Point Cloud-Based Real-Time 3D Object Detection for Predictive Analytics of Safety Incidents in Manufacturing Industry

Autonomous vehicles and the associated technologies have significantly enhanced safety and convenience within the realm of transportation, garnering considerable attention from the research community. Consequently, there's a growing need to investigate their performance across various environments. This paper aims to make a valuable contribution towards minimizing vehicular accidents, optimizing traffic flow and mobility, and enhancing the safety of autonomous vehicles deployed in manufacturing settings. The prevailing 3D object detection algorithms exhibit certain limitations, particularly in their ability to predict safety-related incidents within manufacturing environments. As a result, the development of vehicle safety control technology that can identify risk factors through real-time video analysis becomes imperative. The approach presented in this study offers a novel perspective on predicting safety-related accidents within manufacturing environments. It achieves this by implementing an algorithm that strikes a balance between real-time processing and accuracy. Furthermore, this proposed algorithm demonstrates scalability, paving the way for future research avenues to enhance the model's sophistication and integrate new technologies better suited for manufacturing environments. This ongoing research holds promise in further elevating safety standards within the manufacturing sector.