Attention-based Multihead Deep Learning Framework for online activity monitoring with Smartwatch Sensors

The expeditious propagation of Internet Things (IoT) technologies implanted in different smart devices such as smartphones and smartwatches have a ubiquitous consequence on the modern population. These are employed to collect data aid tracking analyzing users’ daily activities using various human activity monitoring recognition (HAR) techniques. However, most current HAMR approaches rely exploratory case-based shallow feature learning architectures, which endeavor recognize correctly real-world situations. To address this issue, we offer unique strategy for that leverages attention mechanism with multi-head convolutional neural networks (CNNs) Long-Short-Term-Memory (LSTM). accuracy detection is improved presented method by integrating into CNNs followed LSTM better extraction selection. Verification investigations carried out from University California (UCI) repository, publicly available. results show our proposed framework more accurate than frameworks both 10-fold leave-one-subject-out cross-validation. Finally, can real-time, regardless type device.