Augmented Reality for Assistive Maintenance and Real-Time Failure Analysis in Industries
Jaya Sai Kiran Patibandla, Sunit Kumar Adhikary, Janmesh Ukey · 2020
We present a methodology to solve the problem of maintenance for any machinery using augmented reality (AR) guided assistive systems. The same system can be used to implement real time fault analytics system. Given a visual feed of object, our methodology can accurately determines, tracks and maps the object in the world. The objects are categorized as stationary and nonstationary by the user. The first method presented will address the maintenance solution for object which are stationary using spatial mapping and localization technique. The second method presented will address the problem using a deep learning model that predicts accurate pose of the object in world space. The instruction sets and the necessary animations are overlaid on the objects. This methodology of ours to find the object pose is robust to occlusion, lighting conditions and works in real time on computationally inexpensive hardware.