered in only few studies in literature [23] and [24], (3) discussing the practical implementation issues of proposed system which is important in case of further system application on smart devices, and (4) conducting sensitivity analysis of important system components to determine the optimal settings for proposed system
The results show that support vector machines (SVM) and Artificial Neural Network (ANN) have better precision than random forest in some activities
Summary
A person performs diverse set of activities such as standing up, sitting down, walking, climbing stairs, etc. The relatively challenging Wearable Human Activity Recognition Folder (WHARF) dataset is extensively investigated This dataset is collected using a tri-axial accelerometer placed on the right wrist of subjects; it emulates a smart watch. The contribution of this work can be highlighted as follows: (1) introducing RF-based effective and efficient HAR system with average precision of 86.1% and average accuracy of 84.8% which improves the state-of-the-art rate of 79.1% for WHARF dataset, (2) testing the proposed system on the challenging WHARF datase which is considered in only few studies in literature [23] and [24], (3) discussing the practical implementation issues of proposed system which is important in case of further system application on smart devices, and (4) conducting sensitivity analysis of important system components to determine the optimal settings for proposed system.
Sign-in/Register to view highlights & summary Sign-in/Register to access full text options 
Free) 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call
