A Deep Machine Learning Method for Concurrent and Interleaved Human Activity Recognition

Abstract

Human activity recognition has become an important research topic within the field of pervasive computing, ambient assistive living (AAL), robotics, health-care monitoring, and many more. Techniques for recognizing simple and single activities are typical for now, but recognizing complex activities such as concurrent and interleaving activity is still a major challenging issue. In this paper, we propose a two-phase hybrid deep machine learning approach using bi-directional Long-Short Term Memory (BiLSTM) and Skip-Chain Conditional random field (SCCRF) to recognize the complex activity. BiLSTM is a sequential generative deep learning inherited from Recurrent Neural Network (RNN). SCCRFs is a distinctive feature of conditional random field (CRF) that can represent long term dependencies. In the first phase of the proposed approach, we recognized the concurrent activities using the BiLSTM technique, and in the second phase, SCCRF identifies the interleaved activity. Accuracy of the proposed framework against the counterpart state-of-art methods using the publicly available datasets in a smart home environment is analyzed. Our experiment’s result surpasses the previously proposed approaches with an average accuracy of more than 93%.

Body Temperature—Indoor Condition Monitor and Activity Recognition by MEMS Accelerometer Based on IoT-Alert System for People in Quarantine Due to COVID-19

“…It is necessary to realize this activity and update the temperature threshold (T TH ) for alert temperature because the high-temperature warning works depending on the comparison between these temperature parameters. Usually, the human activity recognition (HAR) algorithm is used for the prediction of the movement of a person based on sensor data [ 9 ] by using Artificial Intelligence such as Machine Learning [ 10 ] and Deep Learning [ 11 , 12 ]. It requires much training data and complexity.…”

Activity Recognition in Residential Spaces with Internet of Things Devices and Thermal Imaging

“…The main data source for the task of activity recognition is sensors. There are two types of sensors used for this purpose: one is external and the second is wearable [ 7 ]. Wearable sensors are used in various healthcare applications such as patient monitoring, fitness monitoring, oxygen level monitoring, heartbeat monitoring, etc.…”

Design of a Kitchen-Monitoring and Decision-Making System to Support AAL Applications

“…Another very important aspect of the AAL system is related to the appropriate analysis of the data received from sensors through the decision-making process. Specifically, pure sensor data are sufficient only for triggering alarms for critical situations such as fire detection [ 11 , 14 , 24 , 25 , 26 , 27 , 28 , 29 , 30 ], smoke [ 11 , 14 , 24 , 26 , 27 , 28 , 29 ], and gas leakage [ 11 , 24 , 25 , 29 ]. To obtain information about daily user activity, more complex analysis and decision-making systems are required.…”

“…Modern methods of the machine-learning have been used recently to recognize certain human activities as quickly and as reliably as possible, and to take necessary actions within the system itself. It is especially important to recognize more complex activities that can only be detected by the joint use of different, and often very diverse, sensors [ 29 ]. In addition, advanced and intelligent mining techniques can be used to detect irregularities within a system or to detect the malfunction of individual components [ 31 ].…”