Dalarna University's logo and link to the university's website

This study provides a comprehensive exploration of deep learning architectures for human activity recognition (HAR), focusing on hybrid models leveraging Convolutional Neural Networks (CNN) with Long-Short-Term Memory (LSTM)) and a range of alternative deep learning framework. The main goal is to evaluate the performance and effectiveness of the hybrid CNN-LSTM model compared to independent models such as Gated Recurrent Units (GRU), Recurrent Neural Networks (RNN), and traditional CNN architectures. By examining multiple models, this study aims to elucidate the advantages and disadvantages of each approach to accurately identify and classify human activities. The study examines the nuanced capabilities of each model, exploring their respective abilities to capture the spatial and temporal dependencies inherent in activity data. Our results not only demonstrate the superior accuracy of the hybrid model, but also highlight the potential for real world applications. © The Author(s), under exclusive license to Springer Nature Switzerland AG 2024.