SMARTMOTIONNET: LEVERAGING CNN-LSTM FOR ROBUST HUMAN ACTIVITY DETECTION

Abstract

Human Activity Recognition (HAR) has gained significant attention due to its applications in healthcare, smart homes, and human-computer interaction. Traditional HAR systems often rely on handcrafted features and shallow learning models, which may not effectively capture complex patterns in sensor data. This paper presents SmartMotionNet, a novel hybrid Convolutional Neural Network (CNN) and Long Short-Term Memory (LSTM) architecture designed to enhance HAR performance. The CNN component automatically extracts spatial features from raw sensor data, while the LSTM captures temporal dependencies, enabling accurate recognition of dynamic human activities. Experimental evaluations demonstrate that SmartMotionNet outperforms existing methods in terms of accuracy and robustness, making it a promising solution for real-time HAR applications.