The field of human activity recognition (HAR) is advancing with the integration of cutting-edge technologies such as artificial intelligence and sensor networks. These developments enable accurate detection and interpretation of human behavior, driven by advances in AI and machine learning. A growing number of electronic devices and applications contribute to a comprehensive understanding of the three pillars of HAR – AI, various devices and methodologies and their applications. This review consolidates HAR research and highlights its processes, applications, and device comparisons. Key findings include the importance of non-invasive, real-time sensor-based approaches, particularly in healthcare for the detection of behavioral abnormalities in the elderly. The study highlights Adam as an outstanding optimizer in HAR applications, who predicts continued expansion and development in various industries.particularly in the healthcare sector.