ACTION-AWARE LABOR TIME STUDY: LEVERAGING DEEP ACTION RECOGNITION FOR OPTIMIZED WORKFORCE MANAGEMENT IN SMART WAREHOUSES

This research examines the application of Action-Aware Labor Time Study (AALTS), which is driven by deep action recognition, as a tool of workforce optimization within smart warehouses. Conventional labor time studies are based on labor observers, therefore limiting accuracy and scalability, as well as observation in real-time. In comparison, AALTS detects and classifies worker activities, namely picking, packing and walking, from video footage by using sophisticated computer vision models automatically. An analysis of secondary data from existing implementations reveals several major areas of improvement in task tracking accuracy, labor efficiency, and idle time detection for this research. It also illustrates such challenges as infrastructure cost, employee privacy, and the need for model upgrade. The results indicate that when properly implemented ethically and strategically AALTS can be a powerful tool to increase operational transparency and data-driven decision-making in logistics settings. This paper is adding to the developing discipline of AI-enabled workforce analytics by offering a systematic review of how deep learning could change labor performance measurement in a world of Industry 4.0.