Human Activity Recognition (HAR) is a crucial task in a variety of applications, including healthcare monitoring, smart homes, and fitness tracking. Accurate and real-time detection of human activities enables systems to adapt to user behavior, providing personalized and context-aware services. Traditional methods for HAR often rely on manual feature extraction or shallow learning models, which may not capture complex patterns in sensor data. This paper proposes an approach for improving Human Activity Recognition by utilizing Convolutional Neural Networks (CNNs), a deep learning technique renowned for its ability to learn hierarchical patterns in data.